.
c .3112}.
I... .2 t .5533!
x :u a 3....

 

 

 

 

 

 

 

 

 

LIBRARY
Michigan State
University

 

This is to certify that the
thesis entitled

IN VITRO BIOMECHANICAL ANALYSIS OF SEVERAL
JOINT AND FRACTURE FIXATION TECHNIQUES

presented by
Michael T. Sinnott

has been accepted towards fulfillment
of the requirements for the

MS. degree in Mechanical Engineering

 

 

 

flwaw

major Professor’s Signature

2212/- 4i 0700.?

Date

MSU is an affirmative-action, equal-opportunity employer

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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K lProj/AccaPres/ClRC/DaleDue Indd

IN VITRO BIOMECHANICAL ANALYSIS OF SEVERAL JOINT AND
FRACTURE FIXATION TECHNIQUES

By

Michael T. Sinnott

A THESIS

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

MASTER OF SCIENCE
Mechanical Engineering

2008

ABSTRACT

IN VITRO BIOMECHANICAL ANALYSIS OF SEVERAL JOINT AND FRACTURE
FIXATION TECHNIQUES

By
Michael T. Sinnott

The goal of biological fixation is to achieve a balance between the mechanical
stability of a joint or fracture repair and the biological preservation of the soft tissue
surrounding the site of the repair. The following studies focused on several new
techniques for the repair of joint instability as well as comminuted long bone fractures. In
Chapter 1, the in vitro biomechanical response of pancarpal arthrodesis constructs using
either a limited contact dynamic compression plate (LC-DCP), or a recently developed
hybrid plate (HP) was investigated. This study demonstrated the mechanical advantages
of the HPs over LC-DCPs, making them a viable alternative to LC-DCPs. In Chapter 2, a
tibial gap fracture model featuring a synthetic bone substitute developed by our group
was used to mechanically compare an investigational interlocking nail (ILN) system,
featuring extended modified bolt-pins coupled to a type-IA external skeletal fixator (ILN-
ESF), to standard bolted ILN (ILNb) constructs. Results showed that the substitution of
locking bolts with extended bolts connected to an ESF significantly reduced construct
compliance and overall deformation and eliminated the inherent slack of the ILNb. In
Chapter 3, a novel nail (ILNn), engineered by our group, was investigated. This study
demonstrated that the ILNn may represent a biomechanically more effective fixation
method than stande ILNs for the treatment of comminuted diaphyseal fractures as well
as a valid alternative to plate fixation. The results presented in this thesis demonstrate the

mechanical and biological advantages of several new techniques for biological fixation.

ACKNOWLEDGEMENTS

I would like to thank my academic advisor, Dr. Roger C. Haut for allowing me to
take part in the research at the Orthopaedic Biomechanics Laboratories (OBL) and for
continuously motivating me to succeed. I would also like to thank my research advisor,
Dr. Loi'c M. Déjardin for his expertise, dedication, and leadership throughout my research
at the College of Veterinary Medicine Orthopaedic Investigations Laboratory (COIL).
Above all, I would like to thank Dr. Haut and Dr. De’jardin for providing me with the
opportunity to achieve my goals and for their persistence to see each of these projects to
the end. I would also like to acknowledge Dr. Gary Cloud for his excellence in teaching
and for his guidance while serving on my committee. I would also like to thank Clifford
Beckett of the OBL and Mike McLean of the Engineering Research Machine Shop for
their knowledge, creativity, and willingness to help. Finally, I would like to acknowledge
everyone who worked with me at both the OBL and COIL; including Eric Meyer,
Jennifer Lansdowne, Dennis Ting, Keith Curcio, Eugene Kepich, Lynn Martin, Jerrod
Braman, and Tim Baumer.

I would like to thank my better half, Nicole Tarala, for inspiring me to achieve
and for patiently being at my side for every step of the way. I would like to thank my
siblings, Lory, Nikki, Danny and Monique, for their unspoken lessons and unwavering
support in everything I do. I would also like to thank my parents, Judith and Daniel
Sinnott, for teaching me the importance of an education and for instilling in me the desire
for knowledge and the courage to passionately pursue happiness. I never could have

done this without you.

iii

TABLE OF CONTENTS

LIST OF TABLES .................................................................................... iv
LIST OF FIGURES .................................................................................. vi
LIST OF EQUATIONS ............................................................................. xv
CHAPTER ONE

IN VITRO MECHANICAL EVALUATION OF DYNAMIC COMPRESSION PLATES
FOR PANCARPAL ARTHRODESIS

Abstract ....................................................................................... 1
Introduction ................................................................................... 3
Materials and Methods ..................................................................... 16
Results ........................................................................................ 32
Discussion ................................................................................... 38
Conclusion .................................................................................. 46
References ................................................................................... 47
CHAPTER TWO

MECHANICAL EVALUATION OF AN INTERLOCKING NAIL LOCKED WITH
EITHER BOLTS OR A TYPE-IA EXTERNAL SKELETAL FIXATOR IN A
SIMULATED CANINE TIBIAL MODEL

Abstract ...................................................................................... 49
Introduction ................................................................................. 5 1
Materials and Methods ..................................................................... 61
Results ....................................................................................... 78
Discussion .................................................................................... 87
Conclusion .................................................................................... 97

References98

CHAPTER THREE

COMPREHENISIVE MECHANICAL EVALUATION OF FRACTURE FIXATION
SYSTEMS: COMPARISON OF PLATED VS. SCREWED, BOLTED, AND NOVEL
INTERLOCKING NAIL CON STRUCTS

Abstract .................................................................................... 102
Introduction104
Materials and Methods ................................................................... 108
Results ...................................................................................... 120
Discussion ................................................................................. 129
Conclusion ................................................................................. 136
References ................................................................................. 1 37

 

APPENDIX ONE

FIXTURE SOPS
Chapter One
Plate Bending ..................................................................... 141
Strain Gage Selection and Application ....................................... 143
Chapters Two and Three
Bone Model Drilling ............................................................ 147
Bushing Adjustment ............................................................ 150
ILN Placement ................................................................... 151
ILN-ESF Placement ............................................................. 152
Foam Tube SOP .................................................................. 154
Foam Lathing SOP ............................................................... 155
APPENDIX TWO
TESTING SOPS AND PIDS
Instron Machine General Information ................................................. 164
Torsion Test SOP ......................................................................... 168
Bending Test SOP ........................................................................ 174
Compression Test SOP .................................................................. 180
Testing PIDS ............................................................................... 187
APPENDIX THREE
THEORETICAL ANALYSIS
Chapter One
Theoretical Formulasl90
Chapter Three
Matlab Programming ............................................................ 193
APPENDIX FOUR
ADDITIONAL DATA
Chapter One
Compliance Data ................................................................. 198
p-values ............................................................................ 198
Finite Element Analysis ......................................................... 200
Chapter Two
Overview of Data Provided .................................................... 203
Torsion Test Data and p-values ................................................. 204
Bending and Compression Test Data ......................................... 205
Chapter Three
Overview of Data Provided .................................................... 206
Torsion Test Data ................................................................ 207
p-values .................................................................. 209
Bending Test Data ............................................................... 210
p-values .................................................................. 216

 

LIST OF TABLES

Table 1-1. Maximum angular deformation data indicating significant differences..... ....35
Table 1-2. 10th cycle plate strain data indicating significant differences... ... ... ... .... ....37

Table 2-1. Comparative material properties of human and canine cortical

bone and of the composite material used in this study ............................. 62
Table 2-2. Torsion data indicating significant differences .................................... 83
Table 2-3. Bending data indicating significant differences .................................... 85
Table 2-4. Compression data indicating significant differences .............................. 86
Table 3-1. Nail and plate AMI values .......................................................... 121
Table 3-2. Torsion data ........................................................................... 125
Table 3-3. Bending data .......................................................................... 128
Table Al-l. Bushing selection for desired screw/bolt size .................................. 151
Table A1-2. Bolt/screw selection for desired nail size ....................................... 152
Table A4-l. Chapter One - Compliance data for HP and LC-DCP constructs ............ 198
Table A4-2. Chapter Two - Torsion test data .................................................. 204
Table A4-3. Chapter Two - Torsion data p-values ............................................ 204
Table A4-4. Chapter Two - Mediolateral bending, Craniocaudal
bending, and Compression test data .............................................. 205
Table A4-5. Chapter Three - Torsion test data .......................................... 207-208
Tables A4-6 thru A4-9. Chapter Three - Torsion data p-values ............................ 209
Table A4-10. Chapter Three - Mediolateral and craniocaudal
bending data (long and short specimens) ................................... 210-216
Table A4-11 thru A4-14. Mediolateral bending data p-values ............................. 217
Table A4-15 thru A4-18. Craniocaudal bending data p-values ............................ 218

vi

LIST OF FIGURES

Images in this thesis are presented in color
Figure 1-1. Canine skeleton, with upper extremity bones indicated ........................... 3

Figure 1-2. Lateromedial radiograph of the radiocarpal joint.
The figure on the left is of a normal joint;
the figure on the right shows a severe hyperextension ............................ 4

Figure 1-3. A) Superficial ligaments of the left carpus, palmar aspect.
B) Deep ligaments in the left carpus, palmar aspect ............................... 5

Figure 1-4. Dorsopalmar radiograph of the radiocarpal joint.
The figure on the left is of a normal joint
the figure on the right shows a distal radial fracture..................................6

Figure 1-5. Dorsopalmar and lateromedial radiographs
of a normal canine radiocarpal joint .................................................. 7

Figure 1-6. Lateral and medial view of the carpus and metacarpal bones,
indicating location of the accessory carpal bone ................................... 8

Figure 1-7. Dorsopalmar radiograph showing metacarpal fractures induced
by large metacarpal screw diameter in relation to bone size .................... 10

Figure 1-8. Comparison of the HP and LC-DCP, the HP is shown
on the left and the LC-DCP is on the right ........................................ 1 1

Figure 1-9. Lateromedial radiograph of an implanted LC-DCP bent with a
standard bending press. The standard press allows the bend to
span over an extended length. Reduction of bone-plate contact
area due to the shape of the standard bend was also visible .................... 14

Figure 1-10. Top view of the MSU BP ........................................................... 16

Figure 1-11. Dorsal and palmar aspects of both plates have convex
and concave surfaces, respectively. A HP is shown ........................... 17

Figure 1-12. View of the concave fulcrum with a curvature
matching the dorsal aspect of the plate ........................................... 18

Figure 1-13. Mount and pin, which locked the plate in position,

resisting force from the fulcrum and contact point,
which would result in horizontal displacement ................................. l8

vii

Figure 1-14. View of the locator, which stopped the lever arm
when a specific degree of rotation was completed.
The resultant bend in the plates could be modified
by adjusting the position of the locator .......................................... 19

Figure 1-15. Location of strain gages on plates, HP is shown ................................ 21

Figure 1-16. Dorsopalmar and mediolateral radiographs of
specimen implanted with fully instrumented HP ................................ 22

Figure 1-17. Dorsopalmar and mediolateral radiographs of
specimen implanted with fully instrumented LC-DCP ......................... 23

Figure 1-18. Specially design potting jig to fix the
constructs at 20° of hyperextension .............................................. 24

Figure 1-19. Insert at base of potting jig to prevent contact between

the epoxy and the plate near the metacarpals .................................... 25
Figure 1-20. 500 lb load cell, mounted on the base of the Instron table ..................... 25
Figure 1-21. Specimen mounted in servo-hydraulic testing machine ........................ 26

Figure 1-22. Construct compliance increased in nonlinearity as load
increased. Shown are 40% and 120% BW compliance

curves, corresponding to 100 N and 300 N, respectively.........................28
Figure 1-23. Construct angular deformation .................................................... 29
Figure 1-24. Diagram of single arm Wheatstone bridge circuit .............................. 30

Figure 1-25. Dorsal view of plates, showing consistent location
of bend for the LC-DCP group and HP group .................................. 32

Figure 1-26. View of plates, showing consistent degree of
bend for the LC-DCP group and HP group ...................................... 33

Figure 1-27. Lateromedial radiograph of an implanted LC-DCP
bent with MSU BP and a standard bending press.
Notice the concentrated bend with the MSU BP
occurring between the fourth and fifth screw hole,
while standard bending presses allow the bend to
span over a longer length. Reduction of bone-plate
contact area due to the shape of the standard bend
is also visible ........................................................................ 34

viii

Figure 1-28. Representative angular deformation curves of
HP and LC-DCP constructs at 120% BW ........................................ 35

Figure 1-31. Close-up dorsal view of HP and LC-DCP after
bending with MSU BP. Significant denting still
occurred in the central region of the plate ...................................... 44

Figure 2-1. Comminuted long-bone fracture, before and after
fixation with a medially applied bone-plate ...................................... 51

Figure 2-2. Mediolateral radiograph of a tibial fracture
internally fixated with an interlocking nail ....................................... 52

Figure 2-3. Diagram of mismatch between screw diameter and the ILN hole diameter..53

Figure 2-4. Bolt for ILN, featuring a threaded cis-cortex
section and a smooth, solid core ................................................... 54

Figure 2-5. Comminuted long—bone fracture, before and after being
externally fixated with a type-IA external fixator ............................... 55

Figure 2-6. Type-IA is a one-plane, unilateral fixator with
6 transosseous pins and one connecting bar.
Type-IB is a two-plane bilateral fixator with
8 transosseous pins and four connecting bars .................................... 56

Figure 2-7. Type-[IA and type-IIB ESFs. Notice the full pins in
the type-11A and the full and half pins in type-11B.
The use of half pins results in a slight loss of stiffness

compared to a full-pin configuration .............................................. 56
Figure 2-8. Type-III ESF, notice the bilateral and biplanar configuration .................. 57
Figure 2-9. Geometric parameters influencing fracture stabilization with an ESF ........ 58

Figure 2-10. Radiograph of a tibial specimen implanted with the
investigational hybrid bolt-pin locked with a type-IA
external skeletal fixator. A close-up photograph of the
hybrid 2.7mm ILN bolt/3.2mm ESF pin passing through
one of the nail holes... ................................................................................. 59

Figure 2-11. Schematic depiction of the synthetic bone model with
the various dimensions of each segment and a schematic
depiction of the outline of an actual canine tibia and fibula
superimposed over a completed bone model with a novel
ILN with 2 of the 4 locking SCPs ................................................. 62

ix

Figure 2-12. Partially disassembled drilling fixture, with the top portion
on the left and the bottom on the right. The cupping section
of the bone model is inserted into the bottom portion with the
metaphyseal and diaphyseal sections still visible. A spacer for
minor adjustment is also shown ................................................... 63

Figure 2-13. Top portion of drilling fixture, with symmetrical halves shown
on the left and the removable drill guides shown on the right ............... 64

Figure 2-14. Schematic of correct and incorrect way to drill pilot holes .................... 64

Figure 2-15. Removable drill guide, featuring a geometry matching the slope
of the metaphyseal section of the bone model, a radius matching
the diaphyseal section, and interchangeable bushings for various
bolt and screw sizes ................................................................ 65

Figure 2-16. Assembled bone drilling fixture, with bone model enclosed.
A hand held drill is shown drilling pilot holes .................................. 66

Figure 2-17. Custom designed lathe tool, which created a geometry
matching the inside of the metaphyseal section of the

bone model ........................................................................... 66
Figure 2-18. Foam plug insertion into the bone model ......................................... 67
Figure 2-19. ILNb alignment fixture ............................................................. 67
Figure 2-20. [LN-ESF alignment fixture ........................................................ 68
Figure 2-21. CDI Computorq digital torque wrench

shown tightening ESF clamps ..................................................... 69
Figure 2-22. Disassembled potting fixture, inner shell and outer shell ...................... 70
Figure 2-23. Assembled potting fixture with bone model inserted ........................... 70

Figure 2-24. Potting fixture, which featured perpendicularly oriented thru
holes for locking bone to potting fixture and threaded-holes
for locking potting fixture to testing fixture cups .............................. 71

Figure 2-25. Potting fixture with partially inserted locking pins installed.
The assembled construct is shown ................................................ 71

Figure 2-26. The torsion fixture shown with a loaded [LN-ESF ............................. 74

Figure 2-27. The bending fixture shown with a loaded ILNb ................................ 75
Figure 2-28. The compression fixture shown with a loaded ILNb ........................... 76
Figure 2-29. Construct angular deformation in bending ....................................... 77

Figure 2-30. Polyurethane foam plugs created on a lathe with a custom designed tool...78

Figure 2-31. Fully assembled ILNb and [LN-ESF constructs ................................ 79
Figure 2-32. Typical bimodal torsion compliance curve ...................................... 80
Figure 2-33. Representative torsion curve for ILN b, notice the presence of slack..........80

Figure 2-34. Representative torsion curve for [LN-ESF,
notice the lack of slack and change in compliance ............................. 81

Figure 2-35. Representative mediolateral bending curve for ILNb,
notice the bimodal shape and presence of slack ................................ 84

Figure 2-36. Representative mediolateral bending curve for [LN-ESF,
notice the unimodal shape due to an absence of slack ......................... 84

Figure 2-37. Representative compression curves for ILNb (top) and [LN-ESF ............ 86
Figure 2- 38. Schematic of the load transfer between the bone and ESF.

In vitro setup is shown on the left and beam theory

correlation is shown on the right ................................................. 90
Figure 2-39. Schematic of the nail without applied torque and after

torques is applied. The nail hole axis is green and the

bolt axis is pink, notice the amount of nail rotation prior

to engagement between the bolt and nail ......................................... 92

Figure 2-40. Schematic of the applied moment on the bone model
resulting in deflection of the transosseous pins ................................. 93

Figure 3-1. Modified [LN-ESF alignment fixture to accommodate the br-DCPs ........ 109
Figure 3-2. The plate locking mechanism that restricted movement in all directions....109

Figure 3-3. A) Photograph of ILNn. B) Close-up view of the distal
end of the novel ILN. C) Close-up view of the SCP .......................... 1 12

Figure 3-4. Cross-section of a nail with corresponding x and y-axes,
shown at nail hole level ............................................................ 1 13

xi

Figure 3-5. Schematic of various sections of SCP and [LN ................................. l 13
Figure 3-6. The torsion fixture shown with a loaded ILN n ................................. l 15

Figure 3-7. Stainless steel ball for connection between the

cups and the loading arm ........................................................... 1 16

Figure 3-8. Linkage arm shown attached to the bending cups .............................. 1 16

Figure 3-9. The modified loading arm .......................................................... l 17

Figure 3-10. Loading arm bearings .............................................................. 117
Figure 3-11. A) Partial ball joint for connection between main bar and linkage arms.

B) Linkage arm shown attached to main bar ................................... 1 18

Figure 3-12. The bending fixture shown with a loaded br-DCP ............................ l 18

Figure 3-13. Pilot holes drilled in the bone model with
respect to the plate hole locations ............................................... 122

Figure 3-14. All constructs from left to right:
ILN6s, ILN6b, ILNSS, ILNSB, br-DCP, and ILNn ........................... 122

Figure 3-15. Typical compliance curves; note the bimodal response which was
characteristic of the 6 and 8 mm Hst, and the unimodal response

that characterized the response of the br-DCP and ILNn constructs........ 123
Figure 3-16. Construct compliances in torsion ................................................ 124
Figure 3-17. Construct angular deformation in torsion ....................................... 125

Figure 3-18. Typical compliance curves; note the bimodal response which was
characteristic of the 6 and 8 mm ILNs, and the unimodal response

that characterized the response of the br-DCP and ILNn constructs ....... 126
Figure 3-19. Construct compliance in bending ................................................ 127
Figure 3-20. Construct angular deformation in bending ..................................... 128
Figure Al-l. MSU BP lever arm ............................................................... 141
Figure Al-2. MSU BP plate holder and adjustments ......................................... 142
Figure A1-3. MSU BP with the plate pushed forward ....................................... 142

xii

Figure Al-4. Close-up view of solder joint attaching

the copper wire to the strain gage .............................................. 146
Figure Al-5. Fully prepared implant ........................................................... 147
Figure A1-6. Disassembled drill guide ......................................................... I48

Figure Al-7. Make sure all edges pointed out by the white arrows are flush and fiat... 148

Figure A1-8. View of drill bit insertion into drill guide ..................................... 149
Figure Al-9. Disassembled top portion of drill guide ........................................ 150
Figure Al-10. View of inner housing containing drill guide bushings .................... 150
Figure Al-ll. View of drill guide inner housing profile .................................... 151
Figure A1-12. View of fully assembled construct ............................................. 152
Figure Al-l3. Enclosed pin cutter .............................................................. 152
Figure Al-l4. [LN-ESF in alignment fixture .................................................. 153
Figure Al-lS. Correct screw tightening order ................................................. 154
Figure Al-16. Depuy “hip drill 45° angle” plastic tube ..................................... 154
Figure A1-17. Depuy plastic tube with aluminum foil shell ................................. 155
Figure Al-18. Iron plunger ...................................................................... 155
Figure Al-19. Foam lathing materials .......................................................... 156
Figure A1-20. Lathe, including: A) Lathe head; B) Lathe bit crosshead;

C) Drill bit crosshead ............................................................ 157
Figure A1-21. Foam lathe bit, Version 1 and Version 2 ...................................... 158

Figure A1-22. Close-up view of lathe bit holder and tightening of lathe
bit holder using foam lathe bit chuck. Note that using
T-square while tightening will ensure proper alignment ................... 158

Figure A1-23. Close-up view of drill bit crosshead with drill bit and chuck shown ...... 159

Figure A1-24. Tighten the foam tube into the lathe head using the lathe head chuck....159

xiii

Figure Al-25. Finally, push the foam piece into the bone model ........................... 163

Figure A2-l. General view of Instron machine ............................................... 164
Figure A2-2. A2D, Encoder, and Instron computers ......................................... 166
Figure A2-3. 8500 Board ........................................................................ 166
Figure A2-4. Torsion fixture and required materials ......................................... 169
Figure A2-5. View of construct within torsion fixture ....................................... 169
Figure A2-6. View of torsion fixture within Instron machine .............................. 170
Figure A2-7. Bending fixture and required materials ........................................ 174
Figure A2-8. View of construct within bending fixture ...................................... 175
Figure A2-9. Compression fixture and required materials ................................... 181

Figure A2-10. View of construct within compression

fixture attached to Instron machine ............................................ 182
Figure A2-ll. Order of materials for bottom cup ............................................. 183
Figure A4-l. Cross-sectional view of the solid section of the HP and LC-DCP. ..........200

Figure A4-2. FEA of the HP. Equivalent stresses are shown on the top,
deformation is shown on the bottom ........................................... 201

Figure A4-3. FEA of the LC-DCP. Equivalent stresses are shown on the top,
deformation is shown on the bottom ............................................. 202

xiv

LIST OF EQUATIONS

Equation l-l. Angular deformation .............................................................. 30
Equation [-2. Shunt calibration strain readout ................................................. 31

Equation 2-1. Angular deformation, where positive and negative

are in reference to the applied moment ......................................... 77
Equation 2-2. Stiffness equation for External Skeletal Fixator construct .................. 92
Equation 3-1. Area moment of inertia .......................................................... l 12
Equation A3-1. Shunt calibration strain readout .............................................. 190

Equation A3-2. Longitudinal strain as a function of radius

of curvature and distance from neutral axis ................................ 190
Equation A3-3. Hooke’s Law ................................................................... 191
Equation A3-4. Flexural formula ............................................................... 191
Equation A3-5. Area moment of inertia (AMI) ............................................... 192
Equation A3-6. Inverse relationship between AMI and strain .............................. 192

XV

CHAPTER ONE

IN VI TRO MECHANICAL EVALUATION OF DYNAMIC
COMPRESSION PLATES FOR PAN CARPAL ARTHRODESIS

ABSTRACT

Pancarpal arthrodesis is indicated for severe injuries, degenerative conditions and
instability of the carpus. While various fixation methods have been described in large
dogs, the most common procedure uses a dorsally applied 3.5mm AO/ASIF limited
contact dynamic compression plate (LC-DCP). Due to various clinical issues associated
with the use of LC-DCPs, a new 3.5/2.7mm hybrid pancarpal arthrodesis hybrid plate
(HP), which tapers distally in both width and thickness, has recently been designed. The
purpose of this study was to compare the biomechanical properties of pancarpal
arthrodesis constructs using either an LC-DCP or an HP. The hypotheses of this study
were that pancarpal arthrodesis constructs stabilized with an HP would experience less
angular deformation than LC-DCP constructs under load. It was further hypothesized that
HPs would experience less strain than LC-DCPs at various load levels. Seven pairs of
canine forelimbs were obtained from dogs of similar body weight euthanized for reasons
unrelated to this study. Seven HPs and 7 LC-DCPs were consistently bent to 20°, using a
custom-designed bending press. All plates were instrumented with two strain gages
placed in the area of maximum bending on the medial and lateral side. Arthrodesed
constructs were embedded in epoxy molds and mounted in a servo-hydraulic testing
machine using custom-designed articulated fixtures that allowed rotation in the sagittal
plane. Specimens were successively loaded at 100, 200 and 300N for 10 cycles. Data,

consisting of construct angular deformation and plate strain, were compared using paired

t-tests. The results of this study were that the angular deformation; and plate strain of the
HP constructs was less than the LC-DCP constructs at all load levels. This study
demonstrated the mechanical advantages of the HPs over LC-DCPs under physiological
loading conditions. The smaller HP construct angular deformation and plate strain may
reduce the risk of implant failure and post-operative morbidity. The improved mechanical

properties of the HP make this implant a viable alternative to LC-DCPs.

INTRODUCTION
Arthrodesis (surgical fusion of a joint) is an orthopaedic procedure most often
performed in the carpus and tarsus."4 In canines, the carpus is similar to the cluster of
bones in the human hand between the radius and ulna and the metacarpals. The carpal
bones are not directly associated with the digits, whereas the metacarpal bones are. The

joint between the radius/ulna and the carpus is called the radiocarpal joint (Figure 1-1).

T
humerus A\ ‘t;\.
radius

  

Radiocarpal joint

 

 

Figure 1-1. Canine skeleton, with upper extremity bones indicated.5

Arthrodesis of the carpus is most commonly indicated after an injury caused by

hyperextension (Figure 1-2).2

 

Figure 1- 2. Lateromedial radiograph of the radiocarpal joint. The figure on the left IS of a
normal joint; the figure on the right shows a severe hyperextension. U SA
Falling or other impact traumas that hyperextend the carpus can cause severe
in'ur to the numerous almar li aments that su on the 3 levels of the ca al ‘oints
J Y P g PP rp J

(Figure 1-3).

 

' MSU SACS: Michigan State University Small Animal Clinic Sciences. College of Veterinary Medicine.
Michigan State University. East Lansing, MI

 

Radius

 

 

 

 

 
  
  
  
  

  
 
  

 

Short radial collateral
ligament

Superficial digital flexor

 

 

 

 

 

 

 

 

Deep digital flexor Ulna

 

Palmar radiocarpal
ligament

 

‘ Flexor retinaculum CA

 

 

 

 

 

 

 

 

 

Palmar ulnocarpal
ligament

. Palmar carpal fibrocartilage

 

 

 

 

Figure 1-3. A) Superficial ligaments of the left carpus, palmar aspect. B) Deep ligaments in the
left carpus, palmar aspect.6

These ligaments, along with the palmar fibrocartilage located on the palmar
aspect of the carpometacarpal joint space, are the major supports that permit ~15° of
extension at the antebrachiocarpal joint in the average standing animal.7 Although
immune-mediated arthritis can cause some of these hyperextensionsf‘7 one of the more
common causes is a degenerative condition of the palmar carpal ligaments seen in older,
large-breed dogsg‘9 Other indications for carpal arthrodesis include intractable distal
radial and carpal fractures or dislocations (Figure 1-4), where adequate joint stability
cannot be achieved and severe luxation/subluxation is present. Cases of severe carpal
arthritis and selected neurological defects, where medical therapy does not provide

adequate pain relief and limb function, have also been shown.2J0

    
 

carpal fracture

Figure l- 4. Dorsopalmar radiograph of the radiocarpal joint. The figure on the left IS of a

normal joint; the figure on the right shows a distal radial fracture. SAC
Evaluation of patients with hyperextension, fractures, or dislocations of the carpus
involves physical examination to determine the degree of joint instability, as well as

taking dorsopalmar and lateromedial radiographs (Figure 1-5).

 

Figure 1-5. Dorsopalmar (left) and lateromedial (riLght) radiographs of a normal canine
. . . MS SACS
radiocarpal jomt.

The carpometacarpal joint is most commonly involved due to the support
provided by the palmar carpal fibrocartilage found at this level (Figure 1-3). Instability of
the radiocarpal joint is present in only 10% of cases with hyperextension}7 The main
structure that provides significant radiocarpal support during weight bearing is the flexor

carpi ulnaris muscle, which inserts on the accessory carpal bone (Figure l-6).

 

 
   

 
   

accessory
carpal bone

 

 

 

 

Figure 1-6. Lateral (left) and medial (right) view of the carpus and metacarpal bones, indicating
location of the accessory carpal bone.6

Disruption of the ligaments attaching the accessory carpal bone to the carpal and
or metacarpal bones will result in proximal displacement of the accessory carpal bonef"7
Identification of the joint levels involved in the injury leads to the next question: partial
carpal arthrodesis or pancarpal arthrodesis. Since carpal hyperextension injuries do not
necessarily involve the radiocarpal joint, some argue that if the radiocarpal joint is
preserved only partial arthrodesis should be performed. Partial arthrodesis involves only
the intercarpal and carpometacarpal joints thus resulting in little change in gait.
Furthermore, the procedure involves limited surgical equipment and implants such as
cross-pin stabilization and cancellous bone grafting have been described.7 Pancarpal
arthrodesis, in contrast, produces a change in gait, where animals must learn to

circumduct the limb during the forward swing phase of the gait since the carpus can no

longer be flexed. Partial arthrodesis in the presence of subtle or mild radiocarpal ligament
pathology could lead to a subsequent breakdown and reoccurrence of hyperextension.
Regardless of partial or pancarpal arthrodesis, the procedure may increase stresses on
adjacent joints, potentially resulting in degenerative joint disease. With partial carpal
arthrodesis, there is a high risk of radiocarpal degenerative joint disease (DJD) due to
impingement between the implant and the distal aspect of the radius?“ For these
reasons, pancarpal arthrodesis is most often preferred, regardless of the level of carpal
joint injury.2

While various internal and external stabilization procedures have been described,
”"2 a dorsally applied limited contact dynamic compression plate (LC—DCP) is the
standard technique currently recommended for pancarpal arthrodesis.l Application of a
plate to the carpus for pancarpal arthrodesis presents several technical challenges. First,
basic fracture repair biomechanics dictate that in order to provide adequate support, the
bone plate should be placed on the tension side of the bone (i.e. the palmar aspect of the
carpus). This is due to the fact that upon physiological loading of the bone, compressive
forces will reduce the fracture gap. Muller et al. found that excessive fracture gap, with
no inherent fracture stability, may lead to a hypertrophic nonunion. However, the
approach to the palmar aspect of the canine carpus is difficult due to the important
tendinous and vascular structures located in this area (Figure [-3) Consequently the
compression surface (dorsal aspect) of the carpus is the preferred location.'2 Second,
plates that are sized appropriately for a dog‘s radius accept screws that are often too large

for the carpal and metacarpal bones. Until recently, the choice of bone plates has been

limited to the LC-DCP included in the AO/ASIF (Arbeitsgemeinschaft fur

Osteosynthesefragen / Association for the Study of [ntemal Fixation) system.9 While
these plates are suitable for the radius, they become too wide and thick for the metacarpal
region of most dogs, leading to complications. Mainly, the screw diameter used in a LC-
DCP that is appropriately sized for the radius typically encompasses more than 50% of
the diameter of the metacarpal bone. This results in an increased risk of metacarpal
fracture and subsequent implant failure because the large screw diameter, relative to the

diameter of the bone, acts as a stress riser (Figure l—7).l3

metacarpal fracture

 

Figure l-7. Dorsopalmar radiograph showing metacarpal fractures induced by large metacarpal
screw diameter in relation to bone size. N 5 ACS

Conversely, the use of a smaller plate with screws that are appropriate for the
metacarpal bones sacrifices plate strength and rigidity. Third, there is limited soft tissue
coverage over the metacarpal region, thus closure over the relatively thick LC—DCP is
usually done under excessive tension, leading to incisional dehiscence and secondary

wound complications. Considering these complications, 3 hybrid pancarpal arthrodesis

plate (HP) was developed by Veterinary Instrumentation in Sheffield, UK. (Figure [-8)

Some images in this thesis are presented in colour.

       
  
   
  

 

holes for 3.5 mm screws
for radius attachment

  

holes for 3.5 mm screws
for radius attachment

 

hole for either a 2.7 mm
or 3.5 mm screw for
carpal attachment

 
 

 

 

holes for 2.7 mm screws for
metacarpal attachment

   
   

holes for 3.5 mm screws
for metacarpal attachment

 

 

Figure 1-8. Comparison of the HP and LC-DCP, the HP is shown on the left and the LC-DCP is
on the right.

The HP features nine screw holes spanning its entire length. Hybrid plates that
would be used in a mid-sized dog (~30 kg) includes four 3.5 mm oval dynamic
compression (DC) screw holes in the proximal aspect of the plate, four 2.7 mm DC screw
holes in the distal aspect of the plate, and one central round screw hole, which can
accommodate either a 3.5 mm or 2.7 mm screw. The smaller screws in the distal aspect
may reduce the risk of metacarpal fracture. In addition, the HP is manufactured at a
standard length or with an extended length of 17 mm. A recent study demonstrated that
the risk of a bone fracture at the distal screw hole was reduced if at least 50% of the
length of the metacarpal bone was covered by the bone plate.'3 The added length of the

extended HP could thus further prevent metacarpal fracture. The HP also narrows at the

distal end in both width and thickness. This feature allows for easier closure and less
incisional tension.

Although the introduction of the HP has improved clinical application of the plate
during surgery, controversy remains regarding the angle at which the plate should be
placed in order to provide the strongest construct while preserving a functionally
appropriate carpal angle. Traditionally, canine carpal joints have been fused at a straight
(0°) angle. Although some veterinarians continue to use this technique because a straight
plate usually provides a stiffer construct, MSU SACS experience suggests that forcing
the carpus into a straight position not only results in poor contact of the digital and
metacarpal pads with the ground and subsequent gait abnormalities, it can also cause
tendonitis and bursitis and thus is not an optimal fixation. In order to circumvent this
limitation the HP has been designed with a distal taper in thickness that places the carpus
in approximately 4° to 5° of hyperextension without having to bend the plate. However,
the normal standing angle of the canine carpus is approximately 15° hyperextension.20
While the HP could be advantageous over a standard LC-DCP due to an extended fatigue
life of the plate given that it allows for some angling without having to actually bend the
plate,'5 it still fails to place the carpal joint at the desired angle of hyperextension.

Current veterinary surgical textbooks recommend the fusion of the carpal joint at
10° to 15° of hyperextension."I4 Anecdotally it has been the experience of MSU SACS
that further angulation of the plate up to 20°, allows for better contact between the ground
and the paw at paw strike. This appears to be particularly important at faster gaits such as
the trot.l4 The surgeon’s choice to bend the plates introduces multiple complications to

the integrity of the plate. As mentioned before, bending the plate immediately reduces the

strength of the plate for static reasons. More specifically, the introduction of a bend in the
plate creates a larger moment arm than if the plate were left straight, increasing the
stresses experienced by the plate at the location of the bend. Another problem of clinical
relevance is the actual bending of the plate, and the type of resultant bend. Because the
plate has varying cross sectional areas throughout its length, due to the presence of screw
holes, the ability of the plate to resist deformation is dramatically reduced in some areas
of the plate. It has been the experience of MSU SACS that bending the currently
available LC-DCP consistently produces a bend at the level of a screw hole (Figure [-9)
Basic strength of materials theory states that a region of reduced cross sectional area
(screw hole level) will experience greater stress under a given load and subsequently will
deform prior to a region which has a greater cross sectional area (non screw hole level).'7
In theory, LC-DCPs have equal cross sectional areas (CSA) throughout the length of the
plate due to their undersurface design which features cutouts between screw holes so that
the CSA of the plate at screw hole levels is equal to the CSA of the plate between screw
holes. Although this weak spot should theoretically not exist in LC-DCPs, it has been the
experience of MSU SACS that the bend still tended to occur near a screw hole. The
undersurface design of the HP does not feature any cutouts and therefore should
theoretically be much weaker at the screw hole level and subsequently consistently bend
at that location. The largest problem with the bend of the plate occurring at the cross
section with a hole once again is related to the bending stress in a member, which is
inversely proportional to the area moment of inertia (AMI) of the member. Current
bending presses also allow the bend to span a larger length of the plate, due to inherent

design flaws, specifically the fact that there is no way of locking the plate to the bending

press while bending. This makes bone-plate contact difficult in the area of the joint, due

to the bend being in the shape of an arc (Figure 1—9).

broad bend.
spanning 2-3
holes

 

weakest section of
plate (hole is present)

 

Figure 1-9. Lateromedial radiograph of an implanted LC-DCP bent with a standard
bending press. The standard press allows the bend to span over an extended length. Reduction of
bone-plate contact area due to the shape of the standard bend was also visible. MSU SMAC

So ideally, if a bend is to occur, extreme care must be taken to ensure that the
bend is concentrated in an area of the plate where no hole is present (solid section). Due
to the increased force and precision necessary for this to occur, a bending press was
specifically designed (MSU BP), for the purpose of this study. This press, if used in the
operating room, could also enable a surgeon to produce a concentrated bend at a
determined location on the plate, regardless of the cross section of that location.

With the occurrence of a bend in the plate as large as 20°, the effect of the bend
on the integrity of the plate should be considered. Since a known amount of deformation

is to be given to the plate (20° curvature), the cross section of the plate where the bend is

to occur must be evaluated. The HP and LC-DCP have different cross sections, therefore
they must be considered differently. Under a given curvature, the distance from the
neutral axis to the surface of an implant governs the strain experienced in a given cross
section.'7 The thickness of the cross sections of the two plates (1 mm difference) do not
affect the strain experienced by the plates as much as their geometries do. The cross
section of the LC-DCP is crown shaped in the solid section of the plate, whereas the HP
has a cross section similar to a thin rectangle in the solid section of the plate. To
investigate the significance of their respective cross sections, a finite element analysis
was done on each plate. In this analysis, the cross sectional areas of the plates, where the
bend occurs, were subjected to a point load normal to their dorsal surface. Based on the
results of this theoretical analysis (Appendix 4), it was determined that the HP would
deform less and be subjected to less stress than the LC-DCP under a given load. Li. et al.
showed good clinical outcomes with radiocarpal extension from 5-10°, reporting that
arthrodesis occurred in all thirteen cases. Although a 20° bend may have the potential to
weaken the bone plate and decrease its fatigue life, to date no studies comparing the
biomechanical properties of a LC-DCP to a HP placed at a functionally appropriate 20°
angle have been reported. Therefore, the main purpose of this study was to compare the
biomechanical prOperties of a nine-hole LC-DCP and an extended length 3.5/2.7 mm HP
placed at a 20° bend for canine pancarpal arthrodesis. It was also desired to develop a
clinically usable bending press, which could produce a bend in plates that was both
consistent and functional. The hypotheses were that the HP constructs would undergo
less radiocarpal joint angular deformation and that the HP would experience less plate

strain compared to the LC-DCP, due to differences in the cross sections of the plates.

MATERIALS AND METHODS
Implant Preparation
Plate Bending — Prior to specimen collection, seven LC-DCPs and HPs were accurately
bent to 20°. Due to the difficulty and inconsistency of using current plate bending tools
and techniques, a specially designed bending press was developed (MSU BP [Figure l-

10]).

 

Figure 1-10. Top view of the MSU BP

The bending press uses a design commonly seen in pipe bending, where a lever
arm is rotated about a fulcrum and movement of the lever arm results in the bending of
the pipe. The pipe is essentially being forced into the shape of the fulcrum by the
distribution of force by a contact point. This concept was translated into a design, which
would enable a similar deformation or bending of a veterinary plate. In our bending press,

the plate was placed between two pins, one pin acted as a fulcrum and the other pin acted

as the contact point. Both pins were fixed on a lever arm, which enabled the generation of
large enough forces to bend the steel plates. Without adequate length of the lever arm, the
design of the bending press would be impractical due to the difficulty of generating
enough force to cause the resultant bend. The lever arm and both pins were then fixed to
a location on the base of the bending press. Both plates are convex dorsally and concave

palmarly (Figure 1-11).

 
  

—\, »

-\

Figure 1-11. Dorsal and palmar aspects of both plates have convex and concave surfaces,
respectively. A HP is shown.

When placed into the bending press, the dorsal and palmar aspects of the plate, or
convex and concave respectively, were in contact with both the fulcrum and contact pins,
respectively. To enhance the quality of the bend, the fulcrum and contact points were
fashioned into a concave and convex radius, respectively, which matched the concave

and convex aspects of the plate (Figure 1-12).

 

Figure 1-12. View of the concave fulcrum (shown in white) with a curvature matching the dorsal
aspect of the plate.

The importance of the distribution of forces over the entire width of the plate will
be further explained in the strain gage application section. The plate was fixed to prevent
displacement horizontally by a pin located proximally to the bend of the plate (Figure 1-
[3).

horizontal force ‘ ‘ , .j .
7 horizontal force

(5‘

/

 

hi . . i . '3 mount
lockingpin ‘

Figure 1-13. Mount and pin, which locked the plate in position, resisting force from the fulcrum
and contact point, which would result in horizontal displacement.

The mount that the pin attached the plate to was adjustable in the same horizontal
plane on the base of the bending press, this allowed for changes in the location of the
bend if desired. When the mount and pin were in place and secured, the bending press

was ready for use. By trial and error, the amount of lever arm rotation needed to achieve

the desired bend of 20° was located. The bending press also included a locator, which
would enable a consistent amount of lever arm rotation; this was utilized once a

satisfactory bend was achieved (Figure 1-14).

locator

 

Figure 1-14. View of the locator, which stopped the lever arm when a specific degree of rotation
was completed. The resultant bend in the plates could be modified by adjusting the position of
the locator.

There were differences in the positioning of the mount and locator between the
HP and LC-DCP constructs in order to achieve the desired bend angle of 20° Trial and
error provided to be the most accurate means of finding these points. The location
dissimilarities were due to geometrical differences of the thickness and width between the
HP and LC-DCP, resulting in different amounts of force required to achieve the desired
bend. The plate holes were numbered 1 through 9 from proximal to distal, all plates were

bent between the 4th and 5'h screw hole. See Appendix 1 for MSU BP instructions.

Strain Gage Application - The importance of the distribution of force over the entire

width of the plate, instead of a point load, was to preserve the integrity of the surface of

the plate. Strain gage application requires very specific surface conditions for proper
adhesion and accurate measurements. The presence of a point load (if standard bending
irons are used rather than the MSU BP) would result in large enough compressive forces
to cause plastic deformation of the plate in the form of a dent on the dorsal surface of the
plate, where the gage was applied. This dent or divot would cause a stress concentration
in addition to the already present location of high stress, relative to other areas of the
plate, which we were intending to measure. The already present stress risers were: a hole
located adjacent to the gage, and the gage was placed directly on top of the bend in the
plate. Due to the size of the dent, strain measurement in that area would not have been
quantifiable with our strain gages. By distributing the load over the entire width of the
plate, the MSU BP, prevented this from occurring to an extent, and allowed for an evenly
distributed bend matching the radius of the fulcrum

Strain gages were applied to medial and lateral sides of each plate between the
fourth and fifth screw hole, adjacent to the bend (Figure 1-15). Strain gage selection was
based on the required size of the gage and the estimated levels of strain, which would be
incurred during testing. Strain values were not expected to exceed 4000 as. The selected
strain gage (EA-06-031DE-350 Vishay Micro-Measurements) was rated for 3% strain,
and had a gage length of 0.79 mm and a grid width of 0.81 mm. Gages were applied by
the same person (Keith Curcio, DVM, MSU SACS); see Appendix 1 for full strain gage

application instructions.

20

medial gag ‘ \ /.f
\ V I

 

Figure 1-15. Location of strain gages (outlined in white) on plates, HP is shown.

Specimen Preparation

Seven pairs of anatomically normal canine thoracic limbs were harvested from
dogs, which weighed between 30 - 32 kg and had been euthanized for reasons unrelated
to this study. Additionally, dogs were chosen based on the length (80 i 0.15 mm) and
diameter (8 t 0.15 mm) of the third metacarpal bone, which was determined
radiographically. The limbs were disarticulated at the elbow and metacarpal/phalangeal
joints. All soft tissue structures were then removed from the limb with the exception of
the supporting structures around the carpal joint. The articular surfaces were left intact in
order to maintain consistency. The limbs were then stored in a — 20° C freezer until the
time for plate application and biomechanical testing. All specimen preparation was

carried out by the same person (Keith Curcio, DVM, MSU SACS).

21

Within one pair of limbs, a HP bent to 20° was applied to the dorsal side of the
limb. A 3.5 mm screw was first placed into screw hole number 5 to assure accurate
placement into the radiocarpal bone. Next a 2.7 mm screw was placed in screw hole
number 8 into the third metacarpal bone to make sure the plate lined up on the midline of
the bone. Then a 3.5 mm screw was placed in screw hole number 2 into the radius. The
remaining screws were then filled starting proximally with screw hole number 1 down
distally to screw number 9. The end result was four 3.5 mm screws in the radius, four 2.7
mm screws in the third metacarpal bone, and one 3.5 mm screw in the radiocarpal bone
Figure 1-16). The LC-DCPs bent to 20° were placed on the opposite limb in a similar

fashion. However, in the LC-DCP group all nine screws were 3.5 mm (Figure 1-17).

 

Figure 1-16. Dorsopalmar (left) and mediolateral (right) radiographs of specimen
implanted with fully instrumented HP. 1‘ 5” SACS

22

 

Figure 1-17. Dorsopalmar (left) and mediolateral (right) radiographs of specimen
implanted with fully instrumented LC-DCP. MSU SAC
The proximal portion of the radius and ulna were then transected 4.5 cm above
the proximal aspect of the bone plate to standardize the length of the specimen. The

interosseous ligament remained intact at this level.

Specimen Potting

The extremities of all of the specimens were embedded in an epoxy mold (Fibre
Strand, Martin Senour Corp., Cleveland OH.) conforming to a custom designed
articulated loading cup. Potting consistency and accuracy were enabled with a specially

designed pottingjig (Figure 1-18).

23

 

Figure 1-18. Specially design pottingjig to fix the constructs at 20° of hyperextension.

Caution was taken in positioning the specimens during potting to ensure that the
compressive loading of the specimens during testing caused pure bending. The presence
of any off axis loading of the plates, if the plate itself was not oriented perfectly vertical,
would apply a twisting moment on the plate. The presence of any type of twisting or
torsional force would drastically effect both angular deformation and strain gage data.
The epoxy mold extended just below the distal aspect of the plate without contacting it,
and was perpendicular to the metacarpal bones. This was also achieved by a specially
designed piece of the potting jig, which encompassed the metacarpals and protected the

plate from any contact with the epoxy material (Figure l-l9).

24

 

Figure 1-19. Insert at base of potting jig to prevent contact between the epoxy and the plate near
the metacarpals.
Mechanical Testing and Data Acquisition
Specimens were mounted on a servo-hydraulic testing machine fitted with a 500

lb (2,225 N) load cell (Figure 1-20) (Interface, Scottsdale, AZ).

 

Figure 1-20. 500 lb load cell, mounted on the base of the Instron table.

Specimens were secured to the testing machine via articulated custom designed cups,
which allowed free rotation of the specimen in the sagittal plane (Figure 1-21). This
allowed physiologic flexion of the specimen during loading, and minimized non-
physiologic stress concentrations at either bone-epoxy interface. The rotation axis of each
mount was instrumented with rotary encoders (Renco Encoders Inc., Goleta, CA) to

enable measurement of the specimen angular deformation during testing (Figure 1-21).

25

    
 

articulating rotary
cups encoders

Figure 1-21. Specimen mounted in servo-hydraulic testing machine.

Biomechanical testing was conducted sequentially in each pair. Prior to testing,
the load cell was zeroed and a 5N preload was applied to each construct. Specimens were
then successively loaded at 100 N, 200 N, and 300 N or 40%, 80%, and 120%,
respectively, of body weight (BW) at a rate of 1 Hz for 10 cycles. These load levels were

chosen to simulate post-operative loading conditions at rest, walk and trot respectively.'4

26

Since all experiments were conducted under load-control, construct compliance (rather
than stiffness) was determined. Construct compliance is the inverse of stiffness and is
defined as the slope of the deformation versus load curve. Construct compliance was
determined at the 10th cycle. However, it was determined post-hoe that compliance was
the most definitive measure of construct performance. Construct compliance is a measure
of the ability of the plate alone to resist deformation whereas angular deformation is a
measure of the ability of an entire construct (including bone deformation) to resist
deformation. As shown in Figure 1-22, nonlinearities in the compliance curves increased
with load. For this reason, angular deformation was evaluated instead of construct

compliance (compliance data included in Appendix 4).

. b 'd
Compliance at 40% BW . [330p
0.7 — Linear (LC—DCP)
— Linear (Hybrid)

 

0.6

   
 
 

LC-DCP R2 20.9965

 

0.5

0.4

HP R2 = 0.9946

Displacement (mm)
0
Q)

 

 

 

 

    

  

 

 

o
o 25 so 75 100
Load (N)
. ' H rid
Compliance at 120% BW . LgDCP
0.5 —Linear (LC-DCP)
—Linear (Hybrid)
0.4 i
A LC-DCP R2 = 0.9851
E 0.3 3 '
E ' HP R2=O.9834
9 0.2
E
i
a 0.1 L
.‘L’ 1
r: 1'
o
0 so 100 150 200 250 300 350
Load (N)

Figure 1-22. Construct compliance increased in nonlinearity as load increased. Shown are 40%
(top) and 120% BW (bottom) compliance curves, corresponding to 100 N and 300 N,
respectively. Notice the reduction in R1 value from 40% to 120% BW.

As specimens were loaded, the radial and carpal angles (0t and B, respectively)
with respect to the vertical were recorded via the rotary encoders affixed to each
articulated mount. Angular deformation (Ay) was computed with Equation 1 (Figure l-

23).

28

to 111‘ Y

 

Bo

 

Figure 1-23. Construct angular deformation.

A7 = AOC + AB [1.11

Equation [-1. Angular deformation, where:

Av. construct angular deformation
Au: radial angle recorded by a rotary encoder
AB: carpal angle recorded by a rotary encoder

Strain gage outputs (V) were converted to microstrain (us) via the electrical

bridge circuit of a calibrated strain gage indicator (Figure 1-24).

29

 

30060 ()HM
shunt cal resistor

 

 

 

 
 
  
     
  
 

 

3000 OHM
resistor

 

 

 

_, "
I 4
o/r,‘

y”
. av
/

 

350 OHM
RSG

 

 

 

 

O A—
V In __

 

 

 

 

 

 

3000 OHM
t resistor

350 OHM
resistor

 

 

 

 

 

 

 

 

 

 

0C

 

 

 

V out

 

Figure 1-24. Diagram of single arm ( 1/4 bridge) Wheatstone bridge circuit.

Circuit calibration was used to calibrate the system; this was done by attaching a shunt
calibration resistor (green in Figure 1-24) in parallel with the strain gage (red in Figure l-
24). The purpose of the shunt cal. resistor was to simulate an electrical resistance on the
circuit, so the voltage readout can be adjusted to give accurate readings in relation to
strain. The following equation is used to determine the strain reading when the shunt cal.

resistor was activated.
as = (Rg) / [Fg’*‘(Rg + RS)] [1.2]
Equation 1-2. Shunt calibration strain readout, where:

85: strain readout with shunt cal.
Rg: strain gage resistance (350 Ohms)
Fg: gage factor (2.12 i 1.0%)

Rs shunt cal. resistance (30060 Ohms)

30

Individual strains (medial and lateral, see Figure 1-15) were recorded on separate-
channels. Plate strain between medial and lateral gages was not significant; the two-gage

setup was primarily to ensure a reading if one of the gages were to fail or not trigger

properly.

Data Analysis

Construct angular deformation was recorded via the rotary encoders, linear displacement
of the actuator was recorded via the Instron LVDT, and load was measured using the 500
lb load cell. Statistical analysis was performed by use a 2-factor repeated measures
ANOVA (Sigma Stat). The two factors for all data analysis were the type of arthrodesis
(LC-DCP or HP) and load level (100 N, 200 N, or 300 N). All pairwise comparisons
were calculated with Student-Newman-Keuls post-hoc tests where significance was set at

p<0.05.

31

RESULTS

Bending Press Performance

As indicated in the materials and methods section, movement of the lever arm,
which was restricted by the locator, enforced deformation of the plate. Adjustment of the
locator point between the two groups was necessary to acquire a 20° bend in the plates.
The locator point dissimilarities between the HP and LC-DCP were due to geometrical
differences of the thickness and width between the HP and LC-DCP, resulting in different
amounts of force required-to achieve the desired plastic deformation. The plate holes
were numbered 1 through 9 from proximal to distal, all plates were bent between the 4lh
and 5‘h screw hole (Figure 1-25). Once the locator points were established for both
plates, a consistent bend was easily and efficiently achieved (Figure 1-26). The efficiency
of the bending process was due to the absence of multiple bending reiterations on each

plate to achieve similar bends between plates within each group.

 

Figure 1-25. Dorsal View of plates, showing consistent location of bend for the LC-DCP
group (left) and HP group (right).

s
.‘
v I
A

I_\ 'f. .—>-. 1‘,

; A‘ t . . ‘
57“": A”; 3" 3~\7’9\"v<"r‘;

V

 

Figure 1-26. View of plates, showing consistent degree of bend for the LC-DCP group (top) and
HP group (bottom).

As important as achieving a consistent location and degree of bend, was to
achieve a consistent type of bend. More specifically, the goal was to produce a
concentrated bend, which occurred in a solid section of a plate, without spanning over a
length, which included sections where holes were present. This was consistently a
problem with standard bending techniques, due to their inability to restrict movement of
the plate while bending occurred. This problem was considered in the design of the new
MSU BP, and the efficacy of this design in providing a concentrated, functionally located

bend is shown in Figure 1-27.

acute bend. occurring ' - . ‘ broad bend.
between the Iourth

_. spanning 2-3
and 111111 screw holes

screw holes

 

 

Standard BP

 

Figure 1-27. Lateromedial radiograph of an implanted LC-DCP bent with MSU BP (left)
and a standard bending press (right). Notice the concentrated bend with the MSU BP occurring
between the fourth and fifth screw hole, while standard bending presses allow the bend to span

over a longer length. Reduction of bone-plate contact area due to the shape of the standard bend is
also visible. MSU SACS

Angular Deformation During Compressive Joint Loading

Although maximum angular deformation of the HP constructs was consistently
less than that of the LC—DCP constructs at all loads, these differences were only
significant at 80% and 120% BW corresponding to 200 N and 300 N, respectively
(Figure 1-28, Table l-l). However, the differences between the groups were similar at
all load levels, the HP consistently deformed 18-20% less than the LC-DCP.
Additionally, both of the constructs angular deformation increased with increasing load
levels within their respective groups. From 40% to 80% BW (100 N and 200 N), the HP
exhibited a 268% increase in deformation, similarly, the LC-DCP showed a 267%

increase. Results were comparable in the 80% to 120% load level increase (200 N and

300 N), where increases of 188% and 182% were found for the HP and LC-DCP

34

constructs, respectively. The increases shown within groups were significantly different

between all load levels for both groups (p<0.05).

Angular Deformation at 120% BW

 

5.0

4.0 -.—-w ~~ --—- .-.._ ~- ~ , A—-— ~-

 

 

 

30 L. . . . .. .. ._ C .-. , -... . , . .. L..- _- 2..

 

 

 

2.0 , ~ ~~»e-- -- ._ - ~ ELL _, ,

Deformation (degrees)

 

 

 

0.0

 

4 Time (sec) 6

Figure 1-28. Representative angular deformation curves of HP and LC-DCP constructs at 120%
BW (300 N).

 

Maximum Angular Deformation (°)

40% BW 80% 13w“ 120% Bw”
HP 0.73 i- 0.47 i 1.96 :- 1.33 i 3.68 a 2.48 i

LC-DCP 0.94 a 0.66 * 2.51 J.- 1.64 * 4.56 a 2.82 *

 

 

 

 

 

 

 

 

 

Table 1-1. Maximum angular deformation data (mean : SD) (where 40%, 80% and 120% BW
correspond to 100 N, 200 N and 300 N respectively) indicating significant differences (p < 0.05,
see Appendix 4 for exact values):

# between HP and LC-DCP constructs

i between load levels for HP constructs

K between load levels for LC-DCP constructs
Plate Strain

All plate strain data was taken as the average of the two gages on each plate, since
significance was not seen between the medial and lateral strain readings. Similar to
angular deformation, plate strain in the HP constructs was consistently less than that of

the LC-DCP constructs. The HP experienced approximately 35% less plate strain than the

35

LC-DCP at all load level, this difference, however, was only significant at 80% and
120% BW, or 200 N and 300 N, respectively (Figure 1-29, Table 1-2). Similar to
construct angular deformation, plate strain also increased equally with increasing load
levels within their respective groups. From 40% to 80% BW (100 N and 200 N), the HP
exhibited a 240% increase in deformation, similarly, the LC—DCP showed a 239%
increase. Likewise, in the 80% to 120% load level increase (200 N and 300 N), increases
of 170% and 177% were found for the HP and LC-DCP constructs, respectively. In
addition, the peak plate strain of each treatment group increased significantly as the load

level increased for all comparisons (p<0.05).

Plate Strain at 120% SW — Hybrid
2500 . ._ 7,- — LC-DCP
. L_______

2000 1

 

1500 . ~-~

Strain (microstrain)

       

 

 

4 Time (sec) 6

Figure 1-29. Representative plate strain curves of HP and LC-DCP constructs at 120% BW (300
N). Shown is the average of the medial and lateral strain readings for each plate.

36

 

Plate Strain (us)
40% BW 80% 13wt 120% 13w“
HP 269 a 100 i 646 i 259 i 1098 i438 “
LC-DCP 410 .4: 203 * 980 «.- 478 * 1735 a 808 *

 

 

 

 

 

 

 

 

 

Table 1-2. 10th cycle plate strain data (mean : SD)(where 40%, 80% and 120% BW correspond
to 100 N, 200 N and 300 N respectively) indicating significant differences (p<0.05, see Appendix
4 for exact p-values):

# between HP and LC-DCP constructs
* between load levels for LC-DCP constructs

f between load levels for HP constructs

37

DISCUSSION

The hypotheses of the current study were that the HP constructs would undergo
less radiocarpal joint angular deformation than the LC-DCP, and also that the HP would
experience less plate strain compared to the LC-DCP. The results of this study supported
both of the hypotheses. Firstly, the HP bent to 20°° exhibited significantly less angular
deformation than the LC-DCP, at 80% BW (1.9 : l.3° versus 2.5 : 16°, for the HP and
LC-DCP respectively, at 200 N) and at 120% BW (3.7 i 2.5° versus 4.6 1 28°, for the
HP and LC-DCP, respectively, at 300 N). Secondly, that the HP would experience more

plate strain than the LC-DCP. This occurred at all load levels, but their differences were

significant only at the 80% BW (646 t 259 118 versus 980 i 478 as, for the HP and LC-

DCP, respectively, at 200 N) and at 120% BW (1098 i 438 118 versus 1735 i 808 118, for

the HP and LC-DCP, respectively, at 300 N). The design of the cross sectional area of HP
was the major reason there was an improvement in the biomechanical variables
documented in this study in comparison to the LC-DCP.

Based on recent in vitro studies, the decrease in construct angular deformation
observed in the specimens implanted with a HP has several potential advantages. It
decreases the risk of catastrophic plate failure by increasing the ability of the plate to
sustain higher loads. Similarly, the magnitude of stress from each bending cycle during
daily activity is also decreased, which in turn, may translate into an increase in plate
fatigue life.'8 This becomes important because in an angulated construct, there is an
increase of cyclic loading at the point of maximum bending. With repetitive loading,
construct angulation may promote fatigue failure of the plate. Although the immediate

effect of the decreased HP construct deformation appears minor, on the basis of results

38

from previous studies on tarsal arthrodesis constructs,15 it can be speculated that the
reduced deformation seen in the HP may have a substantial effect over the life of the
implant. Hulse et al. demonstrated on 3 16L cold—worked stainless steel that a reduction in
strain by two-fold could effectively extend the fatigue life of an implant by ten-fold.
Regardless of how many cycles of loading and unloading were experienced, the stress
levels of the HP and LC-DCP in this study were found to be far below the stress levels
associated with the fatigue failure seen in the Hulse study.19 However, in larger breed
dogs, where plates would experience increased levels of stress, the effect of the increased
stress on the fatigue life of the implant requires further consideration.

Several factors, primarily related to implant design, degree of construct
angulation, and load magnitudes placed on the implant may predispose pancarpal
arthrodesis constructs to failure. The latter two factors can be altered little, with the last
being largely beyond the direct control of the surgeon. However, addressing the surgical
challenge by improving the biomechanics of the implant construct has the potential to
minimize complications and maximize success. The implant design of the HP was the
major reason there is an improvement in the biomechanical variables documented in this
study in comparison to the LC-DCP. Despite a 20° bend of the plates, the HP was still
shown to be biomechanically superior. The hypotheses were governed by results of a
finite element analysis (FEA) that was conducted on the solid cross sections of the plates.
Results of the FEA were in good agreement with results calculated using equations given
for deformations in members under pure bending (See FEA data and further discussion in

Appendix 4). Specifically, an equation for strain related to radius of curvature, which

39

states that longitudinal normal strain varies linearly with the distance from the neutral
axis (see Equation B in Appendix 3).

We chose to conduct testing under load control since it better reflected the
biomechanical stresses to which the constructs are subjected during physiological
activities. The results of the current study indicated that in both constructs, angular
deformation consistently increased with load, with significance seen in the LC-DCP
construct at each load level increase. This finding likely resulted from the fact that as the
plate angular deformation increased, the moment arm was increased. The moment arm
was represented by the distance between the bend in the plate and the point of load
application. In this study, the HP experienced less angular deformation at all loads tested,
with significance seen at 80% BW (200 N) and 120% BW (300 N), simulating walk and
trot, respectively. The absence of a significant difference in the 40% BW (100 N) was
most likely attributable to large deviations seen at those load levels due to the reduced
load bearing needs of the plate. Though the LC-DCP showed greater angular deformation
at all load levels, the deviation within groups decreased as load increased. The design of
the central portion of the HP may have contributed to the decrease in angular
deformation. The weakest portion of a bone plate was where a screw hole is present,
while the strongest portion was the solid portion between the screw holes. This was due
to the second moment of area, also known as the area moment of inertia (AMI), of the
plate. The bending stress developed in the plate during loading was inversely
proportional to the AMI, this relationship was given by the flexural formula (see
Equation D in Appendix 3). The AMI is a geometrical property (see Equation E in

Appendix 3), which is related to the amount of material in a given region, therefore, the

40

AMI was higher in the solid portion of the plate than it was in the region where a hole
was present.

In the LC-DCP for example, the AMI in the solid region with respect to the axis
of bending was 28.13 mm4, in comparison to the region where a hole was present, where

the AMI was 10.26 mm4 (Figure 1-30).

     

Figure 1-30. Cross-section of the LC-DCP. The figure on the left is the solid section of the plate
where the bend occurs; the figure on the right is the section where a hole was present.

The bend in the plates for pancarpal arthrodesis occured between the fourth and
fifth screw holes to correspond with the radio-carpal joint. The HP had approximately 8
mm of metal between the fourth and fifth screw holes in comparison to the LC-DCP,
which had half (4 mm) of that length. More importantly though, was the AMI at the
location of the bend, which for the LC-DCP was 28.13 mm4 (provided by Synthes) and
for the HP was approximately 35.44 mm4. The large difference in the AMI of the plates
was attributable to the variation in the cross-sectional areas (CSA), the HP had a CSA in
the area of the bend of 32.64 mmz, and the LC-DCP had a CSA of 26.65 mmz, a
difference of 22.5% with respect to the LC-DCP. This further exemplified the importance
of a functional bending press, which forced the bend to occur on the solid section of the

plate. If standard bending techniques were used, and a bend was placed into the section

41

where a hole occurred, the difference between the plates and the inherent weakness of the
LC-DCP would have been even further magnified.

It should be noted that the deformations documented in the FEA did not correlate
very well with experimental data. However, the FEA was not meant to simulate the
experiment, it was only meant to validate theoretical calculations based on the different
cross sections of the two plates. Since theoretical analysis was the basis of the
hypotheses, the FEA was used as a way to non-destructively examine the mechanical
response of the plates when subjected to bending. Angular deformation was documented
as a way to analyze the entire construct, and not just the plate. The entire construct
consisted of the plate itself, the bones and tissues making up the radiocarpal joint, and the
screws, which attached them to each other. The rotary encoder data showed that the HP
construct deformed approximately 23% less than the LC-DCP construct at all load levels.
The FEA on the other hand, resulted in the HP deforming approximately 44% less than
the LC-DCP under a given load (100 N). The large difference between the experimental
and FEA results (see Appendix 4) could have been attributable to many factors, primarily
involving the connection between the plate, screws, and the underlying bone. One factor,
which could have contributed, was the presence of slippage between the plate and the
bone in both constructs. This would undermine the strengthening effects of the HP
design, and make the difference between the two constructs less defined. Such slippage
could have arose under several conditions; if the screws were not tight enough, if the
palmar aspect of the plate did not match the shape of the underlying bone as well as
anticipated, or if the overall bone to plate contact area was reduced in one construct

versus the other.

42

The plate strain was used to compare only the plate, as opposed the entire bone-
plate construct like the angular deformation data provided. The HP experienced
approximately 65% of the plate strain seen in the LC-DCPs at each load tested with
significance seen at 80% BW and 120% BW, or 200 N and 300 N, respectively (p<0.05).
Based on the linear relationship between stress and strain in the elastic region governed
by the constitutive equation known as Hooke’s Law (see Equation C in Appendix 3), in
combination with the flexural formula, the correlation between the AMI and the plate
strain experienced was also inversely proportional (see Equation F in Appendix 3).

Theoretical results provided by the FEA were in good general agreement with the
experimental results. As stated earlier, the AMI of the HP was approximately 26%
greater than that of the LC-DCP (at solid section level), and the recorded plate strains of
the LC-DCP were approximately 36.7% greater than that of the HP. Although they are
not directly related (axial strain and von Mises stress) the results of the FEA (see
Appendix 4) did correlate very well with the recorded strain measurements on the dorsal
aspect of the plate. In the FEA, the surface von Mises stresses seen in the HP were
approximately 38.3% less than that of the LC-DCP under a given load of 100 N.

The FEA results compared to the experimental results better when analyzing plate
strain and stress than they did when investigating angular deformation. This was due to
the fact that both the experimental data on plate strain and the FEA results on von Mises
stresses did not take construct variables into account, such as slippage and bone to plate
contact area. Variations between the theoretical and experimental results may have been
due to the plastic deformation that occurred to the plates during pre-bending, which

would have altered their cross sections and their resultant AMIs. Further investigation of

43

the correlation between surface strains measured experimentally and theoretically is
suggested.

One of the goals of this study was to develop a clinically usable bending press,
which could produce a bend in plates that was both consistent and functional. The
consistent location and degree of bend was achieved as shown in Figures 1-25 and 1-26.
The increased bone-to-plate contact area enabled with the type of bend produced with the
MSU BP is shown in Figure 1-27. One design flaw of the MSU BP was that, despite
precautions taken to distribute the force of the bend over the entire width of the plate, a

dent on the middle of the dorsal surface was still present (Figure 1-31).

 

 

Figure 1-31. Close-up dorsal view of HP (left) and LC—DCP (right) after bending with MSU BP.
Significant denting still occurred in the central region of the plate (circled in white).

An improvement to the design of the bending press would be to increase the
diameter of the fulcrum to more accurately match the dimension of the dorsal and palmar
aspects of the plates. The current diameter focused the force over too small of an area and
caused the denting. With knowledge of the exact measurements of the plate surfaces,
matching the radii of the fulcrum and contact point to the dorsal and palmar aspect of the

plates, respectively, would also reduce the effect of the high stresses generated, thus

44

eliminating the presence of a dent. This dent proved to not be as significant of a concern
as initially anticipated, because the strain gages were small enough to be placed outside

of the dent area (see Figure l-15).

45

CONCLUSION

The HP has been designed for use in the canine pancarpal arthrodesis procedure
in place of a traditional LC-DCP plate, which is a good general use plate. The HP was
specifically designed to address the complications of wound dehiscence and post-
operative metacarpal fractures associated with pancarpal arthrodesis. One clinical study,
using varying plate sizes with 10° of hyperextension, stated that 74% of 45 dogs returned
to normal walking and running on the leg, with 97% clinically improved to a degree that
the owners expressed satisfaction with the procedure.2 Another study reported that
arthrodesis occurred in all thirteen antebrachiocarpal joints implanted with a HP.9 Based
on Li’s study, the HP has been shown to be an effective implant when performing a
pancarpal arthrodesis in canines and is currently recommended for clinical use.

Based on the results of the current study, it was concluded that the HP yielded a
biomechanically superior construct than the LC-DCP. Results of the current study
indicated that the use of a HP for pancarpal arthrodesis provided a construct that was
more resistant to angular deformation and experienced less plate strain than a LC-DCP.
The biomechanical advantage gained with a HP construct may decrease the likelihood of
implant related complications and lessen the need for long-term postoperative coaptation

with no additional time or complexity in the surgical procedure.

46

REFERENCES

l. Lesser AS. Arthrodesis. In: Slatter D, Textbook of Small Animal Surgery. Saunders,
Philadelphia, 2003; 2174-2176

2. Parker RB, Brown SG. Pancarpal arthrodesis in the dog: A review of 45 cases. Vet
Surg 1981; 10: 35-43

3. Early TD. Canine carpal ligament injuries. Vet Clin North Am Small Anim Pract.
1978; 8: 183-199

4. Gambardella PC, Griffiths RC. Treatment of hyperextension injuries of the canine
carpus. Compend Contin Educ Pract Vet 1982; 4: 127-132

5. Fogle B. The Encyclopedia of the Dog. lSt ed. New York: DK Publishing, 1995
6. Evans HE: Miller’s Anatomy of the Dog, 3rd ed. Philadelphia: WB Saunders Co, 1993

7. Piermatti DL, Flo GL. Handbook of Small Animal Orthopaedics and Fracture Repair.
3rd ed. Philadelphia: WB Saunders, 1997; 344—348, 361-370

8. Pederson NC, Morgan JP, Vasseur PB. Joint diseases of dogs and cats. In: Ettinger
SJ, Feldmen EC, eds. Textbook of Veterinary Internal Medicine. 5111 ed. Vol 2.
Philadelphia: WB Saunders, 2000

9. Li A, Gibson N, Carmichael S, Bennett D. Thirteen pancarpal arthrodesis using
2.7/3.5 mm hybrid dynamic compression plates. Vet Comp Orthop Traumatol
1999; 12: 102-107

10. Okin R. Carpal arthrodesis in a cat with radial nerve damage. Fel Pract 1982; 12: 18-
20

l l. McKee M. Intractably painful joints. In: Houlton JE, Collinson RW, eds. Manual of
Small Animal Arthrology. Gloucestershire: British Small Animal Veterinary

Association, 1994: 122-125

12. Chambers JN, Bjorling DE. Palmar surface plating for arthrodesis of the canine
carpus. J Am Anim Hosp Assoc 1982; 18: 875-882

13. Whitelock RG, et a1. Metacarpal fractures associated with pancarpal arthrodesis in
dogs. Vet Surg 1999; 28: 25-30

14. DeCamp CE, Soutas-Little RW, Hauptman J, et a1. Kinematic gait analysis of the trot
in healthy greyhounds. Am J Vet Res 1993; 54(4): 627-634

47

15

16

17

18

19

20

. Hulse D, Hyman W, Nori M, Slater M: Reduction in plate strain by addition of an
intramedullary pin. Vet Surg 1997; 26: 451-458

. Johnson AL, Hulse DA. Diseases of the joints. In: Small Animal Surgery. Mosby
Inc., St. Louis 2002; 1091-1093

. Hibbeler RC. Mechanics ofMarcrials. 3rd ed. Upper Sadle River: Pearson Prentice
Hall, 2004

. Pohler O, Straumann F. The effect of pin size on reducing bone plate strain. Vet
Comp Orthop Traum 2000; 13: 185-19

. Black J. Biological Performance of Materials. 2nd ed. New York: Marcel Decker,
1992

. Marcellin-Little DJ, Deyoung BA, Doyens DH. Canine uncemented porous-coated
total hip arthroplasty: results of a long-term prospective evaluation of 50
consecutive cases. Vet Surg 1999; 28: 10-20

48

CHAPTER TWO

MECHANICAL EVALUATION OF AN INTERLOCKING NAIL
LOCKED WITH EITHER BOLTS OR A TYPE-IA EXTERNAL
SKELETAL FIXATOR IN A SIMULATED CANINE TIBIAL
MODEL

ABSTRACT

Interlocking nails (lLNs) can be implanted with small remote surgical approaches,
thus preserving soft tissue integrity and minimizing disruption to local vasculature. This
technique is synergistic with the trend toward biological fracture management. Due to
their biological advantages, the use of ILNs has gained popularity amongst orthopedic
surgeons in recent years. Nevertheless, with increased use, several limitations have also
been documented, both clinically and experimentally. Recently, an investigational ILN
system, featuring extended modified bolt-pins coupled to a type-IA external skeletal
fixator (ILN—ESF), has emerged as a technique to augment the stability of an ILN. The
extended bolts are used to lock the ILN, while serving as connecting pins for the ESF
frame. The purpose of this study was to use a tibial gap fracture model, featuring a
synthetic bone substitute to mechanically compare [LN-ESF and standard bolted ILN
(ILNb) constructs in torsion, 4-p0int bending and axial compression. The hypothesis was
that the ILN-ESF would be less compliant and sustain less deformation than a bolted nail
construct.

To reduce specimen variability, custom-made synthetic tibial bone substitutes
were used in this study. The tibial bone substitutes were implanted with standard ILNs

locked with either bolts or extended bolts connected with an ESF. Constructs were tested

49

in torsion, bending and axial compression (n:4/testing mode). Data, consisting of
construct compliance and associated deformation, were compared using t-tests.

The results of this study indicated that the [LN-ESF construct compliance and
deformation was significantly less than those of the ILNb construct in torsion, bending
and compression. Slack was present in the ILNb construct under torsion and bending, but
not in the [LN-ESF construct, regardless of testing mode.

This study showed that the substitution of locking bolts with extended bolts
connected to an ESF significantly reduced the construct compliance and overall
deformation in torsion, bending, and compression. Furthermore, the inherent slack of the
ILNb was eliminated by the use of an ESF in torsion and bending. The improvement in
structural properties of the [LN-ESF constructs could diminish interfragmentary motion

at the fracture site and potentially improve bone healing.

50

INTRODUCTION
Due to an advanced understanding of both human and veterinary fracture healing,
the concept of biological fracture fixation has evolved greatly over the past two decades.
The goal of biological fixation, particularly in comminuted long bone fractures, is to
achieve a balance between mechanical stability of the fracture repair and biological
preservation of the soft tissue."2 From a biomechanical standpoint in particular, the
function of fracture fixation is to provide sufficient stability, meaning that the
interfragmentary movement that occurs under external loading and muscle forces is
reduced to a degree which allows bone healing to occur.
Traditionally, it has been felt that comminuted long bone fractures were best
managed by bone-plating techniques (Figure 2-1). Proper application of bone plates
provides mechanical stability by counteracting axial (compression and tension), bending,

- 3.4
and rotatronal forces.

 

 

bone R

plate

 

 

 

 

Figure 2-1. Comminuted long-bone (tibia and fibula) fracture, before (left) and after (right)
fixation with a medially applied bone-plate.S

51

Despite their potential mechanical advantages, bone-plating techniques often
require extensive soft tissue dissection and operative time. It has also been shown that the
pressure of the plate on the surface of the bone can have detrimental effects on the

- . .932
vascularrzatron.6

The end result may be slower bone healing and increased post-
operative morbidity.

As an alternative to bone plates, interlocking nails (ILNs) have been introduced
for internal fracture fixation. Interlocking nails can be inserted into the bone with minor
surgical approaches, away from the traumatized area; thus preserving soft tissue integrity
and minimizing disruption to local vasculature and fracture hematomaf"9 Interlocking

nails provide a connection between the bone and nail, both proximal and distal to a

fracture site (Figure 2-2).

    
   

proximal
attachment

     

 
 

fracture site

distal
/ attachment

Figure 2- 2. Mediolateral radiograph of a tibial fracture S(left) internally fixated with an
interlocking nail (right). ASC

 

' MSU SACS: Michigan State University Small Animal Clinical Sciences. College of Veterinary Medicine,
Michigan State University. East Lansing, MI.

52

Several large clinical studies have reported good to excellent healing results and
low complication rates with the use of ILNs in both animals and humans.8'l0 Due to their
biological advantages, the use of ILNs has gained popularity amongst orthopaedic
surgeons in recent years. Nevertheless, with increased use several limitations have been
documented, both clinically and experimentally.‘M6 A recent in vivo study with an ILN
showed a significant reduction in bone healing and functional recovery attributed to
torsional and bending instabilities, when compared to an external fixator in an ovine tibial
fracture model.ll Similarly, an in vitro study of nine different human ILN designs
implanted in cadaveric tibiae demonstrated that consistent slack was present in all
constructs in torsion regardless of nail design.l5 Finally, a recent in vitro study'6 showed
that torsional compliance of ILNs was indeed greater than that of a plate and
intramedullary rod (IMR) combination (PRC); a fixation device often used in the
treatment of comminuted fractures. In addition, a mismatch between screw and nail-hole
diameter (Figure 2-3) was also identified in the ILN. This mismatch allowed movement

of the nail prior to engagement with the screw.

 

ILN

 

 

  
  

 

 

cortical
screw

 

 

 

I.

“+fijflikifi'.i .v‘ ~~~~~~~~~

'4 """"""

Figure 2-3. Diagram of mismatch between screw diameter (gray) and the ILN hole diameter
(black).

w’khhulk m W‘i’j 1

  

As a result, significant rotational instability (~25° slack) was still present following the
locking of the screw to the bone. In addition, structural deformation (flattening) of the

screw threads further increased the relative mismatch between the nail hole and screw

53

diameters, thus potentially accentuating the magnitude of the slack over time. These
studies suggest that current human or veterinary ILNs may have inherent torsional
instabilities, which could contribute to delayed bone healing and implant fatigue failure
such as clinically described screw bending or breakage.

To address the instabilities associated with ILNs, several modifications have been
devised. These can be divided into two major groups. The first group includes
modifications in the nail and/or screw designs. Based on a recent mathematical
evaluation of ILNs,l4 replacement of the locking screws with similar sized bolts was
recommended. The goal was to increase the stiffness of the locking mechanism by
providing a solid bolt-nail interface.“7 The bolts are threaded to lock on the cis-cortex of
the bone, but otherwise feature a smooth core that slides into the corresponding nail hole

(Figure 2-4).

 

 

threaded -
cis-cortex

 

 

 

Figure 2-4. Bolt for ILN, featuring a threaded cis-cortex section and a smooth, solid core.

The second group that addresses ILN instability supplements the ILN systems with other

”"9 such as stack pinning, double ILNs, ILN combined with an

modes of fixation,
intramedullary rod, and most recently an ILN with proximal “tie-in” configuration. All of
the above techniques, however, have the potential to increase surgical time, invasiveness,

and morbidity; thus defeating the biological purpose intended with the interlocking nail

technique.

External skeletal fixation (ESF) offers yet another alternative for fracture
reduction and stabilization. An ESF is a device, which stabilizes fractures with utilization

of several trans-osseous pins connected to one or more external bars (Figure 2-5).

 

' \i.

_, external (connecting)
I ' bar

 

 

 

 

transosseous pin

 

 

I

 

 

Figure 2-5. Comminuted long-bone (tibia and fibula) fracture, before (left) and after (right),
being externally fixated with a type-IA external fixator.5
Depending on the configurations of the pins and bars, ESFs are further classified into
type-IA or [B (Figure 2-6), type—11A or IIB (Figure 2-7), or type-III (Figure 2-8) with

- - - - , 5.20
Increasrng number of transosseous prns and connecting bars.

55

 

 

 

connecfing

bar connecung

bars

 

 

 

 

 

 

 

 

 

 

Figure 2-6. Type-IA (on left) is a one-plane, unilateral fixator with 6 transosseous pins (half pins)
and one connecting bar. Type-IB (on right) is a two-plane bilateral fixator with 8 transosseous
pins (half pins) and four connecting bars.S

 

 

full pin

 

 

full pin

 

 

 

 

 

 

 

 

half pin

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2-7. Type-11A (left) and type-IIB (right) ESFs. Notice the full pins in the type-[IA and the
full and half pins in type-IIB. The use of half pins results in a slight loss of stiffness compared to
a full-pin configuration.‘

56

 

 

 

 

 

Figure 2-8. Type-III ESF, notice the bilateral and biplanar configuration.5

Similar to ILNs, ESFs can achieve the goal of biological fixation with a
minimally invasive or percutaneous (closed) surgical approach to help preserve the soft
tissue environment. Studies have shown that the mechanical properties of the ESF are
dependent on the overall frame configuration.2| Geometric parameters influencing the
stability of fractures stabilized with external fixation include: transosseous pin diameter,
connecting bar diameter, free length between the connecting bar and the bone, distance

between the inner pin and the fracture site, and the distance between pins (Figure 2-9).

57

 

 

 

 

 

 

 

 

Ls

 

 

 

 

 

 

 

 

 

 

 

. . . . . . . . 22
FIgure 2-9. Geometric parameters Influencmg fracture stabrlrzatron With an ESF, where:
d: transosseous pin diameter
D: connecting bar diameter

L: free length between the connecting bar and bone

L1,L2: distance between pin and fracture site

Lg: distance between pins

In general terms, the construct stability may be increased by increasing the

2

number of pins and/or connecting bars,5‘20' 3 with type-I configurations (made of half
trans-osseous pins and a connecting bar in one plane only) being the weakest of all ESF
frames. However, the improved stability achieved with the addition of pins and bars in
various planes, may come at the expense of post-operative morbidities such as increased
risk of pin tract infection, extensive muscle impalement, and increased bulkiness of the
apparatus. For this reason, simple type-I ESFs are often used as adjunct fixations to
complement and strengthen a primary mode of fixation, such as an IMR.

Recently, an investigational hybrid ILN, featuring extended modified bolts-pins

coupled to a type-I ESF (ILN-ESF), has emerged as yet another technique to augment the

stability of an ILN.18‘24 The hybrid bolt-pins are used to lock the ILN while serving as

connecting pins for the ESF frame (Figure 2-10).

 

Figure 2-10. (Left) Radiograph of a tibial specimen implanted with the investigational hybrid
bolt- -pin locked with a type- IA external skeletal fixator SCSRS (Right) A close- -up photograph of
the hybrid 2.7mm ILN bolt/3.2mm ESF pin passing through one of the nail holes.

By combining the ILN to an ESF, the instabilities associated with each separate system
should theoretically be eliminated by the other, while retaining their own unique
biological and mechanical advantages. For instance, seated along the neutral axis of the
bone, ILNs have bending and torsional stiffness proportional to the cross section of the
nail. Specifically, the area moment of inertia (AMI) and polar moment of inertia (PMI)
are structural properties that characterize the ability of an implant to resist bending and
torsional deformation, respectively. Design consideration of these structural properties
allows ILNs to counteract bending and torsional forces deleterious to bone healing.
Further, by combining a simple type-I ESF with an ILN, substantial bending and
rotational stability may be achieved without utilization of a type III ESF, which is often

associated with a high rate of postoperative morbidity. This technique was successfully

 

' SCSRF: South Carolina Surgical Referral Service. Columbia, SC.

59

. . . . . . . [9.24
used In a limited number of clinical cases In small animal surgery.

A prospective
study, by Basinger et al., showed encouraging clinical use of the investigational hybrid
ILN-ESF construct in dogs.'9 Despite their clinical success, the proposed mechanical
stability has not been validated in an experimental setting.

The purpose of the current study was to use a tibial gap fracture model featuring a
synthetic bone substitute designed by the Comparative Orthopaedic Investigations
Laboratory (COIL) at Michigan State University, to mechanically compare hybrid ILN-
ESF and bolted ILN (ILNb) constructs in torsion, 4-point bending, and compression.
Several objectives were deemed necessary in the overall preparation and testing of these '
specimens. Because instability has been shown to be present in the ILN systems in

l '6 It was Important to create a srmrlar envrronment for the nail wrth

previous studies,l "
the synthetic bone model. For this to occur, consistent placement of the ILNb and ILN-
ESF constructs was absolutely necessary. Therefore, a goal was to develop several
custom designed fixtures which would allow consistent and accurate placement of the
fixators in the bone model. The ultimate goal was to acquire data that would support
application of the ILN-ESF in a wider range of clinical circumstances, including human
orthopaedics. The hypothesis of the study was that the ILN-ESF would be less compliant
than the ILNb in torsion, bending, and compression. We further hypothesized that the use

of an ESF would dampen or eliminate the acute torsional and bending instability (slack)

characteristic of standard nails.

60

MATERIALS AND METHODS

Specimen Preparation

Bone Model — In an attempt to limit specimen variability and circumvent the need to
procure canine bones, a custom—made synthetic tibial model was used in this study. The
material used for the bone model was 30% glass-filled structural nylon (Part No.
P0304909, 30% GF nylon natural, Alro Plastics, Jackson, MI). This material was chosen
on the basis of its material properties, which are similar to those of cortical bone (Table
2-1). To mimic a gap-fracture model, specimens were manufactured in 2 symmetrical
halves. Each half of the bone model featured a segment used for linkage to the holding
fixtures (cupping section), a tapering segment representing the metaphyseal section, and a
final segment representing the diaphyseal section (Figure 2-11). Thickness of the wall
was 2.5 mm throughout. Dimensions for the synthetic bone model were determined on
the basis of morphometric analysis conducted by Michigan State University Small
Animal Clinic (MSU SACS) on canine tibiae obtained from dogs weighing between 30
and 35 kg. The overall length of the tibial model (between each holding fixture and
including a 50 mm central gap) was 210 mm. Machining of the bone model, reducing it

from a solid rod to the bone model tube, was done by Olympia Tools (St. Johns, MI)

61

Human Canine 30% glass filled

 

cortical bone cortical bone structural nylon
Property

range; average range; average range; average
UTS (MPa) 53 — 135; 106.825 NA 65 - 195; 140
UCS (MPa) 145 —167; 158.8 25 112.8 25 140
1185 (MPa) 68 26 NA 59 - 85; 72
Young’s modulus: E (GPa) 8.2 — 17; 14.9 27 12.26 25 7.2
Poisson’s ratio 0.46 — 0.58; 0.49 27 NA 0.35
Density (g/cm3) 1.9 28 0.84 2" 1.35

Table 2-1. Comparative material properties of human and canine cortical bone and of the
composite material used in this study. Where UTS, UCS and USS is Ultimate tensile,
compressive and shear strength, resectively. ”'29

 

 

 

 

 

 

 

 

 

 

- 40 mm - ‘20 mm
a 50 mm M p
I -;.:.'.-'.': f, i-_}.'_.1-;_3;_:;;~h 2-5 mm
; . —-.__“:!__‘_‘:_- 7:77- h." ..- I... .:_-
2.5 mm ' '-r-"~r---' ' ‘
30 mm 25 mm 10mm 15 mm
.-L_..,_, 22.
“a”- “a: ...w 1“ ' ._,__,_.«_’____ .c '
, 12:: " .. .L_ ‘17;- ”4’“;""'
4 60 mm e
_______ \ +,_ Cup +9 Metaphysis , ->< >Diaphysis
L— " ' It, ---:7-.—"t":"‘~2 ‘ ' ”A: "."fl'w- :; i i i ‘3
E H Kg.~-A-—-4~—.w-m‘ ......... 4--- . - . 4"“ 5” ‘
I . | -,'.-.~ . . . ‘_ '
" ----- r'J'HL'WI-W “'"r' ----------------------- ~---- _. _ ;::'-;.:j ‘
C“*::'::" L--L2,;-'--’- "J 1” “T‘s-$.33; "7' r " 2:”:

Figure 2-11. Schematic depiction of the synthetic bone model (30% glass-filled nylon composite)
with the various dimensions of each segment (top) and a schematic depiction of the outline of an
actual canine tibia and fibula (red and gray dotted lines, respectively) superimposed over a
completed bone model with a novel ILN with 2 of the 4 locking SCPs (bottom).

Drilling Fixture — To allow accurate and consistent placement of the bolts and hybrid
bolt-pins, a dedicated custom-designed drilling fixture was fabricated to drill 2 pilot holes
in all bone models (Figure 2-12). The centers of the holes were exactly 1 1 mm apart to

match the nail holes.

 

Figure 2-12. Partially disassembled drilling fixture, with the top portion on the left and the
bottom on the right. The cupping section of the bone model is inserted into the bottom portion
with the metaphyseal and diaphyseal sections still visible. A spacer for minor adjustment is also
shown.

The drilling fixture featured a design made of three main pieces, consisting of a top
portion, which split into two symmetrical halves, and a bottom portion. There was also a
spacer made to compensate for minor variations in bone length due to machining
tolerances. The three-piece fixture allowed the bone to be placed into the fixture by

removing the symmetrical halves of the top portion, which featured removable drill

guides (Figure 2-13).

63

 

Figure 2-13. Top portion of drilling fixture, with symmetrical halves shown on the left and the
removable drill guides shown on the right.

Beyond locating two points which were 11 mm apart to drill the pilot holes, it was
important to ensure that these pilot holes were drilled perpendicular to the axis of the

bone, and not just normal to its slanted metaphyseal surface (Figure 2-14).

correct .
Incorrect

 

 

bone
- — — - ——————————————————————————— — model
axis

 

 

 

 

 

 

Figure 2-14. Schematic of correct (green) and incorrect (blue) way to drill pilot holes.

This was enabled by the design of the removable drill guides, which featured a geometry
matching the slope of the metaphyseal section of the bone model (Figure 2-15). The drill
guides also featured a radius, which matched that of the diaphyseal section of the bone.
An additional feature of the drill guide fixture was that the removable guides were

designed so that bushings of various sizes could be interchanged to accommodate for

64

various screw and bolt sizes.

 

Figure 2-15. Removable drill guide, featuring a geometry matching the slope of the metaphyseal
section of the bone model (white line), a radius matching the diaphyseal section (white rectangle),
and interchangeable bushings for various bolt and screw sizes (white ovals).

The bone was constrained axially and transversely by the radii of the removable drill
guides and the base of the bottom portion of the fixture, which matched the cupping
section of the bone model. The bottom portion of the fixture also featured two pairs of
steel bushings oriented perpendicular to each other, which restricted movement of the
bone model within the fixture. Other applications for the locking holes in the bone model
will be further explained later. When assembled, the drill guide fixture completely
enclosed the bone model, and the drill guides ensured that the drill bit was directed
throughout its path into the bone model (Figure 2-16). This standardized procedure
allowed all nails to be precisely and reliably centered longitudinally within the bone

model. Detailed drilling fixture operating instructions are found in Appendix 1.

65

   

pilot holes
( 1 1 mm apart)

‘ locking holes

Figure 2-16. Assembled bone drilling fixture, with bone model enclosed. A hand held drill is
shown drilling pilot holes.

Construct Preparation

Foam Plugs — Custom-made polyurethane foam plugs were inserted at the ends of each
bone model to maintain the ILNS in a centralized position during testing. The foam plugs
were meant to simulate cancellous bone in the metaphyseal section of the tibia. The foam
plugs were made of a two-part mixture molded into tubes. These tubes were then reduced
to dimensions fitting the inside of the metaphyseal section of the bone model by a custom

designed lathe tool (Figure 2-17).

   

Figure 2-17. Custom designed lathe tool, which created a geometry matching the inside of the
metaphyseal section of the bone model (shown in white).

66

The foam plugs were designed to grasp only 1-2 mm of the proximal and distal ends of
the nail. After lathing, the foam plugs were inserted into the ends of the bone (Figure 2-

18). Detailed foam plug instructions are given in Appendix 1.

 

Figure 2-18. Foam plug insertion into the bone model.

Construct Alignment Fixtures — An alignment fixture was used during implantation to
ensure that all of the nails were also axially aligned within the bone models (Figure 2-19).
The alignment fixture for the ILNb constructs featured two sliding frames which held the

bone model, and a fixed nail-clamping piece that matched the diameter of the nail.

nail clamping
tellon strips piece

   

 
  
 

a}.

  
   

     

guide rails
sliding
l'ramcs

Figure 2-19. ILNb alignment fixture.

The basic construction of the ILNb alignment fixture consisted of a base with two Teflon
strips which were in contact with the base of the sliding frames. The sliding frames

served as a method of holding the potting, which was directly attached to the bone model

67

(potting discussed later). The sliding frames, equipped with brass bushings and clamps,
were attached to two guide rails and a third rail, which could be clamped to secure
location of the cups. When the nail was in the proper position with respect to the bone,
the bolts or hybrid bolt—pins were inserted.

Another fixture was used to ensure proper alignment and distance relationships

between the ILN and external fixator (3.5 cm) (Figure 2-20).

 

ESF hybrid bolt-pin grooves

 

3.5 cm

Figure 2-20. ILN-ESF alignment fixture.

The ILN-ESF alignment fixture consisted of one solid block of aluminum. The block was
milled to place the ESF at the desired location with respect to the bone and ILN. Location

of the ESF was insured by grooves in the aluminum block, which restricted movement of

68

the ESF connecting bar, the ESF clamps, the ESF hybrid bolt-pins, the bone model, and

the ILN. Detailed alignment fixture instructions are given in Appendix 1.

ILN-ESF Tightening — The ESF connecting bar was attached to the hybrid bolt-pins via
four clamps (SKTM). To ensure consistency, all clamps were tightened to the same torque
(4.5 Nm). This was done using a digital torque wrench (CDI Computorq II, City, State

[Figure 2-21]).

 

Figure 2-21. CDI Computorq digital torque wrench shown tightening ESF clamps.

Potting Fixture — A uniform potting system was created in an effort to reduce specimen
variation. The potting was essentially the connection between the specimen and the
mechanical testing environment. In previous studies,'6 an epoxy had been used to create a
solid environment in which load could be transferred from the machine to the construct.
The actual procedure of potting specimens proved to be quite tedious, messy, and prone

to variation. With this in mind, an aluminum potting fixture was custom designed to fit

69

the bone models. The potting fixture featured two sections, an inner and outer shell

(Figure 2-22).

 

Figure 2-22. Disassembled potting fixture, inner shell (left) and outer shell (right).

When joined, the inner and outer shells created an inner and outer diameter matching the

cup section of the bone model with minimal tolerance (<0.01 mm [Figure 2-23]).

 

Figure 2-23. Assembled potting fixture with bone model inserted.

The aluminum pots were equipped with two pairs of axially aligned thru holes, equipped
with hardened steel bushings, which spanned the entire diameter of the potting. The

potting fixture also featured four threaded holes (oriented 90° from each other) with steel

70

threads which were used to lock the potting fixtures into various cups of the mechanical

testing fixtures (Figure 2-24).

   

   

thru holes

Figure 2-24. Potting fixture, which featured perpendicularly oriented thru holes for locking bone
to potting fixture and threaded—holes for locking potting fixture to testing fixture cups.

The drill guide fixture and potting were both designed so that the holes drilled in the
cupping section of the bone were aligned with the thru holes in the potting fixture.
Locking pins were then placed into the holes to rigidly attach the bone model to the

potting (Figure 2-25).

locking
pins

 

Figure 2-25. Potting fixture with partially inserted locking pins installed. The assembled
construct is shown.

71

Study Design

All ILNs in this study were 6 mm (diameter) by 185 mm (length) nails with 2
proximal and 2 distal 2.7 mm (diameter) holes. The corresponding locking bolts (2.7 mm
in diameter) and the ILNs were obtained from Innovative Animal Products, Rochester,
MN.

Each ESF frame was composed of 4 hybrid 2.7 mm ILN bolt/3.2 mm ESF pins
and 1 small SKTM titanium connecting bar (6.3 mm [diameter] by 150 mm [length]). The
hybrid bolt-pins were coupled to the connecting bar via 4 small SKTM clamps. All ESF
frames were obtained from [MEX Veterinary Inc., Longview, TX.

Implants were applied to the synthetic tibiae in 2 groups for a total 0f 12
specimens per group (n=4/testing mode). Group 1 (ILNb) was made up of 6 mm x 185
mm ILNs with four 2.7 mm bolts. Group 2 (ILN-ESF) was made up of 6 mm x 185 mm

ILNs with four hybrid 2.7 mm ILN bolt/3.2 mm ESF pins.

Mechanical Testing

Using custom designed loading fixtures, all specimens were mounted in an
Instron servo-hydraulic testing machine (Instron model 1331, Instron Corp, Canton, MA)
coupled to a 2500 lb (11,120 N) axial load cell (Model 1010AF-2.5K-B, Interface Inc.,
Scottsdale, AZ). The loading fixtures dedicated to bending and torsion were instrumented
with rotary encoders to document angular deformation of the constructs. In addition, the
torsion fixture was instrumented with a 1200 lb*in (135 N*m) torque load cell (Model
5330-1200, Interface Inc., Scottsdale, A2) to further document the magnitude of the

applied torques. All specimens were tested non-destructively in torsion, 4-point bending,

and compression (n = 4 per testing protocol). All tests were run in load control for 10
cycles. The actuator displacement, rotary encoder output, and corresponding load/torque
were documented in the tenth cycle. See Appendix 2 for detailed operating procedures for

the torsion, bending, and compression tests.

Torsion Tests — Torsion tests were run using a 0.125 Hz sinusoidal waveform at a torque
level of 1 5 Nm. This frequency was established in pilot tests and was based on the
frequency that generated minimal electrical noise in the response of the constructs during
this type of loading (where slack is present, such as in ILNb constructs). The torque level
chosen for this study was identical to that used in previous biomechanical investigations.
"’30 Torsion was achieved using a rack and pinion assembly that converted linear
displacement of the actuator into rotation of the proximal cup, while the distal cup
remained static. The torsion fixture was equipped with a rotary encoder (Baumer Electric,

Southington, CT) used to record angular deformation and a torque load cell (Figure 2-

26).

73

torque . .
. mnum
Itirttl t:t:ll

I'UIJIV

encoder

     

Figure 2-26. The torsion fixture shown with a loaded ILN-ESF.

Bending Tests — For purposes of this study, the constructs were only tested in
mediolateral bending (see Appendix 4 for data). Bending tests were run using a 0.125 Hz
sinusoidal waveform at a bending moment of i 3.5 Nm. The bending moment chosen for
this study was identical to that used in previous biomechanical evaluations of fixation

- - .30
devrces m a gap fracture model.'6

A specially designed fixture allowed for application
of a pure bending moment over the entire bone model/implant construct, as well as
alternate bending in the medial to lateral and lateral to medial directions. The bending
fixture had two cups, which articulated in one plane (mediolateral), equipped with rotary
encoders (Renco Encoders Inc., Goleta, CA) used to record specimen angular
deformation (Figure 2-27). The cups were attached to a loading arm, made up of a main

cross bar and two linkage bars, which allowed rotation at two points (at the connection to

the cup and at the main cross bar). The loading arm was attached to the axial load cell.

74

rotary
encoders

 

Figure 2-27. The bending fixture shown with a loaded ILNb.

Compression Tests — Compression tests were run using a '1 Hz haversine waveform to a
peak load of 176 N. This load level was chosen to simulate post-operative loading
conditions of a 30 kg dog. Therefore it was slightly less than the force acting on a healthy
canine limb during trot.3| Compression tests applied an axial load on the ILN through
displacement of the actuator. The compression fixture consisted of the axial load cell and

two compression cups to hold the potting fixtures (Figure 2-28).

75

load cell

  
 

compression
cups

 

Figure 2-28. The compression fixture shown with a loaded ILNb.

Data Acquisition

A load cell and/or a torque load cell coupled to the Instron actuator and to the
torsion fixture, respectively, recorded applied loads (N) and torques (Nm) over time.
These and the actuator displacement (mm) were recorded at a sampling rate of 500 Hz
(compression) and 250 Hz (bending and torsion). The difference in sampling rates
between testing conditions was dictated by computers limitations in handling the size of
the respective data files. Since all constructs were tested under load control, construct

compliance (slope of deformation versus load curve) was evaluated in the 10‘h cycle.

76

Construct compliance in bending and torsion was calculated as the mean compliances
during positive and negative loading. Construct angular deformation (torsion and
bending) was computed from the geometrical dimensions of the torsion and bending
fixtures as well as from rotary encoder data, while construct linear deformation
(compression) was obtained from actuator displacement data. See Appendix 2 for
calculations for torsion and bending tests.

As specimens were loaded in bending tests, the proximal and distal angles (or and

B, respectively) with respect to the horizontal were recorded via the rotary encoders

affixed to each articulated mount. Angular deformation (7) was computed with Equation

loading direction

apositive , Bpositive

2-1 (Figure 2-29).

   

  

 

 

 

“negative Bnegative

Figure 2-29. Construct angular deformation in bending

 

Yminimum = I anegative + Bnegative
Ymaximum = apositive + Bpositive [2'1]
YTotal (full cycle) = Ymaximum + Yminimum

Equation 2-1. Angular deformation, where positive and negative are in reference to the
applied moment, and:

7: construct angular deformation
a : deformed construct angle recorded by a rotary encoder
B : deformed construct angle recorded by a rotary encoder
The results of compliance and associated deformation of ILNb constructs and ILN-ESF

constructs were compared using unpaired t-tests. Student-Newman-Keuls post-hoe tests

were used whenever significant differences were identified (p < 0.05).

77

RESULTS
Performance of Construct Development Fixtures

The goal of developing and using a synthetic bone model was two—fold. One
reason was to eliminate the need of procuring biological specimens, which can require
significant time, effort, and expense. Experimentally, the elimination of biological
variation between specimens could potentially provide data, which has much less
variance between specimens. For this to occur however, a very reliable, efficient, and
consistent method of preparing the bone-implant constructs had to be developed. With
this in mind several fixtures were specially designed which were responsible for
everything from preparing the bone for implantation to mounting the construct in the
loading fixture. Theoretically, the goal was to eliminate the degree of human variation
while preparing specimens for mechanical testing.

The foam plugs, which were meant to simulate cancellous bone at the tibial
metaphysis, were created using a custom designed lathe tool. Consistency and efficiency
were the focal points in the design of the lathe tool. Figure 2-30 shows the consistent

final product prior to implantation into the bone model.

Figure 2-30. Polyurethane foam plugs created on a lathe with a custom designed tool.

 

The use of a drilling and alignment fixture was deemed successful based on the "final

bone-implant construct consistency in both the ILNb and ILN-ESF groups. Figure 2-31

78

shows the achievement of extremely consistent construct overall length (drilling fixture)

and orientation of the nail within the bone model (alignment fixture).

v

' l i- ’ . -
__ __ __ ____

BILNh

 

    
 

AILN-HSF

Figure 2-31. Fully assembled ILNb (top) and ILN-ESF (bottom) constructs, where:

A = overall construct length
AILNb = 270 mm = AlLN-ESF

B = testing length (length between loading cups)
BILNh= 2'0 mm = BlLN-ESF

C = osteotomy length (fracture gap)
CILNh = 50 mm = CILNJ-ZSF

79

Torsion Tests

Construct Compliance — Construct compliance curves for both the ILNb and ILN-ESF

were consistently bimodal in torsion (Figure 2-32).

Angular Deformation (degrees)

 

ll unloading
_ ........................................ ----—-----—--—-------------;—-----—,-.--.. 2,.
:I / [HI P;&,I
I ’4’ _,—r- ,W'fl‘fl‘
construct r’“ T: """ ..--' positive loading
maxrmum angular compliance “f: ----------- i -—
deformation —— 5 ;
I f I '
' 3l l
T i” iconstruct ;
. I ’ t
- I Tor ue N m
; ‘ islack q ( )
V

I
v
I
I
I
I
I
0
I
I
I
I
I

-..Jn-af-T'; ..........

1
negative loading 1”“? — 1;"- ------------- _.

 

I . — '4
C‘ ' I
/1 -r’ , "‘
'~. -..d
- 2

'
J.
r .2 d: ,4

"/4'

-IO-.1—II-Ol-lI-lflq

uriloading

 

ll

Figure 2-32. Typical bimodal torsion compliance curve.

The graphs of the ILNb (Figure 2-33) indicated no quantifiable torque in the central

region, corresponding to the change in the direction of applied torque.

10

ILNb

 

 

NON-#0303

gular Deformation (degrees)
0') 4'5 '

I
(D

 

 

 

 

An
8

-4

0
Torque (Nm)

Figure 2-33. Representative torsion curve for ILNb, notice the presence of slack.

80

This central region (torque = 0) represented the slack in the construct and reflected an
abrupt change in angular deformation without resistance to applied torque. The bimodal

appearance of the ILN-ESF curves was different (Figure 2—34).

ILN-ESF

 

 

 

 

 

Angular Deformation (degrees)

 

-6 -4 -2 O 2 4 6
Torque (N*m)

Figure 2-34. Representative torsion curve for ILN-ESF, notice the lack of slack and change
in compliance.

With ILN-ESF, there was no characteristic vertical region of the curve where the
direction of torque was changed, thus indicating that no slack was present. There was,
however, a visible and quantifiable change in the compliance of the construct throughout
its cycle (Figure 2-34). Therefore, both initial (IC) and terminal (TC) compliances were
computed for the ILN—ESF constructs. The initial (ILN-ESF construct only) and terminal
(ILN-ESF and ILNb constructs) compliances were calculated as the slope of the
deformation versus torque curve from 0-1 Nm and 2.5-5 Nm, respectively. Compliance

was calculated as the mean of the positive and negative loading curves because there was

81

not a significant difference between the positive and negative compliances in either group
for either test (p > 0.05).

There was a significant difference in the initial (1.88 :I: 0.29 O/Nm) and terminal (0.87
i 0.1 °/Nm) compliances for the ILN—ESF constructs (p < 0.05 [Table 2-2]). Terminal
compliance of the ILN-ESF constructs was significantly less than that for the ILNb
constructs (p < 0.05). Terminal compliance of the ILN-ESF construct was approximately
25% less compliant than that of the ILNb constructs (0.87 i 0.1 °/Nm and 1.16 i 0.01

°/Nm, respectively).

Construct Angular Deformation — Maximum angular deformation for the ILN-ESF
constructs was approximately 30% less than that for the ILNb constructs (12.2 :t 087°
and 17.74 i 0.45°, respectively). Construct slack, where present, was computed as the
difference between the y-axis intercept of the compliance slope on the positive and
negative loading curves (Figure 2-32). ILNb constructs exhibited 6.12 : 05° of slack.
While the ILN-ESF had no region of slack, there was both an initial and terminal
compliance in the constructs (Figure 2-34). At high torques, ILN-ESF constructs
deformed approximately 5% more than the ILNb constructs (12.2 : 087° and 11.62 :L-
0.06°, respectively), once the slack was overcome at high torques. These differences were

not statistically significant (p=0.234)

82

 

Angular Def. Def. — Slack
(°) ” (°)
ILNb NA 1.16 i 0.01 6.12 $0.5 17.74 i 0.45 11.62zt0.06

Construct IC(°/Nm) TC (°/Nm)# Slack (°)

 

 

ILN-ESF 1.88zt0.29* 0.87:I:O.l* none 12.2 :t 0.87 12.2 d: 0.87

 

 

 

 

 

 

 

 

Table 2-2. Torsion data (mean i standard deviation), indicating significant differences (p<0.05)

# between ILNb and ILN-ESF constructs
* within ILN-ESF group

(see Appendix 4 for exact p-values)

Bending Tests

Construct Compliance — Construct compliance curves for the ILNb were also
consistently bimodal in bending (Figure 2-35), whereas ILN-ESF curves were unimodal.
The bimodal graphs of the ILNb indicate that there was no quantifiable moment in the
central region, which corresponded to the change in the direction of the applied bending
moment. For all practical purposes this region represented slack in the construct and
reflected an abrupt change in angular deformation without resistance to the applied
bending moment. Conversely, with unimodal shape of the ILN-ESF compliance curves
(Figure 2-36), there was no measurable region of slack for these specimens. Construct
compliances were calculated as the slope of the deformation versus moment curve from
1.5-3.5 Nm. Compliance in the ILN-ESF constructs was significantly less than that for
the ILNb constructs (p < 0.05 [Table 2-3]). Specifically, the ILN-ESF construct was
approximately 60% less compliant than the ILNb constructs (0.29 i: 0.01 0/Nm and 0.7 :t

0.05 O/Nm, respectively).

83

ILNb

 

 

-2 j .
-4

 

 

 

 

Angular Deformation (degrees)

-4 -2 0 2 4
Moment (N‘m)

Figure 2-35. Representative mediolateral bending curve for ILNb, notice the bimodal shape
and presence of slack.

 

 

 

 

 

 

 

ILN-ESF

8
8 6 -
Q d
a. 4 -
e .
3 , .
C 2 _L
.2 t
g 0 t)—+—I-—I-—+—I—I—I=I=I=u-=F==I=—I—I_o—I—I—o—l
E .___ II
o '2 cl
‘5
o -4 -
a
'5 '5 ‘
en 4
5; -8

-4 -2 0 2 4

Moment (N*m)

Figure 2-36. Representative mediolateral bending curve for ILN-ESF, notice the unimodal
shape due to an absence of slack.

Construct Angular Deformation — Maximum angular deformation developed in the ILN-
ESF constructs was approximately 80% less than that in the ILNb constructs (2.09 i

0.004 O and l 1.38 i 0.45 0, respectively). ILNb constructs exhibited slack (6.51 i 0.05 0)

84

(Figure 2-35), whereas the ILN-ESF had a relatively constant resistance to load
throughout their bending cycle (Figure 2-36). Angular deformation for the ILNb
constructs, once slack was overcome (at high torques), was approximately 60% greater

than that of the ILN-ESF constructs (4.88 i 0.44 0 and 2.09 t 0.01 0, respectively).

 

 

 

 

. o # o Angular Deformation
Construct Compliance ( le) Slack ( ) Deformation (o) # _ Slack (0)4
ILNb 0.7 5.- 0.05 6.51 t 0.05 I 1.38 i 0.45 4.88 :1: 0.44
ILN-ESF 0.29 i 0.01 none 2.09 :t 0.004 2.09 :1: 0.01

 

 

 

 

 

Table 2-3. Bending data (mean : standard deviation), indicating significant differences (p <
0.05):

# between ILNb and ILN-ESF constructs.

Compression Tests

Construct Compliance and Displacement - Construct compliances were computed as the
slope of the displacement versus load curves from 50—176 N (Figure 2-37). Compliance
of the ILN-ESF construct was significantly less than that of ILNb constructs (p < 0.001).
ILN-ESF constructs were approximately 45% less compliant (5.35 1: 0.89 mm/N*E-04
and 9.78 t 0.86 mm/N*E-04, respectively) than ILNb constructs and displaced
approximately 40% less than ILNb constructs (9.28 i 1.8E mm*E-02 and 16 i 0.7

mm*E-02, respectively [Table 2-4]).

85

 

ILNb

 

0475

045

0425

Displacement (mm)
0 0
<3 53 ‘0
01 0'1 -t

 

 

 

 

 

0025
0
O . 50 100 150 200
Load (N)
ILN-ESF
0475
045

O
N
01

F’
—L

005

Displacement (mm)
8
V
01

0025

 

 

 

 

0 50 100 150 200
Load (N)

Figure 2-37. Representative compression curves for ILNb (top) and ILN-ESF (bottom).

 

Construct Compliance (mm/N*E-04) # Displacement (mm*E-02) #

 

ILNb 9.78 i 0.86 16 i 0.7
ILN-ESF 5.35 i 0.89 9.28 i 1.8

 

 

 

 

 

 

Table 2-4. Compression data (mean i standard deviation), indicating significant differences (p <
0.05):

# between ILNb and [LN-ESF constructs.

86

DISCUSSION

The hypothesis of the present study was that the [LN-ESF construct would be less
compliant than the ILNb in torsion, bending, and compression. The results of the torsion
tests in the current study indicate that there was a significant variation in compliance of
the ILN-ESF construct as it was loaded. The initial and terminal compliances of the
construct were 1.88 i 0.29 °/Nm and 0.87:0.1 O/Nm, respectively. Furthermore, the
terminal compliance of the ILN-ESF constructs were 25% less compliant than the ILNb
constructs (0.87 and 1.16 °/N m, respectively) and deformed approximately 30% less. The
results of the bending tests indicated that the ILN-ESF constructs were approximately
60% less compliant than ILNb constructs (0.29 i- 0.01 °/Nm and 0.7 i 0.05 °/Nm,
respectively) and underwent approximately 80% less deformation (7.97 '1: 005° and 9.45
1 022°, respectively). Finally, as seen with torsion and bending, the compression tests
showed similar trends, specifically, the ILN—ESF constructs were approximately 45% less
compliant than the ILNb constructs (5.35 i- 089 and 9.78 i 0.86 mm/N*E-04,
respectively. The substitution of standard ILN locking bolts with an ESF in the current
study significantly reduced the construct compliance and overall deformation in torsion,
bending and compression, thus creating a more stable construct. This added stability
could create the biomechanical environment necessary for improved fracture healing, by
eliminating excessive interfragmentary motion.

Although the exact level of fracture stability necessary to optimize bone healing
has yet to be determined, numerous studies have suggested that, while controlled axial
micromotion promotes bone healing, excessive interfragmentary shear motion often

21.3

results in delayed ossification of the fracture site. 3'36 In my study, the substitution of

87

locking bolts with extended bolts connected with a type [A ESF significantly reduced
construct compliance and overall deformation in torsion, bending and axial compression
to create a more stable construct. Perhaps more importantly, the use of an ILN-ESF
appeared to have successfully eliminated construct slack. By eliminating excessive and
uncontrolled interfragmentary motion, the added stability provided by ILN-ESF could
create a biomechanical environment more favorable for fracture healing.

Because tibial fractures in people are inherently unstable in torsion and

37'“) the fixation method should efficiently counteract torsional and

mediolateral bending,
bending stresses to which these fractures are subjected post-operatively. However, both
in vitro and in vivo experimental studies have documented torsional and bending

11.12.15 -
The recent comparison of a human

instability with various human ILN designs.
ILN to an ESF in an ovine tibial fracture model showed significantly more
interfragmentary motion with inferior bone healing and incomplete return to function in
the ILN treated sheep compared to the ESF group.H In an effort to improve fracture
repair stability, experimental angle-stable nails have been recently designed in human and
veterinary orthopaedics.‘”'42 Both systems showed that through re-engineering of the
locking system, construct stability is greater than that offered by conventional nails,
which in turn results superior bone healing and more complete functional recovery.42 One
can speculate that because of the comparable muscle distribution between humans and
dogs, tibial fractures in dogs may undergo similar deformation patterns if ineffectively
stabilized. Ironically, the limited ability of current veterinary nails to counteract torsion
and bending parallels that of human nails. Indeed, torsional slack was first reported in an

. . . . . . . . . . |(
m wtro evaluation of screwed Interlocking na11s us1ng a canlne t1b1al fracture model. ’

88

Concurrently, intra- and post-operative instability have been documented in 12-14% of
fracture cases treated with standard veterinary ILNs.9"8"9 To provide adequate stability
and reduce the risk of delayed or non-union, various supplementations of current
veterinary ILN systems with other modes of fixation have been devised.'7"9‘24 The
development of an [LN coupled to a type IA ESF represents one of the recent techniques
used to more effectively stabilize diaphyseal fractures in dogsm‘w'24

While the angular deformation of both constructs in response to applied torques
was biphasic, the torsional compliance of the bolted nail constructs at low torques was
not measurable. A similar behavior has been documented in screwed ILNs and was
attributed to construct slack inherent to the design of current ILN systems.” In contrast,
the initial compliance of the ILN-ESF constructs at low torque levels was quantifiable.
This anticipated behavior was attributed to the resistance of the external fixator frame
alone, mainly through early and continuous flexural deformation of the extended locking
bolts. From a mechanical point of view, the external fixator serves as a pathway for
accepting the torsional loads applied to the bone through bending of the transosseous
pins. A torsional force on the bone model is like a load applied perpendicular to the

longitudinal axis of the pin. This load results in bending of the transosseous pin,

comparable to a cantilever beam in bending (Figure 2- 38).

89

If

{y lulqllt

 

Figure 2- 38. Schematic of the load transfer between the bone and ESF. In vitro setup is shown
on the left and beam theory correlation is shown on the right.

In beam theory, the stiffness of a cantilever beam can be calculated with:

P 3 El
y D [2.2]
Equation 2-2. ESF stiffness (k) equation, where:
P = applied load
L = length

E = elastic modulus
I = area moment of inertia
y = deflection

Huiskes et al., developed a more complete model of ESF stiffness, taking mechanical
properties of the connecting rod as well as the pin configuration into account. That model
was also based on cantilever beam theory and gave results that were in good agreement

with both experimental and finite element analysis.43 However, the Huiskes model was

90

based on axial compression tests, without the connection of an ESF to an ILN, therefore
it could not be directly applied to the current study.
The initial compliance of the ILN-ESF constructs was directly influenced by the working
length of the extended locking bolts, i.e. the length between the connecting bar and the
outer edge of the bone model.23'3O A previous study showed that the stiffness (the inverse
of compliance) of an ESF construct in torsion significantly increases when the distance
between the connecting bar of the ESF and the bone is decreased from 1.5 cm to 0.5
cm.30 Therefore, to reduce inter-specimen variability, the distance between the ILN and
the connecting bar was standardized in this study. As the deformation of the ILN-ESF
constructs increased in response to increased torque levels, the extended locking bolts
began to engage the nail. The added contribution of the nail in resisting construct
deformation resulted in a significant decrease in the ILN-ESF construct compliance at
high torque levels and accordingly, the terminal compliance of the ILN-ESF constructs
was significantly smaller than that of the ILNb constructs.

The slack in the ILNb construct was due to a mismatch between bolt and nail hole
diameter. This mismatch allowed the nail to move a considerable amount prior to

engagement with the bolt (Figure 2-39).

91

   

 

torsion
applied to
nail

 

 

 

 

Figure 2-39. Schematic of the nail without applied torque (left) and after torques is
applied (right). The nail hole axis is white and the bolt axis is black, notice the amount of nail
rotation (9) prior to engagement between the bolt and nail.

The significantly smaller torsional deformation of the ILN-ESF constructs was
mainly due to the elimination of construct slack. Indeed, once the slack inherent to the
bolted nail system was overcome, both ILN-ESF and ILNb constructs underwent similar
overall deformation (12.2 1 087° and l 1.6 i 05°, respectively). Interestingly, a previous
study16 showed that the torsional deformation of tibial constructs stabilized with a plate-
rod-combination was similar in magnitude (1 1.66 1 259°) to that of ILN-ESF constructs.
This suggested that this hybrid system should be effective in counteracting fracture forces
in clinical cases. While the current study suggested that the use of bolts instead of screws,
as used in the previous study,'6 substantially reduced the amount of slack, it also
demonstrated that construct slack can be eliminated by using extended bolts connected to
an ESF frame. Similar observations have been reported in a clinical study where intra-
operative torsional slack associated with the use of an ILNb was eliminated with the

addition of a type IA ESF.l8

92

In bending tests, the differences seen in the angular deformations can again be
attributed to slack in the ILNb constructs (6.51 i 005°). However, the differences in
compliance were most likely due to the external fixator acting as a pathway for accepting
the applied loads through bending of the transosseous pin. While the addition of an ESF
reduced construct torsional compliance by approximately 25%, the same procedure had a
more dramatic effect in bending where the use of an ESF resulted in a 60% decrease in
construct compliance. As in torsion, this bending deformation resulted in the transosseous

pins being subjected to loading perpendicular to their longitudinal axes (Figure 2- 40).

', E1 lu-i ‘

 

 

 

 

 

 

 

 

 

 

 

   

Figure 2-40. Schematic of the applied moment on the bone model resulting in deflection of the
transosseous pins.

Furthermore, the bending compliance curves of the ILN—ESF constructs were
consistently unimodal. This indicates that ILN-ESF constructs resisted applied bending

16

moments in a continuous manner similar to that of bone plates. The effectiveness of the

93

ESF in mediolateral bending was also reflected by the small angular deformation, as well
as by the absence of slack compared to the bolted nails.

Compression showed that the ILNb and ILN-ESF constructs both had small
deformations (approximately 0.16 and 0.09 mm, respectively). Several in vivo studies
have shown that controlled axial micromotion may enhance bone healing,2"33‘36 therefore
with this amount of deformation (less than 2 mm), the use of either construct could be
advantageous for healing of fractures from an axial micromotion perspective.

Because standard and extended bolts have identical diameters, it was speculated
that extended bolts alone would not be capable of providing construct stability. To test
this theory (see Appendix 4 for data), the connecting bar was removed from the ILN-ESF
specimens following torsion and bending tests and retested in torsion. As expected, the
compliance, angular deformation and slack of the ILN constructs with extended bolts
alone were identical to those of the ILNb constructs. Therefore, one must recognize that
the integrity of the connecting clamps, and not the extended bolts, is essential to prevent
torsional and bending slack in ILN-ESF constructs. Should the ESF connecting bar
become lose, the stability of the repair would be compromised.

A limitation of this study was the use of a synthetic bone model as a substitute for
cadaveric tissue. However, the bone model greatly limited specimen variability and
thereby allowed for better evaluation of the implant itself. The shape and material
properties of bones used in a human cadaveric study 44 have standard deviations (SDs)
that exceeded 100%. These SDs are likely to be even greater in dogs on the basis of the
large number of breeds and conformations. Therefore the canine tibia model was created

to limit interspecimen variability and thereby allow for a better evaluation of the implant.

94

37.45
however,

Simple tubes of various materials have been used for fatigue testing;
complex shapes may be more appropriate to model physiological loads when devices for
fracture fixation are evaluated.46 In an attempt to create a more realistic tibial model,
tapered ends created on the basis of measurements obtained by MSU SACS were
incorporated into the model to mimic the larger metaphysis. In another experimental
study 45 that used a bone model, ILNs were more susceptible to failure when implanted in
the center of larger aluminum bone models. Accordingly, to subject the ILN-ESF to the
most stringent conditions and allow meaningful comparison between groups, all nails
were locked near the ends of the model (outside diameter, 26 and 23 mm) and were
maintained in a central location by use of a polyurethane plug.

In nonfailure tests, mechanical properties that directly affect any alteration of the
screw-bone interface during loading have not been determined. Therefore, general
mechanical properties of human and canine cortical bone were used as guidelines for
choosing an acceptable material. Although models of human tibiae are commercially
available (Sawbones, Pacific Research Laboratories Inc., Vashon, WA), they are too
large to represent tibiae of giant breeds (405 to 375 mm in length), and were cost
prohibitive. The 30% glass-filled structural nylon composite chosen for use in the study
reported here was commercially available and relatively inexpensive.

The development of the bone model required the development of numerous
dedicated fixtures, which could allow the implants to be effectively and consistently
implanted within the bone model. These fixtures produced constructs lacking the
variability that could be introduced with implantation, such as implant positioning and

overall length differences between constructs. In addition, the results of the study also

95

illustrated the effectiveness of not only the bone model itself, but also the fixtures that
enabled its use. In a previous study,'6 angular deformations under torsion had standard
deviations as high as 22% of the mean; standard deviations in the current study of similar
tests were 7% of the mean. Minimum standard deviations in the current study were as
low as 0.2% of the mean (ILN-ESF angular deformation in bending), compared to
minimum deviations of 7% of the mean (ILN constructs locked with screws) in similar
studies.'6 The reduction in standard deviations compared to the previous study was
mainly attributable to the development of dedicated fixtures, which enabled the bone
model to be implanted with the ILNs and ILN-ESFs consistently. This was based on the
fact that both studies were conducted in the same laboratory (Orthopaedic Biomechanics
Lab) with the same equipment (Instron), and almost identical loading fixtures. Thus, the
current study with this newly developed bone model yielded numerous advantages over

natural specimens for the functional evaluation of construct designs.

96

CONCLUSION

The addition of an ESF to currently available ILN systems was meant to eliminate
the instabilities associated with each system when used individually, while retaining their
own unique biological and mechanical advantages. This technique was successfully used
and documented in a limited number of clinical cases in small animal surgery.'9'24 One
such study showed encouraging clinical use of the investigational hybrid ILN-ESF
construct in dogs.”

The results of the current study suggested that an ILN locked with hybrid bolt-
pins coupled with an ESF, compared with an ILN locked with standard bolts,
significantly reduced construct compliance and overall deformation in torsion, bending,
and compression. Furthermore, the inherent slack of the bolted ILN was eliminated by
the supplementation of an ESF in torsion and bending. The improvement in construct
stability documented in the ILN-ESF constructs could diminish interfragmentary motion
at the fracture site and potentially improve bone healing. Accordingly, ILN-ESF may
represent an effective alternative to standard ILN fixation, particularly in the treatment of

diaphyseal tibial fractures.

97

REFERENCES

. Field JR, Tomkvist H. Biological fracture fixation: a perspective. Vet Comp Orthop
Traumatol 14: 169-178, 2001

Aron DN, Palmer RH, Johnson AL. Biologic strategies and a balanced concept of
repair of highly comminuted long bone fractures. Comp Cont Ed Pract Vet 17:
35-38, 1995

Bernarde A, Diop A, Maurel N, et al. An in vitro biomechanical study of bone plate
and interlocking nail in a canine diaphyseal femoral fracture model. Vet Surg 30:
397-408, 2001

Roe SC. Mechanical characteristics and comparisons of cerclage wires. Introduction
of the double-wrap and loop/twist tying methods. Vet Surg. 1997 Jul-Aug; 26 (4):
310-6

Piermattei DL, Flo GL. Fractures: classification, diagnosis, and treatment, in
Handbook of Small Animal Orthopedics and Fracture Repair (ed 3rd).
Philadelphia, W. B. Saunders, 1997, pp 24-146

Moses PA, Lewis DD, Lanz 01, et a1. Intra-medullary interlocking nail stabilization
of 21 humeral fractures in 19 dogs and one cat. Aust Vet J 80: 336-343, 2002

Durall I, Diaz-Bertrana MC, Morales I. Interlocking nail stabilization of humeral
fractures: initial experiences in seven clinical cases. Vet Comp Orthop Traumatol
7: 3-8, 1994

Dueland RT, Johnson KA, Roe SC, et a1. Interlocking nail treatment of diaphyseal
long bone fractures in dogs. J Am Vet Med Assoc 214: 59-66, 1999

Duhautois B. Use of Veterinary Interlocking Nails for Diaphyseal Fractures in Dogs
and Cats: 121 Cases. Vet Surg 32: 8-20, 2003

10. Klemm KW, Bomer M. Interlocking nailing of complex fractures of the femur and

tibia. Clin Orthop Relat Res. 1986 Nov; (212): 89-100

1 l. Klein P, Opitz M, Schell H, et a1. Comparison of unreamed nailing and external

fixation of tibia diastases-mechanical conditions during healing and biological
outcome. J Orthop Res 22: 1072-1078, 2004

12. Laflamme GY, Heimlich D, Stephen D, et a1. Proximal tibial stability with

intramedullary nail fixation using oblique interlocking screws. J Orthop Trauma
17: 496-502, 2003

98

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

Knothe U, Knothe TML, Klaue K, et al. Development and testing of a new self-
locking intramedullary nail system: testing of handling aspects and mechanical
properties. Injury 3 1: 617-626, 2000

Lin J, Lin SJ, Chen PQ, et a1. Stress analysis of the distal lockinig screws for femoral
interlocking nailing. J Orthop Res 19: 57-63, 2001

Schandelmaier P, Krettek C, Tscherne H. Biomechanical study of nine different
tibia-locking nails. J Orthop Trauma 10: 37-44, 1996

Von Pfeil DJF, De’jardin LM, DeCamp CE, et al. In vitro biomechanical comparison
of plate-rod combination-construct and an interlocking nail-constructs for
experimentally induced gap fractures in canine tibiae. Am J Vet Res 66: 1536-
1543, 2005

Suber JT, May BM, Basinger RR. A comparison of bending and gap stiffness
between interlocking nails and interlocking nails supplemented with stack pins.
Vet Comp Orthop Traumatol, In press

Basinger RR, Suber JT. Two techniques for supplementing interlocking nail repair of
fractures of the humerus, femur, and tibia: results in 12 dogs and cats. Vet Surg
33: 673-680, 2004

Basinger RR, Suber JT. Supplemental fixation of fractures repaired with interlocking
nail: 14 cases. In: Proceedings, 29'h An Meet Vet Orthop Soc, The Canyons, UT:
27 2002

Marcellin-Little DJ. External skeletal fixation, in Slatter D (ed): Textbook of Small
Animal Surgery (ed 2). Philadelphia, Saunders, 2003, pp 1818-1834

Goodship AE, Watkins PE, Rigby HS, et al. The role of fixator frame stiffness in the
control of fracture healing. An experimental study. J Biomech 26: 1027-1035,
1993

Claes LE. Biomechanics of fracture repair and fracture fixation. In: Mow VC, Hayes
WC, eds. Basic Orthopaedic Biomechanics and Mechano-Biology. Philadelphia:
Lippincott Williams and Wilkins; 2005

Johnson WD, Fischer DA. Skeletal stabilization with a multiplane external fixation
device. Biomechanical evaluation and finite element model. Clin Orthop Rel Res
180: 34-43, 1983

Nanai B, Basinger RR. Use of a new investigational interlocking nail supplement in
the repair of comminuted diaphyseal tibia fractures in two dogs. J Am Anim Hosp
Assoc 41: 203-208, 2005

99

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

Yamada H. Mechanical properties of locomotor organs and tissues. In: Evans FG, ed.
Strength of biological materials. Baltimore: The Williams & Wilkins Co, 1970;
20, 54

Reilly DT, Burstein AH. The elastic and ultimate properties of compact bone tissue.
J Biomech 1975; 8: 393—405

Carter DR, Spengler DM. Mechanical properties and composition of cortical bone.
Clin Orthop Relat Res 1978; 135: 192—217

Silbernagel JT, Kennedy SC, Johnson AL, et a1. Validation of canine cancellous and
cortical polyurethane foam bone models. Vet Comp Orthop Traumatol 2002; 15:
200—204

Mat-Web. http://www.matweb.com/search/SpecificMaterial.asp?
bassnumzPTSXBIZ 1

Bouvy BM, Markel MD, Chelikani S, et a1. Ex vivo biomechanics of Kirschner-
Ehmer external skeletal fixation applied to canine tibiae. Vet Surg 22: 194-207,
1993

DeCamp CE, Soutas-Little RW, Hauptmann J, et a1: Kinematic gait analysis of
the trot in healthy greyhound dogs. Am J Vet Res 54:627—634, 1993

Claes LE, Eckert-Hubner K, Augat P. The effect of mechanical stability on local
vascularization and tissue differentiation in callus healing. J Ortho Res 20: 1099-
1 105, 2002.

Claes LE, Heigele CA. Magnitudes of local stress and strain along bony surfaces
predict the course and type of fracture healing. J Biomech 32: 255-266, 1999

Claes LE, Heigele CA, Neidlinger-Wilke C, et al. Effects of machanical factors on
the fracture healing process. Clin Orthop Relat Res Suppl 355: 8132-5 147, 1998

Lewallen DG, Chao EY, Kasman RA, et a1. Comparison of the effects of
compression plates and external fixators on early bone-healing. J Bone Joint Surg

Am 66: 1084-1091, 1984

Wu J], Shyr HS, Chao EY, et a1. Comparison of osteotomy heling under external
fixation devices with different stiffness characteristics. J Bone Joint Surg Am 66:
1258-1264, 1984

Gaebler C, Berger U, Schandelmaier E, et a1. Rates and odds ratios for
complications in closed and open tibial fractures treated with unreamed, small
diameter tibial nails: a multicenter analysis of 467 cases. J Orthop Trauma 2001;
15: 415-423

100

38.

39.

40.

41.

42.

44.

45.

46.

Freedman EL, Johnson EE. Radiographic analysis of tibial fracture malalignment
following intrmedullary nailing. Clin Orthop 315: 25-33, 1995

Mercado EM, Lim EV, Stern P], et a1. Exchange nailing for failure of initially
rodded tibial shaft fractures. Orthopedics 24: 757-762, 2001

Safran O, Liebergall M, Segal D et a1. Proximal tibial fractures - should we nail
them? Am J Orhtop 30: 681-684, 2001

Déjardin LM, Lansdowne J L, Sinnott MT, et al. In vitro mechanical evaluation of
torsional loading in simulated canine tibiae for a novel hourglass-shaped
interlocking nail with a self—tapping tapered locking design. Am J Vet Res
67:678-685, 2006.

Kaspar K, Schell H. Seebeck P et a1. Angle stable locking reduces interfragmnetary
movements and promotes healing after unreamed nailing. Study of a displaced
osteotomy model in sheep tibiae. J Bone Joint Surg Am 87: 2028-2037, 2005

. Huiskes R, Chao EYS. Guidelines for external fixation frame rigidity and stresses. J

Orthop Res 4: 68-75, 1986

Hulse D, Hyman W, Nori M, et a1. Reduction in plate strain by addition of an
intramedullary pin. Vet Surg 1997;26: 451—459

Aper RL, Litsky AS, Roe SC, et al. Effect of bone diameter and eccentric loading on
fatigue life of cortical screws used with interlocking nails. Am J Vet Res 2003;
64: 569-573

Szivek JA. Synthetic materials and structures used as models for bone. In: An YH,
Drauhgn RA, eds. Mechanical testing of bone and the bone-implant interface.
Boca Raton, Fla: CRC Press LLC, 2000; 159-171

101

CHAPTER THREE

COMPREHENSIVE MECHANICAL EVALUATION OF FRACTURE
F IXATION SYSTEMS: COMPARISON OF PLATED VS. SCREWED,
BOLTED AND NOVEL INTERLOCKING NAIL CONSTRUCTS

ABSTRACT

Interlocking nails (ILNs) are attractive alternatives to plate fixation for repair of
comminuted fractures due to several mechanical and biological advantages. While the use
of [INS has yielded favorable clinical outcomes in both human and animal patients,
several limitations have been documented clinically and experimentally. Specifically,
current ILN systems do not counteract torsional and bending forces as much as initially
anticipated, which could explain the complications reported clinically. In an attempt to
reduce instability inherent to ILN constructs, a novel interlocking nail and locking
mechanism were designed (ILNn) by our research group. The novel nail featured a
tapered self-locking mechanism engineered to improve implant stability along with an
hourglass profile, which tapered from 8-mm extremities to a 6-mm central core and was
designed to preserve the biological advantages of ILN fixation. The purpose of this study
was to use a previously described tibial fracture model to provide a comprehensive
evaluation of the torsional and bending characteristics of currently available 6-and 8-mm
screwed and bolted ILNs (ILN6s, ILN83, ILN6b, ILN8b, respectively), and to compare
the structural properties of these ILNs to those of an H.Nn and a standard plate (br-DCP).
The study hypotheses were that the compliance of ILNn constructs would be in between
that of 6-mm and 8-mm ILN constructs and smaller than that of 3.5mm br-DCP
constructs, that the angular deformation of screwed constructs would be greater than that

of bolted constructs and that ILNn constructs would sustain the least deformation of all

102

constructs. Another hypothesis was that all standard ILN constructs would have slack,
whereas the ILNn and br-DCP construct deformation would occur without slack.

Forty-eight bone models treated with 6-mm and 8-mm nails locked with screws or
bolts (ILN6s, ILN85, ILN6b, ILN8b, respectively), ILNn, and a 3.5-mm broad-DCP (br-
DCP) (n=4/testing mode) were tested in torsion or 4-point bending. Construct
compliance, deformation and slack were statistically compared.

Results of this study showed that regardless of testing mode, construct compliance
was greater with smaller ILNs. Screwed constructs were more compliant than bolted
ones, with a significant difference between ILN6s and ILN6b in torsion. Plated constructs
were significantly more compliant than the ILNn. Angular deformation was consistently
greater with smaller ILNS. Screwed ILN constructs sustained approximately twice the
torsional deformation of bolted ones. Comparatively, ILNn constructs underwent
significantly less torsional and bending deformation than other constructs. While standard
ILN constructs had slack in both modes, ILNn and br-DCP construct deformations
consistently occurred without slack. This study suggested that while using bolts rather
than screws improved the ILNs mechanical behavior, neither locking mechanism
effectively counteracted torsion and bending forces. Conversely, the ILNn angle-stable
locking system eliminated torsional and bending slack resulting in comparable
mechanical performances between ILNn and plated constructs. Considering the
potentially deleterious effect of uncontrolled motion on bone healing, the ILNn may
represent a biomechanically more effective fixation method than standard ILNs for the
treatment of comminuted diaphyseal fractures, as well as a valid alternative to plate

fixafion.

103

INTRODUCTION

As a result of biological and mechanical advantages, interlocking nails (ILNs) are
considered the standard of care for use in the treatment of most diaphyseal fractures of
the humerus, femur, and tibia in people.l However, as the use of ILNs has increased
clinical limitations have been reported which prompted several experimental studies}8
One in vivo study,2 revealed a significant delay in bone healing and functional recovery
when results for an ILN were compared with those for an external fixator in
experimentally induced tibial fractures using sheep. This delay in bone healing was
attributed to torsional and bending instability. A similar in vitro study9 revealed that
torsional compliance and deformation of tibial constructs implanted with ILNs were
greater than those of constructs implanted with a plate-rod construct (PRC). In that study
ILN constructs under torsional loads had up to 28° of slack, whereas PRC constructs
underwent continuous deformation throughout the testing period. Similarly, an in vitro
study7 of 9 ILNs implanted in human tibiae revealed that there was both torsional and
bending slack in all constructs, regardless of nail design. These studies suggest that
current human and veterinary ILNs do not provide torsional and bending stability as
much as initially anticipated, thus potentially contributing to complications, such as
delayed healing or nonunion of fractures.“0

Contrary to ILNs used in humans, which often are implanted after the medullary
cavity is reamed, ILNs are used in veterinary medicine without reaming of the medullary
cavity. Reaming allows for implantation of larger, stronger nails and increases Contact
between the nails and endocortices, thereby potentially improving stability of the repair.'0

However, reaming severely impairs the medullary blood supply and has been associated

104

with a higher incidence of infection and fat emboli.2‘12"3 The use of thinner ILNs, without
reaming of the medullary cavity, has potential biological advantages, such as preservation

of the endosteal and medullary blood.2‘7"3'l4

The procedure however, places the construct
at a mechanical disadvantage by reducing the nail-bone contact areall and increasing the
working length of the locking device. This suggests that while reaming may be preferable
from a biological standpoint, they risk the integrity of the locking mechanism from a
mechanical standpoint. I 5

Current ILN designs require a mismatch between the screw/bolt and the nail hole
diameters. While this mismatch permits the insertion of the locking device, it also allows
some movement between the nail and the screw or bolt prior to rigid interaction between
these two components of the ILN. Indeed, it has been reported9 that in vitro torsional and
bending instability results from the discrepancy between the screw or bolt diameters and
that of the nail hole (See Chapter 2). Though not quantitatively reported, the same
occurrence is likely responsible for the “play” described in similar in vivo studies.3‘7‘8"O
Furthermore, it has been suggested9 that this discrepancy may be exacerbated by
flattening of the screw threads and by structural damages to the edges of the nail holes
induced by the screw.

To increase stability, ILN constructs have been augmented with various
implants.l0 Although improved stability has been subjectively reported'0 for use of such
techniques, these augmentations are potentially time consuming and invasive, thereby
offsetting the biological advantages of interlocking nails. In addition to providing

adequate stability, implants must be sufficiently strong to withstand loads during the

early postoperative period, particularly when cortical continuity is not achieved or local

105

instability may result in a prolonged healing time.2 The second moment of area, or area
moment of inertia (AMI) of an implant characterizes its ability to resist bending and is
inversely proportional to the stress of a section under bending.I6 Because a larger AMI
results in smaller stress values for a specific load, AMIs have been used to determine the

”"8 and predict the risk of stress fractures.19 The AMIs

theoretical fatigue life of implants
for several implants, including broad dynamic compression plates (br—DCPs) and ILNs
currently used in veterinary medicine have been reported and used as a basis for
comparison of theoretical bending stiffness.”I8

To circumvent the mechanical shortcomings of current nail designs while
preserving the biological benefits from use of smaller nails without reaming of the
medullary cavity, a novel nail and locking system was developed (Loic M. Déjardin, US
patent application No. ll/243,725, Office of Intellectual Property, Michigan State
University). The nail and locking system (ILNn) were created to address several issues.
Firstly, the ILNn was meant to substantially reduce relative motion between the locking
device and nail, thereby effectively locking the construct. Another goal was to increase
the nail-bone contact area in the metaphyseal region while limiting contact between the
nail and endocortices throughout the medullary cavity of the diaphysis. And finally, the
11an was designed to provide bending strength similar to that of existing ILNs, based on
AMI. It was proposed that this design would substantially reduce torsional and bending
instabilities, compared with traditional ILNs of similar size.

The purpose of the current study was then to mechanically compare this novel

hourglass-shaped ILN and locking system to currently available ILN, br-DCP, and ILNn

constructs in torsion and 4-point bending using a canine tibial gap fracture model

106

featuring the previously described synthetic bone substitute (Chapter 2). It was
hypothesized that the bending and torsional compliances of novel ILN constructs would
be significantly less than those of the currently available 6 mm ILN constructs, and that
bending and torsional angular deformation of the novel ILN constructs would be less than
currently available 6 and 8 mm ILN constructs. Because the new nail and locking system
design was meant to eliminate slack, it was also hypothesized that angular deformation of
the novel ILN and br-DCP constructs would be continuous throughout loading, whereas
there would be slack during angular deformation of currently available ILN constructs.
The slack present in currently available ILN constructs is due to a mismatch in screw and

nail hole diameters.

107

MATERIALS AND METHODS
Specimen Preparation .
Bone Model — In an attempt to limit specimen variability and circumvent the need to
procure canine bones, a custom-made synthetic tibial model was used in this study. See

Chapter 2 for a detailed description of this model.

Fixtures — Several fixtures were developed to provide consistent specimens and help
reduce specimen variation. These fixtures guided specimen preparation from drilling pilot
holes in the bone model to loading the specimens into the mechanical testing fixtures. See
Chapter 2 for detailed information regarding these fixtures (drilling fixture, alignment
fixture, foam lathing tool, potting fixture), and Chapter 1 for detailed information on the
bending press (MSU BP), which was slightly modified to accommodate the br-DCPs.
Accommodation of the MSU BP consisted of modifying the fulcrum and contact point to
match the convex and concave aspects of the br-DCP.

In addition to the previously mentioned fixtures, an additional modification was
made to the ILN-ESF alignment fixture mentioned in Chapter 2. Several additional pieces
were designed to allow for consistent and accurate placement of the br-DCPs onto the
bone model. Figure 3-1 shows the additional pieces of the ILN-ESF alignment fixture

used to stabilize the pre-bent (with the MSU bending press) br-DCPs.

108

plate locking
mechanism

   
 

Figure 3-1. Modified ILN-ESF alignment fixture to accommodate the br-DCPs.

The function of the modified alignment fixture was to stabilize the plates so that pilot
holes could be accurately drilled into the bone model for consistent plate placement.
Movement in all directions had to be restricted to stabilize the plates. This was

accomplished by a specially designed locking mechanism (Figure 3-2).

 

Figure 3-2. The plate locking mechanism that restricted movement in all directions.

The plate locking mechanism restricted front to back movement (shown in blue) of the
plate by the shape of the cutout being the same size as the width of the plate. Left to right
movement (shown in red) was restricted by two pins, which passed through holes in the
plate and inserted into the base of the alignment fixture. Vertical movement (shown in
green) of the plate was restricted by the bone model on the bottom and the cutout surface

on the top.

109

Design of Novel ILN - Dogs affected with long bone fractures seen at Michigan State
University Small Animal Clinic (MSU SMAC) commonly are midsize dogs weighing
approx 30 to 35 kg. This patient population is often treated for comminuted fractures of
the tibial diaphysis with a 6 mm or an 8 mm ILN (ILN6 and ILN8, respectively). Based
on morphometric analyses of canine tibiae, a 185 mm long ILN is most often appropriate
to fit the tibial diaphysis of dogs within that weight range. Therefore, prototypes of a
novel ILN system were designed to be comparable to the commercially available 6 or 8 X
185-mm ILN (models 11-06-185-02-32.7 and ll-08-l85-02-3.5, respectively, Innovative
Animal Products, Rochester, MN).

Prototypes of the locking device and novel nail were designed with several
primary constraints. The locking device had to provide rigid interaction with the nail. The
shape of the nail had to limit interference with endocortices and facilitate fracture
reduction, while limiting the risk of joint infraction attributable to perforation of the distal
subchondral bone plate. Finally, the shape and size of the locking device had to enhance
successful insertion during surgery.

Secondary constraints were the stiffness and strength of the locking device and
nail. The AMI of the novel locking device had to be greater than that of a 4.5-mm bone
screw and comparable to that of a commercially available 3.5-mm locking bolt. The AMI
of the novel nail at the level of a locking hole had to be similar to that of an ILN8 with
3.5-mm bone screws in both the mediolateral and craniocaudal planes. Finally, the AMI
of the weakest part of the solid central section of the novel nail had to be similar to that of

the solid section of a 3.5-mm br-DCP.

110

To address the above constraints, an hourglass-shaped nail was designed. It
featured an oblong bullet-like distal tip and a novel screw-cone-peg (SCP) locking
device. The nail was fabricated with 316L stainless steel certified to ASTM F138
standards (Hourglass ILN/SCP, 20-TB08-TB-0A, BioMedtrix, Boonton, NJ [Figure 3-
3]). This material was similar in chemical composition and mechanical properties to that
used to manufacture the ILN8 used in the study (316L ASTM F139). Stainless-steel 316L
ASTM F 138 was chosen for the manufacture of the SCP locking device and the ILN
components. This was largely because it is the material of choice for medical
manufacturers of such implants. The outside diameter of the nail ends was 8 mm,
whereas the central portion featured a reverse entasis that reduced the middle portion of
the nail to an outside diameter of 6 mm. Two tapered holes (diameter, 4.0 and 3.2 mm,
respectively) were placed 11 mm apart in each end of the nail. The most proximal and
distal nail holes were separated by 155.5 mm. The SCP was designed as a self-tapping,
cortical type screw (core diameter, 4 mm) with a central Morse taper that matched the
nail hole and a solid distal tip (outside diameter, 3.2 mm). Prototype SCPs were

manufactured in lengths of 26 and 30 mm.

111

 

 

 

 

Figure 3-3. A) Photograph of ILNn. B) Close-up view of the distal end of the novel ILN. C)
Close-up view of the SCP.

Measurement of AMI - AMIs were calculated using basic strength of materials theory,

namely Equation 3-l, corresponding to Figure 3-4.

Ix = l ysz Iy = l XZdA [3.1]

Equation 3-1. Area moment of inertia (AMI), where:

Ix: AMI with respect to x-axis (mediolateral AMI)

Iy: AMI with respect to y-axis (craniocaudal AMI)

y, x: equation of the line forming the boundary of the cross section

A: area of the cross section

112

nail hole

 

Figure 3-4. Cross-section of a nail with corresponding x and y-axes, shown at nail hole level.

AMI values were calculated for various sections of the SCP (screw, cone and peg) and

the ILNs (nail hole level, and for the central section) (Figure 3-5).

 

nail hole level

 

 

 

v

Figure 3-5. Schematic of various sections of SCP and [LN

AMIs were calculated with a specially designed computational program20 which
produced automated results based on inputs of ILN diameter (central section and nail

hole) and screw size. See Appendix 3 for detailed Matlab programming.

Study Design
Implants were applied to the synthetic tibiae in 6 groups for a total of 8 specimens
per group (n=4/testing mode; torsion, bending). Group 1 (ILN6s) was made up of 6 mm x

185 mm ILN with four 2.7 mm screws. Group 2 (ILN6b) was made up of 6 mm x 185

113

mm ILN with four 2.7 mm bolts. Group 3 (ILNBS) was made up of 8 mm x 185 mm ILN
with four 3.5 mm screws. Group 4 (ILN8B) was made up of 8 mm x 185 mm ILN with
four 3.5 mm bolts. Group 5 (br-DCP) was made up of broad dynamic compression plates
with eight 3.5 mm screws (four on each bone segment). Group 6 (ILNn) was made up of
the 185 mm novel interlocking nail and four 4/3.2 mm SCPs. All standard ILNs (ILN6s,
ILN6b, ILN8S, ILN8B), plates (br-DCP), and locking bolts/screws used in this study
were obtained from Innovative Animal Products, Rochester, MN. ILNn and SCPs used in

this study were obtained from BioMedtrix, Boonton, NJ.

Mechanical Testing

Using custom designed loading fixtures; all specimens were mounted in an
Instron servo-hydraulic testing machine (Instron model 1331, Instron Corp., Canton, MA)
coupled to a 2500 lb (11,120 N) axial load cell (Model 1010AF-2.5K-B, Interface Inc.,
Scottsdale, AZ). The torsion and bending loading fixtures were instrumented with rotary
encoders (torsion encoder- Bez BHW 16.05A72000, Baumer Electric, Southington, CT;
bending encoders-RHSZOD-SOOO—l/2-5/LD-l-M4, Renco Encoders, Goleta, CA) to
document angular deformation of the constructs. In addition, the torsion fixture was
instrumented with a 1200 lb*in (135 N*m) torque load cell (Model 5330-1200, Interface
Inc., Scottsdale, A2) to further document the magnitude of the applied torques. All
specimens were tested non-destructively in torsion and 4-point bending (n=4 per testing
protocol). All tests were run in load control for 10 cycles. Actuator displacement, rotary
encoder output, and corresponding load/torque were documented in the tenth cycle. See

Appendix 2 for detailed operating procedures for the torsion and bending tests.

114

Torsion Tests — Torsion tests were run using a 0.125 Hz sinusoidal waveform at a torque
level of t 5 N*m. The torque level chosen for the current study was identical to that used
in previous biomechanical investigations.9'2| The torsion fixture was equipped with a
rotary encoder to record angular deformation and a torque load cell (Figure 3-6). See

Chapter 2 for detailed information on torsion tests.

torque rack (pinion 9
load cell not vmhle) rotary
encoder

n

I c
\ fl '
a

-'

distal torsion cup proximal torsion
(remains static) cup (rotates)

 

Figure 3-6. The torsion fixture shown with a loaded ILNn.

Bending Tests — Bending tests were run using a 0.125 H2 sinusoidal waveform at a
bending moment of -1_- 3.5 N *m. The bending moment chosen for this study was identical
to that used in previous biomechanical evaluations of fixation devices in a gap fracture
model?“2| A specially designed fixture allowed for application of a pure bending moment
over the entire bone model/implant construct. See Chapter 2 for detailed information on
bending tests. Slight modifications were made to the bending cups and loading arm of the
bending fixture to improve both performance and ease of use. In previous studies
(Chapter 2), the loading arm had been connected to the bending cups via a steel shaft that

was rigidly attached to the cups. The steel shaft was difficult to use because if there was

115

any out-of—plane misalignment between the cups and the linkage arms, they could not be
connected. Another effect of misalignment when using the steel shafts was that binding
would sometimes occur at various points throughout the loading cycle. The addition of
stainless steel balls to the attachment point (Figure 3-7) between the loading arm and

loading cups allowed for ease of connection between tests (Figure 3-8).

 

Figure 3-7. Stainless steel ball for connection between the cups and the loading arm.

   
    
    
 

    

rl/ -
$-

steel ball attached 2
to shaft "

  

Figure 3-8. Linkage arm shown attached to the bending cups.

116

In addition to modification of the connection points from the loading arm to the bending

cups, the loading arm itself was also modified (Figure 3-9).

if-
/

redesigned main bar

redesigned
linkage arms

 

Figure 3-9. The modified loading arm.

Improvements in the loading arm included reducing the overall weight of the arm and
minimizing friction between connection points. The loading arm was remade out of
aluminum and cutouts were made to reduce its weight. The loading arm was also
equipped with bearings to reduce friction at the connection between the loading arm and

the actuator (Figure 3-10).

 

Figure 3-10. Loading arm bearings.

Another modification was made at the point of connection between the main bar and the
linkage arms. Instead of a steel shaft, which allowed rotation in one plane, the linkage

arm was now attached to the main bar with a partial ball joint (Figure 3- 1 l-A). This joint

117

allowed full rotation in one plane, as with the shaft, and also slight out-of-plane
movement. The out-of-plane movement was restricted by the spacing between the two
cross bars which make up the main bar (Figure 3-11-B). Figure 3-12 shows the fully

assembled bending fixture.

 

Figure 3-11. A) Partial ball joint for connection between main bar and linkage arms (the main bar
is shown disassembled so the ball is visible). B) Linkage arm shown attached to main bar
(spacing shown in black).

    
    

’;
. 4:1»...
rotary
encoders

 

bending
cup

   
 

Figure 3-12. The bending fixture shown with a loaded br-DCP.

118

Data Acquisition

A load cell and/or a torque load cell coupled to the Instron actuator and to the
torsion fixture, respectively, recorded applied loads (N) and torques (N*m) over time.
These and actuator displacement (mm) were recorded at a sampling rate of 250 Hz. Since
all constructs were tested under load control, construct compliance (slope of deformation
versus load curve) was evaluated in the 10th cycle. Construct compliance in bending and
torsion was determined as the mean compliances during positive and negative loading.
Construct angular deformation (torsion and bending) was recorded by rotary encoders in
addition to redundant measurements based on the geometrical dimensions of the torsion
and bending fixtures. Angular deformation of constructs in bending was calculated as
described in Chapter 2. See Appendix 2 for detailed static analysis and calculations for
torsion and bending tests.

The results of compliance and associated deformation between constructs were
compared using a one factor ANOVA. Student-Newman-Keuls post-hoc tests were used

whenever significant differences between groups were indicated (p < 0.05).

119

RESULTS
Values for AMI

The AMI of the 2.7 mm screw, 2.7 mm bolt, 3.5 mm screw, 3.5 mm bolt, 4.5 mm
screw, and SCP was calculated. The AMI of 2.7 mm screws was approximately 25% of
the 2.7 mm bolt (0.639 mm4 and 2.607 mm4). The AMI of various sections of the SCP
(threaded section, 12.57 mm4 [core, 4.0 mm]; tapered midsection, 8.24 mm4 [core, 3.6
mm]; and smooth solid section, 5.15 mm4 [core, 3.2 mm]) were always larger than those
of the 3.5—mm (1.63 mm4) or 4.5-mm (3.98 mm4) screws (core diameter of the 3.5 and
4.5 mm screws, 2.4 and 3.0 mm, respectively“). In addition, the AMI for the threaded
section of the SCP was 171% greater than that of a 3.5-mm bolt (7.36 mm4). Whereas the
AMI of the smooth section of the SCP was 31.6% and 129% greater than that of a 3.5 or
4.5 mm screw, respectively, it was 30% less than that of a 3.5 mm bolt.

Values of AMIs for the ILN6 (with 2.7 mm nail hole), ILN8 (with 3.5 and 4.5 mm
nail holes), ILNn and 3.5-mm br-DCP were calculated (Table 3-1). The AMI values for
the novel ILN at the nail holes were similar to those of an ILN 8 with 3.5 mm screws and
larger than those of an ILN8 with 4.5 mm screws in both the mediolateral and
craniocaudal planes. The central section of the novel ILN had an AMI greater than that of

a 3.5 mm br-DCP.

120

 

 

 

 

 

 

 

 

 

 

 

seam" (2.711111: irate) (3.511114: liole) ILN“ (4.511;: inure) br'DC'"
S°“° °.°“"°‘ 63.6 201.06 63.62 201.06 ~ 59
SCCIIOH
Nail hole
Mediolaterall 19.8 65.6 | 62.1 37.94 ~ 32
Craniocaudall 54.4 174.2 I 171.4 146.5 ~ 500

 

 

 

 

 

Table 3-1. Nail and plate AMI values (mm‘).

* Values determined on the basis of a plate cross-section of
approximately 3.9 X 1 1.95 mm and a hole cross-section

of approximately 3.9 X 5.55 mm, assuming rectangular
cross sections for the plate and hole.

Performance of Construct Development Fixtures

The goal of developing and using a synthetic bone model was two-fold. One
reason was to eliminate the need of procuring specimens, which can require significant
time, effort, and expense. Experimentally though, the elimination of biological specimens
would potentially provide data which had much less deviation from specimen to
specimen. For this to occur however, a reliable, efficient, and consistent method of
preparing the bone-implant constructs had to be developed. With this in mind, several
fixtures were specially designed, for preparation of the bone for implantation to mounting
the construct in the loading fixture. Theoretically, the goal was to eliminate the
possibility of human variation while preparing the specimens for mechanical testing.

The modification of the ILN-ESF alignment fixture was done to produce
consistent br-DCP specimens by drilling pilot holes at the same location on the bone
models from specimen to specimen. The plate locking mechanism ensured the plate’s

position with respect to the bone. Drilling of pilot holes was then carried out by locating

121

appropriate position with respect to the standardized plate hole locations; Figure 3-13

shows a sample bone model with pilot holes drilled using the alignment fixture.

 
      

t
locking
‘1 “ mechanism

9
"-
I J

J1...

Figure 3-13. Pilot holes (white) drilled in the bone model with respect to the plate hole locations.

The ILN alignment fixture was used to create consistent ILN constructs. Figure 3-14

shows the consistency in overall length and osteotomy length between groups.

 

Figure 3-14. All constructs from left to right: ILN6s (yellow), ILN6b (orange), ILN8S (green),
ILNBB (red), br-DCP, and ILNn (blue).

122

Torsion Tests
Construct Compliance — Construct compliance curves for the ILN6 (ILN6s and ILN6b)
and ILN8 (ILNSS and ILNSB) constructs were bimodal, whereas compliance curves for

the br-DCP and ILNn constructs were unimodal (Figure 3—15).

 

 

   
 

 

 

 

 

 

 

Torsion

8

6 -
a
2 4
U!
0
3 2
1:
3% o L/1
g .6 Slack l 6
'5 '2 '_JL Angular
o .
.. Deformation
S -4 _
3
U!
C
< -6 -

-8

 

Torque (N'm)

Figure 3-15. Typical compliance curves; note the bimodal response (red-ILNSB) which was
characteristic of the 6 and 8 mm ILNs, and the unimodal response (blue-ILNn) that characterized
the response of the br-DCP and ILNn constructs.

In the bimodal curves for the 6 and 8 mm ILN constructs, there was no
quantifiable torque in the central region, which corresponds to the change in direction of
torque. This region represented the slack in the construct and reflected an abrupt change
in angular deformation without resistance to applied torque. Conversely, the unimodal
shape of the compliance curves for br-DCP and ILNn constructs reflects a lack of slack,
and a continuous resistance to deformation throughout and during reversals in construct
loading. Construct compliances were calculated as the slope of the deformation versus

torque curve in the range of 2.5-5 N*m. Since there was not a significant difference

123

between the positive and negative compliances within any of the groups (p > 0.05)
compliance was calculated as the mean of the positive and negative loading curves. The
ILNn constructs were significantly less compliant than that of the ILN6s and ILN6b
constructs ([p < 0.05] - 0.87 i 0.02 °/N*m, 1.65 3: 0.05 0/N*m, and 1.12 :1: 0.04 °/N*m,
respectively [Figure 3—16]). Construct compliance for the ILNn constructs was greater

than that of both 8 mm ILN constructs (Table 3-2). See Appendix 4 for all pairwise

 

  
 
 
  

comparisons.
Torsion Compliance 1:1 ILN6s
1'8 I ILN6b
1.6 I ILNBS

I ILNaB
[:1 br-DCP
I ILNn

Compliance (degrees/Mm)
O
on

 

Figure 3-16. Construct compliances in torsion.

Construct Angular Deformation — Maximum angular deformation of the IIan
constructs was significantly less than all other constructs ([p < 0.05] Appendix 4 [Figure
3-17]). Construct slack, where present, was computed as the difference between the y-
axis intercept of the compliance slope on the positive and negative loading curves (Figure
3-15). There was also no slack present in the ILN and br-DCP groups. Within groups
where slack was present, the ILN6s construct had the most, and the ILN8B construct had

the least (19.79 :1: 24° and 5.6 :1: 014°, respectively). Once slack was overcome (at high

124

 

torques), angular deformation for the ILNn constructs, was less than that for the 6 mm

ILN constructs and greater than that for the 8 mm ILN constructs (Table 3-2). See

Appendix 4 for all pairwise comparisons.

 

 

 

 

 

 

 

 

 

 

Torsion Angular Deformation 1:11LN6s
45.0 I ILN6b
A 40.0 I 1LN88
1'1
2 35.0
at
o
33, 30.0
1:
.9 25.0
H
a
E 20.0
.2
a 15.0
in
2 10.0
:1
g, 5 0
< .
0.0
Figure 3-17. Construct angular deformation in torsion.

Construct Compliance (°/N*m) Slack (°) Angular Deformation (°) AD — Slack (°)
ILN6s 1.65 1 0.05 19.79 1 2.1 36.44 1 2.68 16.66 1 0.58
ILN6b 1.12 1 0.04 6.85 1 0.44 18.11 1 0.11 11.26 1 0.35
ILN8S 0.76 10.04 16.31 1 0.75 23.93 1 0.39 7.61 1 0.42
ILN8B 0.68101 5,610.14 12.41 1 1.14 6.811142
br-DCP 1.28 1 0.07 0 11.93 1 0.51 11.93 1 0.51

ILNn 0.868 1 0.02 0 8.29 1 0.22 8.29 10.22

 

 

 

 

 

Table 3-2. Torsion data (mean 1 standard deviation).

125

 

Bending Tests
Construct Compliance — Similar to torsion tests, bending tests compliance curves for the
ILN6 (ILN6s and ILN6b) and ILN8 (ILN8S and ILN8B) constructs were bimodal,

whereas compliance curves for the br-DCP and ILNn were unimodal (Figure 3-18).

Bending

 

   

Angular

 
 

Deformation

l

Angular Deformation (degrees)
0)

Moment (N*m)

Figure 3-18. Typical compliance curves; note the bimodal response (red-ILN8B) which was
characteristic of the 6 and 8 mm ILNs, and the unimodal response (blue-ILNn) that characterized
the response of the br-DCP and ILNn constructs.

In the bimodal curves for the 6 and 8 mm ILN constructs, there was a region
where the construct did not resist the applied moment. This region was defined as the
slack in the construct. Due to the unimodal appearance of the compliance curves for the
br-DCP and ILNn constructs, there was a continuous resistance to deformation
throughout loading and reversal of loading. Construct compliances were computed as the
slope of the deformation versus moment curve in the range 1.5-3.5 N*m. Since there was
not a significant difference between the positive and negative compliances in any of the

groups (p > 0.05), compliance was computed as the mean of the positive and negative

126

loading curves. In bending tests, ILNn constructs were significantly less compliant than 6
mm bolted constructs, however, ILNn and ILN6s constructs were not significantly
different (Figure 3-19)(Table 3-3). ILNn constructs were significantly more compliant

than that of the H.N8S and ILN8b constructs. See Appendix 4 for all pairwise

 

 
 
 
  

 

comparisons.
Bending Compliance 1: 'LNSS
1.0 I1LN6b
IILNBS
08 IILN8B .
E ' Bbr-DCP
§ 6 I ILNn
O.
i
CI
0
3
g 0.4
O
C
.5
g- 0.2
O
o
0.0

Figure 3-19. Construct compliance in bending.

Construct Angular Deformation — Maximum angular deformation of the [1an
constructs was significantly less than all 6 and 8 mm ILN constructs in bending ([p <
0.05] Figure 3-20 [Table 3-3 ]). There was also no slack present in the ILN and br-DCP
groups in either mode of testing. The constructs locked with screws (ILN6s and ILNSS)
had more slack (9.79 1 121° and 9.03 1 106°, respectively), than the constructs locked
with bolts (6.81 1 04° for ILN6b and 4.32 1 058° for ILN8B). Once slack was overcome
(at high moments), angular deformation for the ILNn constructs was less than that of the
6 mm ILN constructs and greater than that of the 8 mm ILN constructs (Table 3-3). See

Appendix 4 for all pairwise comparisons.

127

Bending Angular Deformation

El ILNGS

 

  
 
 
  

 

 

 

 

 

 

 

 

 

15.0 I lLN6b
A I |LN8S
to I lLNBB
§ 12.0 1
a) 13 br-DCP
i 1| ILNn
g 9.0
a
111
E
e 6.0
o
D
13
3 3.0
U)
c
<

0.0
Figure 3-20. Construct angular deformation in bending.

Construct Compliance (°/N*m) Slack (°) Angular Deformation (°) AD — Slack (°)
ILN6s 0.58 1.03 9.79 1 1.21 13.87 1 1.18 4.08 1 0.24
ILN6b 0.71 1 .08 6.81 1 0.4 11.77 1 0.8 4.96 1 0.56
ILN8S 0.43 1.04 9.03 1 1.06 12 1 0.79 2.98 1 0.33
ILN8B 0.4 1.03 4.32 1 0.58 7.07 1 0.72 2.75 1 0.15
br—DCP 0.92 1 .02 0 6.15 1 0.07 6.15 1 0.07
ILNn 0.6 1 .05 0 4.03 1 0.25 4.03 1 0.025

 

 

 

 

 

Table 3-3. Bending data (mean 1 stande deviation).

128

 

DISCUSSION

This study compared the mechanical properties of a novel interlocking nail
designed for the treatment of diaphyseal fractures in mid-size dogs (body weight of
approximately 30 to 35 kg) to those of standard 6-mm and 8-mm ILNs available in the
USA. It was hypothesized that the bending and torsional compliances of novel ILN
constructs would be significantly less than those of the currently available 6 mm ILN I
constructs, and that bending and torsional angular deformation of the novel ILN
constructs would be less than currently available 6 and 8 mm ILN constructs. It was also
hypothesized that angular deformation of the novel ILN and br-DCP constructs would be
continuous throughout loading, whereas there would be slack during angular deformation
of currently available ILN constructs. The results of the torsion and bending tests in the
present study support the above hypotheses. Specifically, the compliance of the ILNn
construct was significantly less than all other constructs tested in this study. Furthermore
results of the torsion and bending tests on the ILNn constructs indicated a continuous
resistance to deformation throughout loading and reversal of loading directions, whereas
slack was present during reversal of loading in currently available ILN constructs. On the
basis that there was no slack for the novel ILN in torsion or bending, it was determined
that the 11an described here had better torsional and bending stability, compared with
that for the currently available ILNs. What is also important to consider, is that the
loading curves of the ILNn constructs were similar to the br-DCP construct (no slack,
constant resistance to applied load). This suggested that the novel ILN design may

improve the biomechanical environment for fracture healing.

129

Clinical studies have revealed intraoperative or postoperative instability of
interlocking nails.“7 In these studies, 12% to 14% of the animals have delayed healing or
require supplementation of the initial ILN repair to provide adequate stability. Similarly,
delayed union rates as high as 18% have been reported in humans treated by use of tibial
nails without reaming of the medullary cavity.22 Although the optimal mechanical
environment favorable to bone healing remains controversial, it is generally accepted that
excessive interfragmentary motion delays bone healing, in contrast to studies where

(123.25 In particular, the effect of shear motion on

results for axial micromotion is controlle
fracture healing continues to be debatable. Numerous studies in dogs and sheep have
indicated deleterious effects of torsional and shear motions on early bone healing and

2"5‘26‘37 Another study,28 on rats has revealed that the callus 2 and 4

functional recovery.
weeks after surgery was significantly larger in tibial fractures subjected to local shear
strains than those with rigid stabilization of the fracture.

Results of such studies, supported by experimental evidence of construct
instability with current ILNs,7‘9 emphasize the need for a more effective ILN design. In an
attempt to investigate potential correlations between construct stability and functional
recovery, an in vitro and in vivo study was conducted to evaluate an experimental nail
featuring bolts threaded into the nail holes using a tibial gap fracture model in sheep.29
The study revealed that although experimental and conventional nails had similar
stiffnesses in mediolateral bending and axial compression, the experimental nail was
significantly stiffer in shear and craniocaudal bending. Additionally, the experimental

nail had a slightly higher stiffness in torsion. As a result, the experimental nail group had

significantly smaller interfragmentary motion throughout the 9 weeks of the experiment.

130

This was suspected to explain the resultant superior bone healing (as evaluated by use of
histomorphometry, radiography, and ex vivo biomechanical testing) and faster, more
complete functional recovery of the patient (as evaluated by use of gait analysis).29

To increase construct stability by eliminating motion at the interface between the
nail and locking device the SCP was designed with a central Morse taper. A Morse taper
features a matching trunnion (male component) and bore (female component) and is
commonly used to join modular components during total hip arthroplasty.30 The tapered
design also facilitates the insertion of the SCP by providing a self-centering feature, even
when there is slight misalignment between the cortical pilot hole and nail hole. This may
help reduce the incidence of missed locking devices due to misaligned holes during
treatment of fractures with ILNs.

The strength and failure pattern of ILNs depend in large part on the design and

.3 .32 - -
6 ' In comrmnuted fractures, implants are

size of the nail hole and locking device.
mainly. subjected to bending stresses.l8 Therefore, the AMIs used in the study reported
here were calculated with respect to theoretical bending conditions. The AMI of an
implant is a structural property that characterizes the geometric distribution of a material
with respect to the axis of loading. Based on strength of materials bending theory, an
implant with a larger AMI will sustain lower stresses during cyclic loading, which in turn
will extend its fatigue life.'7 One limitation of early ILNs designed for veterinary use was
the weakness of the nail holes, which, because of the sharp local decrease in AMI, acted
as stress concentrators that led to nail failure.'7'33‘34 On the basis of the AMIs, ILNs are

weakest at the nail hole in mediolateral bending. Reducing the diameter of the screw hole

from 4.5 mm (early 8 mm ILN design) to 3.5 mm (current 8 mm ILN design) results in a

131

5.7-fold increase in local AMI, which translates into an 8-fold increase in fatigue life of
an 8 mm nail.33 This design change improves structural properties of a nail to the
detriment of a thinner or weaker screw as indicated by approximately a one third decrease
in AMI. 3“

Because of the failure patterns of 6-mm nails and screws and increased strength of
8 mm ILNs, the novel ILN reported here was devised so that the AMI of the locking
device was equal to or greater than that of the 4.5 mm screw and the AMI of the nail at
the holes was similar to that of the ILN8 with 3.5 mm screws. Design of the SCP was
such that its smallest AMI (ie, AMI for the peg section) had to be greater than that of a
4.5 mm bone screw; AMI for the peg section was 5.15 mm4, whereas AMI for the 4.5
mm bone screw was 3.98 mm4, which represented an increase of approximately 30%. At
the same time, the smallest AMI for the SCP could be smaller than that of a 3.5 mm bolt
(7.36 mm4). This design constraint was selected on the basis of the fact that failure of 4.5
mm screws has not been reported for the first generation of 8 mm ILNs. Furthermore,
although the corresponding AMI of the novel ILN (62.1 mm4) was slightly smaller than
that of the ILN8 with a 3.5 mm screw (65.6 mm4), it is 64% greater than that of an ILN8
with a 4.5 mm screw (37.94 mm4), thus suggesting that the noVel ILN should have an
estimated fatigue life similar to that of the currently available 8 mm 11st with 3.5 mm
screws. Moreover, the improved stability of the novel ILN (no slack) could potentially
result in shorter healing time, thereby rendering the slight theoretical decrease in fatigue
life of the novel ILN clinically irrelevant.

The hourglass design of the nail provides several potential benefits. First, this

design should contribute to preserving the endosteum and improve restoration of the

132

medullary blood supply after implantation, 2 that, in turn, may enhance the rate of bone

13.14

healing. This shape should also facilitate implantation of the nail in curvilinear bones
by limiting contact between the nail and endocortices. The use of a relatively large nail is
often guided by the necessity to have a strong locking mechanism.33 Consequently, it is
recommended that the largest possible nail be used, principally on the basis that the
fatigue life of an 8 mm nail is 10 times that of a 6 mm nail.33 Alternatively, a comparable
fracture could be treated successfully by use of a 3.5 mm br-DCP.35 Because the AMI of
the solid section of a 6-mm nail (approx 64 mm4) is similar to that of a 3.5-mm br-DCP
(approximately 59 mm4), it can be argued that the strength of an ILN 8, which has an AMI
for the solid section of 201 mm4, is not warranted. Furthermore, the use of larger
medullary implants can substantially impede the cortical blood supply, which in turn
could negatively affect bone healing.2

Finally, the bullet-shaped distal tip of the nail was designed to facilitate reduction
of the fracture, particularly with regard to restoring length of the bone without increasing
the risk of penetration of the distal joint associated with the use of trocar points. The
oblong tip of the novel ILN could facilitate insertion through the proximal metaphysis
and permit deep anchorage in the distal metaphysis, compared with results for the flat
truncated tip in currently available ILNs. This may allow for treatment of a greater
variety of fractures including those affecting the metaphyses.

Results of the torsion and bending tests indicate that the compliance of the [1an
is less than the 6 mm constructs but greater than the 8 mm constructs. In comparison to

the 6 mm ILNs, the ILNn has the same AMI in the central section of the nail (6 mm

diameter); therefore the nails should have the same ability to resist deformation.

133

However, the load applied to the bone is essentially transferred to the nail through
bending of the locking mechanism.

In the current study, the torsional and bending forces on the nail result in applied
bending and shearing forces to the locking mechanism. Therefore, given an equal
diameter of the solid central section of the nails (ILN6s, and ILN6b, and ILNn),
differences in compliance are only attributable to mechanical properties of their
respective locking mechanisms. This is shown under torsion, where the ILN6s, ILN6b
and ILNn constructs decreased in compliance (approximately 1.65, 1.12, and 0.87 °/N*m,
respectively) as the AMIs of their corresponding locking mechanisms increased. It should
be noted however, that the torsional compliance of an implant is inversely proportional to
its polar moment of inertia (approximately 1.28 mm4, 5.22 mm4, and 16.48 mm4,
respectively) as opposed to its AMI (0.64 mm4, 2.61 mm4, and 8.24 mm4), which is
inversely proportional to the bending compliance of an implant. The larger torsional and
bending compliance of the novel ILN construct, compared with that of the 8 mm ILN
constructs (ILN8S and ILN8B), was attributed to the difference in the core diameter of
the solid section of the nails (6 and 8 mm for the novel ILN and ILN8 constructs,
respectively), which resulted in a > 3-fold increase in AMI. Interestingly, the torsional
compliance of the ILNn in the current study was very similar to the compliance of a 6
mm ILN locked with an external fixator, as tested in Chapter 2 (both are approximately
0.87 °/N*m).

Primarily because of slack in the 6 and 8 mm ILN systems (ILN6s, ILN6b,
ILN8S, and ILN8B) angular deformation of the currently available ILN constructs in

bending and torsion was greater than that of the novel ILN construct. Slack, which

134

corresponds to a lack of resistance to applied load, has been associated with looseness
between the locking mechanism and nail.9 When slack was removed, at high torsional
and bending loads, differences in angular deformation decreased drastically. Consider the
comparison of the overall angular deformation of the ILN6s construct to that of the ILNn
construct in bending. The ILN6s constructs deformed approximately 350% more than the
ILNn constructs, including deformation incurred due to slack in the system. After slack
was removed, the angular deformation of the ILN6s construct (approx. 4.075°) was no
longer significantly different than the ILNn construct (approx. 4.025° [p = 0821]).
Construct slack results in uncontrolled, acute motion at the fracture site, which in turn
likely generates high local strains. Since shear strains can be detrimental to bone healing,
36-38

slack in torsion or bending or both could explain some of the cases of delayed and

non-union seen clinically after repair of fractures with standard ILNs.

135

CONCLUSION

The study reported here suggested that re—engineering of the locking mechanism
in a novel hourglass-shaped ILN has eliminated the instability associated with the use of
current ILNs inserted, without reaming of the medullary cavity. Furthermore, the
improved torsional and bending stability of the construct matched that of a comparable
br-DCP construct, a device routinely used to treat comminuted diaphyseal fractures. In
contrast to plate osteosynthesis, an ILN can be applied at remote fracture sites and be
used to reduce and stabilize fractures by use of a closed technique. This less invasive
approach to fracture repair improves early bone healing.39 Because of the potential
combined mechanical and biological benefits of this novel hourglass shape, the novel
ILN could represent an effective and safe alternative to plate osteosynthesis while
preserving the advantages of nails that do not require intramedullary reaming.39
Additional biomechanical studies to evaluate fatigue strength of the novel ILN are
warranted. In vivo studies of the novel ILN should also be conducted to fully assess the

potential use of the new hourglass-shaped ILN.

136

_
0

REFERENCES

Kyle RF, Schaflhausen J M, Bechtold JE. Biomechanical characteristics of

interlocking femoral nails in the treatment of complex femoral fractures. Clin
Orthop Relat Res 1991: 169-173

Klein P, Opitz M, Schell H, et a1: Comparison of unreamed nailing and external
fixation of tibial diastases--mechanical conditions during healing and biological
outcome. J Orthop Res 22: 1072-1078, 2004

Duhautois B: Use of veterinary interlocking nails for diaphyseal fractures in dogs and
cats: 121 cases. Vet Surg 32: 8-20, 2003

Knothe U, Knothe Tate ML, Klaue K, et a1. Development and testing of a new self-
locking intramedullary nail system: testing of handling aspects and mechanical
properties. Injury 2000; 31: 617-626

Laflamme GY, Heimlich D, Stephen D, et a1. Proximal tibial fracture stability with
intramedullary nail fixation using oblique interlocking screws. J Orthop Trauma
2003; 17:496-502

Lin J, Lin SJ, Chen PQ, et a1. Stress analysis of the distal locking screws for femoral
interlocking nailing. J Orthop Res 2001; 19: 57-63

Schandelmaier P, Krettek C, Tscherne H. Biomechanical study of nine different tibia
locking nails. J Orthop Trauma 1996; 10: 37-44

Suber JT, Basinger RR, Keller WG. Two unreported modes of interlocking nail
failure: breakout and screw bending. VCOT 2002; 15: 228-232

von Pfeil DJ, Déjardin LM, DeCamp CE, et al: In vitro biomechanical comparison of
a plate-rod combination-construct and an interlocking nail-construct for
experimentally induced gap fractures in canine tibiae. Am J Vet Res 66: 1536-
1543, 2005

10. Basinger RR, Suber JT: Two techniques for supplementing interlocking nail repair of

fractures of the humerus, femur, and tibia: results in 12 dogs and cats. Vet Surg
33: 673-680, 2004

1 1. Mayr E. [Tibial fractures]. Chirurg 2002; 73: 642-661

12. Heim D, Schlegel U, Perren SM: Intramedullary pressure in reamed and unreamed

nailing of the femur and tibia--an in vitro study in intact, human bones. Injury 24
Suppl 3: $56-63, 1993

137

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

Krettek C. [Principles of intramedullary fracture stabilization. 1]. Unfallchirurg
2001; 104: 639-651

Runkel M, Wenda K, Stelzig A, et a1: Bone remodeling after reamed and unreamed
intramedullary nailing: A histomorphometric study. Unfallchirurg 97:385-390,
1994

Duda GN, Mandruzzato F, Heller M, et a1. Mechanical boundary conditions of
fracture healing: borderline indications in the treatment of unreamed tibial nailing.

J Biomech 2001; 34: 639-650

Hibbeler RC. Mechanics of Materials. 3rd ed. Upper Sadie River: Pearson Prentice
Hall, 2004

Dueland RT, Vanderby R, Jr., McCabe RP. Fatigue study of six and eight mm
diameter interlocking nails with screw holes of variable size and number. Vet
Comp Orthop Traumatol 1997; 10: 194-199

Muir P, Johnson KA, Markel MD. Area moment of inertia for comparison of implant
cross-sectional geometry and bending stiffness. VCOT 1995; 8: 146-152

Milgrom C, Giladi M, Simkin A, et al. The area moment of inertia of the tibia: a risk
factor for stress fractures. J Biomech 1989; 22: 1243-1248

MATLAB and Simulink. Matlab Version 7.0.4. The MathWorks, Inc. Natick, MA.
2003

Bouvy BM, Markel MD, Chelikani S, et a1. Ex vivo biomechanics of Kirschner-
Ehmer external skeletal fixation applied to canine tibiae. Vet Surg 22: 194-207,
1993

Goldhahn S, Moser R, Bigler R, et al. [Treatment methods and outcomes of tibial
shaft fractures in Switzerland. A prospective multicenter study of the Swiss AO].
Swiss Surg 2000; 6: 315-322

Goodship AE, Kenwright J. The influence of induced micromovement upon the
healing of experimental tibial fractures. J Bone Joint Surg Br 1985; 67: 650-655

Lewallen DG, Chao EY, Kasman RA, et a1. Comparison of the effects of
compression plates and external fixators on early bone-healing. J Bone Joint Surg
Am 66: 1084-1091, 1984

Wu J], Shyr HS, Chao EY, et a1. Comparison of osteotomy heling under external
fixation devices with different stiffness characteristics. J Bone Joint Surg Am 66:
1258-1264, 1984

138

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

Aron DN, Palmer RH, Johnson AL. Biologic strategies and a balanced concept of
repair of highly comminuted long bone fractures. Comp Cont Ed Pract Vet 17:
35-38, 1995

Augat P, Burger J, Schorlemmer S, et a1. Shear movement at the fracture site delays
healing in a diaphyseal fracture model. J Orthop Res 2003; 21: 101 1-1017

Claes LE. Biomechanics of fracture repair and fracture fixation. In: Mow VC, Hayes
WC, eds. Basic Orthopaedic Biomechanics and Mechano-Biology. Philadelphia:
Lippincott Williams and Wilkins; 2005

Kaspar K, Schell H. Seebeck P et a1. Angle stable locking reduces interfragmnetary
movements and promotes healing after unreamed nailing. Study of a displaced
osteotomy model in sheep tibiae. J Bone Joint Surg Am 87: 2028-2037, 2005

Pennock AT, Schmidt AH, Bourgeault CA. Morse-type tapers: factors that may
influence taper strength during total hip arthroplasty. J Arthroplasty 2002; 17:
773-778

Gaebler C, Stanzl-Tschegg S, Heinze G, et a1. Fatigue strength of locking screws and
prototypes used in small-diameter tibial nails: a biomechanical study. J Trauma
1999; 47: 379-384

Gaebler C, Stanzl-Tschegg S, Laube W, et al. The fatigue strength of small diameter
tibial nails. Injury 2001; 32: 401-405

Dueland RT, Johnson KA, Roe SC, et a1. Interlocking nail treatment of diaphyseal
long-bone fractures in dogs. J Am Vet Med Assoc 1999; 214: 59-66

Brinker WO, Olmstead ML, Sumner-Smith G, et a1. Manual of Internal Fixation in
Small Animals. Ist ed. Berlin: Springer-Verlag, 1998; 125

Piermatti DL, Flo GL. Brinker, Piermattei, and F [0 's Handbook of Small Animal
Orthopedics and Fracture Repair. Third ed. Philadelphia: W.B. Saunders
Company, 1997

Claes LE, Eckert-Hubner K, Augat P. The effect of mechanical stability on local

vascularization and tissue differentiation in callus healing. J Ortho Res 20: 1099-
1 105, 2002

Claes LE, Heigele CA. Magnitudes of local stress and strain along bony surfaces
predict the course and type of fracture healing. J Biomech 32: 255-266, 1999

Claes LE, Heigele CA, Neidlinger-Wilke C, et al. Effects of mechanical factors on
the fracture healing process. Clin Orthop Relat Res Suppl 355: S 132-8147, 1998

139

39. Grundnes O, Reikeras O. The importance of the hematoma for fracture healing in
rats. Acta Orthop Scand 1993; 64: 340-342

140

APPENDIX ONE
FIXTURE SOPS
CHAPTER ONE
Plate Bending - Depending on the plate you are bending, you will need to change the
arm, the plate holder, and the hardened steel piece (fulcrum). The 3.5mm LC-DCP has a

different set than 3.5mm broad DCP. They are marked appropriately.

 

Figure Al-l. MSU BP lever arm.

0 Make sure all the adjustment screws are loose and the block should slide easily.

0 Place the plate in the plate holder and place the peg and tighten and plate holder
screws with the plate. One peg is enough. Do not use two pegs, see figure below.

0 Once the appropriate angle and level of the bend are obtained, loosen only the

screws on the plate holder to take the plate in/out.

141

’0‘
J
\-

  

_late holder and screws

1,) l I
T‘\ b <1 /‘u

_. . ,1 , ., block
\ i adjustment

Figure Al-2. MSU BP plate holder and adjustments.
0 Slide the block so the plate is at the correct level for bending. Spacer can be used
if needed to. Double check to make sure the plate is pushed forward and the
bending level is correct. The plate may be bent at this time.

0 Once the correct level is identified, tighten all the adjustment screws.

 

Figure Al-3. MSU BP with the plate pushed forward.

0 After the plate is bent, loosen only the screws on the plate holder (not the

adjustment screws).

142

0 Inspect the angle of the bend; adjust the stopper as necessary to obtain the desired
angle. Once the appropriate angle and level of the bend are obtained, loosen only
the screws on the plate holder to take the plate in/out.

0 You may start bending the rest of the plates. Remember to make sure the bending

level is correct and the plate is pushed forward each time you load the plate.

Strain Gage Selection and Application - Strain gage application was performed
according to manufacturer’s specifications (Vishay Micro-Measurements Inc.). Clean
latex gloves were worn during the surface cleaning process and application of the strain
gages so the bonding surface was not contaminated. The first step was to clean and
prepare the surface of the implant (Surface Preparation, Vishay Micro-Measurements
Inc.) This was done using M-Prep Conditioner A and 320 grit sand paper. The whole
surface of the implant was wet sanded with special attention given to the areas where the
gages would be applied. Next, the sanding residue was removed from the surface by
rinsing it with M-Prep Conditioner A and wiping it clean with cotton gauze. This
procedure was repeated with clean gauze until there was no more sanding residue visible
on the gauze after wiping the surface. The surface was then rinsed with M-Prep
Neutralizer 5A and wiped again with a clean piece of cotton gauze.

Strain gage selection was based on the required size of the gage and the estimated
levels of strain, which would be incurred during testing. Strain values were not expected
to exceed 4000 us. The selected strain gage (EA-06-O31DE-350 Vishay Micro-
Measurements) was rated for 3% strain, and had a gage length of 0.79 mm and a grid

width of 0.81 mm.

143

The overall size of the gage, which includes the plastic backing that the gage is
attached to, was 6.9 mm and 3.0 mm in length and width, respectively. This presented a
problem due to the size of the gage with respect to the area of the plate where it was to be
adhered. Therefore, the gage backing had to be carefully trimmed, without interrupting
the actual grid, to allow the gage to fit properly. After trimming of the backing, the
overall size of the gage was 4 mm by 1mm in length and width, respectively. Cellophane
tape was then placed over the gage to allow for accurate positioning of the gage on the
plate.

Positioning of the gage on the specimens was also difficult, once again due to the
small tolerances between the features of the plates and the strain gage area. Special
attention was given to make sure that there were no comers or edges of the gages hanging
over the edge of a hole or the edge of the plate. Because strain gages are uniaxial, and
measure an average strain over a finite area, it was important to align them on the axes
where maximum strain was thought to occur. Therefore, prior to placing the gages on the
plates, a line was sketched to mark the longitudinal and transverse axes. The longitudinal
axes were located 1 mm from the edge of the plate. The transverse axis was identified as
the line along which the bend was centralized. Neither line jeopardized the integrity or
cleanliness of the plate surface. The gage was then carefully aligned with both axes and
taped to the surface of the plate.

Once desired gage location was achieved, M-Bond 200 Adhesive and M-Bond
200 Catalyst C were used to attach it to the specimen. The tape was slowly peeled back at
a large angle, making sure the gage was still attached to the tape, and M-Bond 200

Catalyst C was delicately applied directly to the bottom of the gage. The catalyst was

144

allowed to dry for 30 seconds. M-Bond 200 Adhesive was then applied sparingly (1—2
drops) to the plate surface. The gage was then immediately pressed onto the surface of
the plate applying gradual pressure to the length of the gage to squeeze out excess glue,
but care was taken not to press too hard and squeeze out all of the glue. Firm pressure
was applied to the gage overnight using clamps. It should be noted that the clamps had
soft tips and applied pressure over the entire area of the gage so as not too damage the
grid of the gage with any point loads. The tape was removed at a sharp angle ensuring
that the gage was adhered to the plate surface and no longer attached to the cellophane
tape. The process of gluing the strain relief tabs to the plate was exactly the same, without
the extreme attention given to their exact location. This process was carried out on both
the medial and lateral sides of the plate, resulting in dual uniaxial strain measurement
(medial and lateral) of the specimens. The strain gages were check for a their appropriate
resistance readings (350 1 0.2% Q) with an ohmmeter prior to moving on to the next
step.

The next step in preparation of the plates was to prepare the wires and to solder
the relief wire to the strain gage tabs and the strain relief tabs. Single strand, solid copper
wire with a polyurethane enamel coating (American Wire Gauge size 34) was used as the

strain relief wire.

145

 

Figure Al-4. Close-up view of solderjoint attaching the copper wire (dashed white lines) to
the strain gage (strain relief tabs not shown).

Relief wire soldering followed standard soldering procedures at a temperature of
500°. The next step was to solder the 3-conductor cable (M-Line accessories, 326-DSV
6507, Measurements Group) to the strain relief tabs. Once again, standard soldering
procedures were followed. The final step in preparation of the plates was to apply a
coating material over the strain relief tabs to secure the 3-c0nductor cable and the relief
tab connection (M-Coat B Nitrile Rubber Coating, M-Line Measurements),
manufacturer’s application specifications were followed. The fully prepared implant is
shown below. The 3-wire cable was checked again for appropriate resistance of the gages
according to manufacturer’s specifications. It should be noted that the 3-conductor cable
also had an inherent resistance, which was taken into account on the final measurements

of resistance.

146

 

Figure Al-S. Fully prepared HP implant (dashed white lines are electrical lines).

CHAPTERS TWO AND THREE

Bone Model Drilling - Based on the 30% glass-filled structural nylon (Alro Plastics,
Jackson, MI), custom made by Olympia Tool, St. Johns, MI (contact person Todd
Deitrich 989-224-4817). Before starting, make sure the bushings are the correct size.
Change the bushings based on the SOP provided.

0 Fit the larger end of the bone model (30mm) in the fixture. Make sure it is well
seated with no wiggle room, yet you should be able to take them out with mild
force. Remove any drilling debris from previous drilling.

0 Note: some bone models are made slightly larger. In this case, you want to sand
the larger end down with 220-grade sandpaper to a point where it can just fit in

the fixture like in step one.

147

 

Figure Al-6. Disassembled drill guide.
0 Fit the top and bottom pieces together and tighten the screws just slightly. Adjust

the pieces until all the edges are smooth and the surface is flat.

 

Figure Al-7. Make sure all edges pointed out by the white arrows are flush and flat.
0 Once they are lined up, tighten the screws sequentially (finger tight).
0 When drilling the interlocking screw holes, make sure your drill goes in and out
the bushing easily and perpendicular to the bushing. Make sure you don’t hold the
fixture too tight, so the drill can go down the path of least resistance inside the

bushing without damaging the wall of the bushing or the drill bit.

148

0 Since the interlocking screw holes are perfectly aligned, you may drill one side at

a time.

 

Figure Al-8. View of drill bit insertion into drill guide.

0 Use 5.5mm drill bit for the linkage hole. You may drill one side and flip around to
drill the other side.

0 Loosen all the screws, but remove the two most distal screws and take the fixture
apart. Make sure you get rid of all the plastic debris in the fixture and inside the
bushing before drilling the next bone model.

0 Over-drill the cis cortex with the following:

o For 2.7mm bolt, over drill with 7/64 (2.76mm in metric system) drill bit
0 For 3.5mm bolt, over drill with 9/64 (3.55mm in metric system) drill bit
- Note: Make sure you label the size of the hole you drilled if you are going to put

various size holes on the same bone.

149

Bushing Adjustment — Change bushings based on the chart provided below for different
size screw or bolt.

0 Remove the top two halves as shown on the picture.

 

Figure Al-9. Disassembled top portion of drill guide.
0 Next, remove the inner housing for the bushing. The inner housing has two set-
screws holding the bushing down.

0 Loosen the set—screws and change the bushings to the desired size.

 

Figure A1-10. View of inner housing containing drill guide bushings.

0 Make sure when the bushings are inserted, the notch on the bushing line up with
set-screws and the bushings should not protrude from the slanted side (this is

where the bone contact would be).

150

 

Figure Al-ll. View of drill guide inner housing profile.

Desired size Drill bit size Bushing size
6mm ILN, 2.7mm screw 2.0mm 2.0mm
6mm ILN, 2.7mm bolt 2.7mm 2.7mm
8mm ILN, 3.5mm screw 2.5mm 2.5mm
8mm ILN, 3.5mm bolt 3.5mm 3.5mm
MSU ILN cis-cortex 4.0mm 4.0mm
MSU ILN trans-cortex 3.0mm 3.0mm

Table Al-l. Bushing selection for desired screw/bolt size.

ILN Placement -

Insert the bolts from the cis cortex, half way in.

Place the nail in the bone model.

Insert the bolts all the way in.

Make sure that the nail hole is centered.

Similar placement for the rest of the bolts.

For screw placement, the only difference is that you need to tap the bone first.
Tap 2.7mm screw with 2.0mm tap, 3.5mm screw with 2.5mm tap.

Looseness should be present if the traditional ILN is used. If not, the bolts/screws

may be bent or the holes are not lined up.

151

 

Figure A1-12. View of fully assembled construct.

Size of nail Bolt/Screw size
6mm 2.7mm
8mm 3.5mm

Table A1-2. Bolt/screw selection for desired nail size.

ILN-ESF Placement -
0 Similar to ILN bolt placement, but instead of the bolts, place the 2.7/3.2mm
hybrid pins into the pre-drilled holes.
0 1 cm should be removed from the most distal and proximal pins with enclosed
pin cutter. This should be done before placement into the bone model. The

shorter pins will allow easier placement.

 

Figure A1-13. Enclosed pin cutter.

152

o All the hybrid pins should be perfectly aligned from all angles.
0 Pre-place all the clamps and rod, but do not tighten them at this point.
0 Place the construct in the ILN-ESF alignment fixture (Figure Al-l4) with

distal aspect facing left.

 

Figure A1-14. ILN-ESF in alignment fixture.

Make sure that all the bones, clamps, and rods seat properly. There should not be
any wiggle room.
Tighten all the top screws first (the lag screw), then the bolts with torque wrench

(4.5 N*m is the ideal torque).

153

 

Figure Al-15. Correct screw tightening order.
Foam Tube SOP -
0 Take out about 30” of aluminum foil. Roll the aluminum foil around the DePuy
“hip drill 45° angle” plastic tube. Make sure one side is flush against the top of

the tube.

 

Figure Al-l6. Depuy “hip drill 45° angle" plastic tube.

0 Tape down the edge with several pieces of scotch tape or masking tape. Gently
smash the side of the aluminum roll that is longer than the plastic tube, but make

sure that the end of the tube is still round.

154

‘. \.

 

Figure Al-l7. Depuy plastic tube with aluminum foil shell.

0 Flatten the bottom from inside with the iron plunger.

 

Figure Al-18. Iron plunger.

0 Take the plastic tube out.

0 Mix lScc of mixture A with 15cc of mixture B in the urine cup. Mix them very
fast with a tongue depressor for about 10-15 seconds. Make sure you add A to B,
not B to A.

0 Pour the mixture down the tube and set it straight up for about 20 min.

0 Remove all the foil. Remove as much foil at the distal end as possible. Don’t
worry if you have leftover foil on the distal end. Cut approximately 1/3 of the

proximal end off since the density will be a bit different.

Foam Lathing SOP

Purpose and Use - The foam is meant to simulate cancellous bone in the synthetic bone
model. It is inserted into the metaphyseal region of the bone and it’s sole purpose it to
provide otherwise non-existent vertical stability (parallel to the screws) to keep the ILN

centered throughout testing. If the foam is not present, the nail will fall to the bottom of

155

the inner diaphyseal diameter, the off axis position of the nail will erroneously affect the

positive and negative loading data.

The foam lathe bit has been designed to fill approximately 10 mm proximal to the
beginning of the taper and 15-20 mm distal to the beginning of the taper (30 mm was the
original length, this was then reduced to 25 mm). The distal dimensions were specifically
chosen to provide support for the nail only, and to not reach the nail screw interface.

Materials

 

Figure Al-l9. Foam lathing materials, including:

A) hand saw

B) foam cylinder

C) drill bit chuck

D) lathe head chuck

E) foam lathe bit chuck

F) t-square

G) foam lathe bit (30 mm version shown)
H) drill bit

156

     

 
   

Lathe, including:

 

E

A) Lathe head B) Lathe bit crosshead C) Drill bit crosshead

Procedure - Determine desired amount of nail support, and decide which version of the
foam lathe bit is appropriate.
0 Version 1 produces a 30 mm (overall) piece of foam, with 20 mm distal to the
beginning of the taper (use when more support of the nail is desired).
0 Version 2 produces a 25 mm (overall) piece of foam, with 15 mm distal to the

beginning of the taper (use when less nail support is desired).

157

 

Notice the additional piece on
Version 2. This reduces the
overall size of the foam piece

 

 

   

Figure Al-21. Foam lathe bit, Version 1 (left) and Version 2 (right)

Place foam lathe bit into lathe bit holder located in the lathe bit crosshead

between the foam lathe bit and the base of the bit holder, there should be a
large black washer, approximately 9/32” thick and 1 3A” in diameter.

0 use a T-square to make sure that the foam lathe bit is oriented 90 degrees to
the rotating surface.

0 Tighten the foam lathe bit into the bit holder using the foam lathe bit chuck.

_ foam lathe I. E
-

Figure Al-22. Close-up view of lathe bit holder (left) and tightening of lathe bit holder using
foam lathe bit chuck. Note that using T-square while tightening will ensure proper alignment.

 

158

0 Insert drill bit into drill bit crosshead (use a #3 drill bit for 6 mm nails, use a
19/64” for 8 mm nails)

0 Tighten the drill bit into the bit crosshead using the drill chuck.

 

Figure A1-23. Close-up view of drill bit crosshead with drill bit and chuck shown.

Insert the foam tube into the lathe head
0 align the foam tube with the rotational axis of the lathe as much as possible,
this may take some adjustment and getting used to (some tubes are better than
others).

0 when tightening lathe grip, be sure that the teeth of the lathe head sink in the

foam at least 1A” to 3/8”, if not, the tube will come loose when trying to lathe.

Va" to 3/8"
penetration

     

Figure Al-24. Tighten the foam tube into the lathe head using the lathe head chuck.

159

Begin lathing, several tips to keep in mind though:
Take the first (shiny) layer of tube off by moving the lathe bit

crosshead from right to left, not outside to inside (front to back).

 

Continue moving the lathe bit crosshead from right to left until the

shiny layer of the foam tube for the foam piece is removed.

 

Once the shiny layer is removed, the lathe crosshead can then be

moved towards the center of the tube (away from your body).

Continue to move the lathe crosshead towards the center of the

tube, do this slowlylll!

 

160

 

 

Take measurements of the tube as you reduce the diameter.
When a desirable diameter is achieved (slightly > 0.25 mm),
take note of the position of the crosshead indicator. This will

ensure consistent foam pieces.

The drill bit crosshead can be adjusted
with this lever, located on the back of the

lathe.

The entire drill bit crosshead is then slid
along the lathe base and into the central
axis of the foam, do this slowly and

carefully!

 

Continue to move crosshead until the
entire length of the foam piece (25-30

mm) has been drilled.

161

Using the handsaw, oriented with the teeth
facing up, place the tip of the saw (under the
foam tube) on the crosshead. In a very slow,

controlled motion, lift the handle of the saw

 

towards the foam tube.

Continue lifting the handle of the handsaw
towards the center of the foam tube.
Continue lifting until the tube is almost

cut. Be careful as the foam piece is cut

   

from the tube.

Do not grab the foam piece. It is best to allow the

piece to fall.

 

162

 

When the foam piece is cut from the tube, there will be some residue from the tube. Just

lightly run the foam piece over some sandpaper until the residue is gone.

 

Figure Al-25. Finally, push the foam piece into the bone model

163

APPENDIX TWO
TESTING SOPS AND PIDS

INSTRON MACHINE GENERAL INFORMATION
Location: Orthopaedic Biomechanics Laboratories

4th floor, East Fee Hall
Technician/Operator: Cliff Beckett

A41 1 East Fee Hall

353-3122

0 Must be present during operation.

Parts and Terms

       

crosshead
guides

crosshead
adjustment

 

Figure A2-l. General view of Instron machine.

Actuator: consists of the working parts of the machine: the piston, the servo-hydraulic
valve, and the LVDT (linear variable displacement transducer).

Piston: only part of the machine that will move during testing.

Axial Load Cell: not actually part of the Instron machine, but attached to the end of the
piston. Essential to record load data. Load cells of various shapes and sizes can be fitted
to the Instron Machine. Do not over tighten when attaching load cell to piston, can
damage [and cell.’.’.’.’.’ Usually about an inch of threaded length is adequate to securely fix
the load cell to the piston.

Crosshead: the moveable base connecting the actuator to the base of the Instron machine
vie the guide rails.

Guides: allow for the very top, or crosshead of the Instron machine, to be adjusted
vertically. An orange clamp has been put on the top of the left guide to prevent the head
from going to far up and hitting the ceiling, be cautious of this.’!.’.’.’

Crosshead Adjustment: controls crosshead height. There are three handles from left to
right they are: raise, lower and clamp. The hydraulics must be on to move the crosshead;
the on button is located on the 8500 board attached to the Instron. The clamp handle
needs to be loosened first to allow any further adjustment. After the clamp handle is
loosened, loosening of either the raise of lower handle will result in motion of the
crosshead in that direction. The speed at which the crosshead moves is proportional to
how much the raise or lower handles are loosened.

Base: all fixtures will be mounted to the base of the Instron in some way shape or form.
8500 Board: triggers and controls all data acquisition by the computers. The 8500 board

will be further discussed in the data acquisition section of the SOP.

165

Data Acquisition

Computers

I

 

:pyr.(m via-4.“. ; .-

:1; (. L____,g

 

Figure A2-2. A2D, Encoder, and Instron monitors and computers.

8500 Board

 
  

 

   
   
 
 
  
  
  

hydraulic

load and
Controls

position
display

1
actuator
controls

 
 

 

 

 

H :6
. . actuator
positron panel jog
controls

line
adjustment

remote button

 

Emergency STOP

Figure A2-3. 8500 Board.

166

 

 

Instron Computer: used for all tests. All programming is done on this computer, instead
of the 8500 board, making them easier to modify and edit. All the data from the Instron
machine, which goes through the 8500 board, is sent to this computer and stored as a
*.csv file, a copy of this data is also sent to the A2D computer for interpretation into
digital data.

Encoder Computer: this computer is only used for tests that involve additional encoders,
such as the bending tests, which collect data from two rotary encoders. The program
Labview is installed on this computer, it allows us to tell the computer to start recording
data when triggered. The trigger is usually set to be the first bit of movement by the
piston. A copy of this data is also sent to the A2D computer for further interpretation.
A2D Computer: this computer takes the analog data that it receives from both the Instron
computer and the Encoder computer and converts it into digital format. This computer is
also equipped with Labview and must be run simultaneously with the Encoder computer.

Detailed use of the 8500 board is included within each test procedure.

167

TORSION TEST SOP

Test

Sine wave (load control) ------- controlled byl35 N *m torque load cell,
previously controlled by 2500 lb axial load cell

0 N*m preload to 1 5 N*m.

10 cycles @ 8 seconds a cycle (0.125 H2)

PIDS

Load and position PIDs are construct dependent, see Testing PIDS
Data

- Torque is recorded via the torque load cell

- Actuator displacement measured by the LVDT is recorded (converted to angular
displacement in the excel macro).

- Angular displacement is recorded via the rotary encoder

168

Materials

 
 
 
 

  

torque load
cell not shown

  
 
   

 
  

fixture
locking
bolts

potting
sleeves

 
 

rotary encoder
not shown

 
    

 

sleeve to potting
locking bolts

 
   

Figure A2-4. Torsion fixture and required materials.

Setup

trocar [101111

_

untapered bolts
1 ill!
1
tapered bolts

Figure A2-5. View of construct within torsion fixture.

I
I
x

i‘

 

169

lixtur'c locks to table
by two bolts here

c-clamp attachment
not shown

 

Figure A2-6. View of torsion fixture within Instron machine.

Torsion Test Background Information

The torsion tests are intended to simulate a torque of 5 N*m being applied. The
amount of torque being generated is completely proportional to the size of the moment
arm where the load is being applied. The current torsion fixture is equipped with a gear of
radius 0.0381 m. To achieve the desired torque of 5 N*m, a 131.2 N load must be applied
by the piston. This was calculated by:

(0.0381 m)*(131.2 N) = (5 N*m)

Prior to using the torque load cell, the axial load cell was programmed to 1 131.2 N.

Procedure

Attaching the torsion fixture to the base of the Instron machine - The primary concern
when mounting the fixture is to ensure that there is no friction or contact between the
rack and the walls of the fixture itself. To do this, the axial load cell must first be

calibrated with the rack of the fixture attached. This is a very important step in recording

170

accurate data. If this is not done, large amounts of friction can develop between the rack
and the torsion fixture. This could result in erroneous data or damage to the Instron
machine. So, attach the rack to the axial load cell and calibrate by pressing the “calibrate”
button on the load panel of the 8500 board, then select “cal”, then “auto”, the screen will
ask you to ensure that the transducer is at zero, double check, then select “go”. The
calibrate button will rapidly blink will it is calibrating itself, you will see it go to its
maximum values and then it will zero itself. The load cell is now calibrated to take into
account the rack of the torsion fixture when the red light stops blinking.

With the rack attached to the calibrated load cell, jog the actuator to its highest
position, or high enough to allow the rack to fit between the walls of the torsion fixture.
Crosshead adjustment may be necessary, if so, adjustment of the crosshead position is
controlled by the 3 black knobs on the silver box under the Instron table on the left side.
The “clamp” knob must be loosened (turn counter clockwise) first, then the “lower” knob
must be slowly loosened to adjust the crosshead positioning to an appropriate height
(notice that the speed of the crosshead position adjustment depends on how much you
turn the “clamp” and “lower” knobs, GO SLOW, BE SAFE. DO NOT forget to tighten
(turn clockwise) the “clamp” and “lower” knobs before doing anything else, notice
though that these knobs do not need to be tightened extremely tight, be snug but don’t
overdo it.

With the rack hanging from the axial load cell, move the actuator up and down
and look for points where the axial load increases rapidly. Ultimately, you should be able
to have less than a couple of Newton felt by the load cell throughout the entire stroke of

the Instron. Once this is achieved, the base of the torsion fixture is then mounted tot eh

171

base of the Instron machine. The fixture is mounted to the Instron machine by two bolts
and a C-clamp. These two bolts are referred to as the fixture locking bolts.

Preparing the specimen - After mounting the specimen in the potting cups, it can then be
mounted into the potting sleeves. Using eight sleeve-to-potting locking bolts, lock the
potting into the sleeve. Be sure to orient the potting sleeves as indicated on the sleeve
(arrows should point away from the middle of the specimen towards the fixture mounts).
Edge of potting and sleeve will be flush if properly placed.

Mounting the specimen - The potting sleeves are then inserted into the torsion fixture.
Sleeve to fixture locking bolts secure its position.

0 The sleeve-to-fixture bolts are different. There are 8 total, 4 with tapered ends for
locking, 4 with blunt ends for removing slack between the potting sleeves and the
torsion fixture cups.

0 The tapered end bolts must be placed in the holes on the fixture, which are l in.
from the edge. This will allow proper alignment of the specimen and the fixture.

Tighten all bolts until the potting sleeves are stable within the torsion cups

Calibrating torque load cell - The torque applied to the specimen is measured with a
torque load cell. The calibration of the torque load cell must be done with Cliff Beckett
present. The PIDS given in the Testing PIDS section are given for the axial load cell and
are different from those implemented when using the torque load cell.

Adjusting actuator position - When running torsion tests, the goal is to start at a point
when no torque is applied to the specimen. This is done by adjusting the actuator position
until the reading of the torque load cell is zero (or very close). So, to zero the torque on

the specimens prior to beginning the test, first go into high pressure by pressing the

172

“high” button on the 8500 board. You will then manually adjust the position control on
the 8500 board until a torque reading close to zero is achieved. The Instron machine can
now be placed into load control. It is important to realize that once the machine is in load
control, position is no longer a factor, the machine now will move in whatever direction it
needs to, to achieve a desired load (in other words, BE CAREFUL, load control is
dangerous! ).

Running the program - The torsion test (torqsin.b1k) program is located in the
CzMavemaker/Lansdown directory. After the load is set to zero by adjusting the position
set point, double check that you are in “high” pressure, and begin the test.

After the test - After the test is run, be sure to switch the Instron machine back into

position control.

173

BENDING TEST SOP

Test

Sine wave (load control) ------- controlled by 2500 lb axial load cell

0 N preload to 1- 151.7 N. (equivalent to moment of 3.5 N*m, based on specimen length)
10 cycles @ 8 seconds a cycle (0.125 H2)

PIDS

Load and position Ple are construct dependent, see Testing P[Ds (calibrated for use of
2500 lb axial load cell)

Data

— Load on the axial load cell is recorded.

- Actuator displacement measured by the LVDT is recorded.

- Angular displacement measured by the 2 rotary encoders is recorded

 

Materials

steel balls sticking out of the cup.
this is where the arm is attached

     

old bending arm

4 Sleeve to
fixture
locking

bolts

Figure A2-7. Bending fixture and required materials.

174

Setup

 

 

 

partial ball joint
bet“ een linkage arm
and main bar

  
 
 

 
 
 
 

hearings in main bar

 

 

 

removable pin

  
  

bending cup-to-
potting bolts

 
   
    
 

end.
channel 1

  
  

 

lixture locks to table by
steel bar here

 
 

 

 

 

 

 

steel balls on loading cups

 

 

 

| steel bushings in linkage arm

 

Figure A2-8. View of construct within bending fixture.

175

Bending Test Background Information

The bending tests are intended to simulate a moment of 3.5 N *m being applied to
the specimen. The moment being generated is completely proportional to the size of the
design of the individual bending cup fixtures. Because four—point bending is the desired
tests, the generation of a 3.5 N*m moment under four-point bending must be considered.
The current bending fixture is equipped with two individual bending cups, each of these
cups is allowed to rotate in one plane. This axis of rotational is approximately 0.04629 m
away from the point of applied load (the steel balls). Based on these dimensions, a load of
75. 6 N must be applied to each bending cup to achieve a desired load of 3.5 N*m.
Because the fixture is designed to transfer one axial load through the main bar into two
equal loads on each linkage arm, an axial load of 151.2 N is equivalent to creating a 3.5

N*m pure bending moment on the specimen. This is calculated by:
(75.6 N)*(0.04629 m) = (3.5 N*m)

Once again, this bending moment is applied to both sides of the bending fixture; therefore

the bendsine program is designed to achieve a maximum axial load of 151.2 N.

Procedure

Attaching the bending cups to the base of the bending fixture - The first step is to attach
the two bending cups to the base of the fixture. Be sure to attach the fixed cup (ch. 0) on
the left and the free cup (ch. 1) on the right. The cups do not need to be locked to the base
yet.

Aligning and connecting the bending fixture and the Instron machine - Prior to

aligning the cups, a loosely placed specimen should be put into the bending cups, making

176

sure the surface of the potting is flush with the surface of the bending cups. The main bar
is attached to the Instron actuator via a removable pin. Once the main bar of the bending
fixture is attached to the Instron, the next step is to lower the crossheads of the Instron
machine and to lower the actuator position so that the linkage arm steel bushings line up
with the steel balls of the bending cups. In order to achieve this, the two individual bases
of the bending cups must be perfectly aligned with one another.

Putting a specimen into the cups, then rotating the cups until all of their surfaces
are parallel or perpendicular to one another, does this. Adjustment of the crosshead
position is controlled by the 3 black knobs on the silver box under the Instron table on the
left side. The “clamp” knob must be loosened (turn counter clockwise) first, then the
“lower” knob must be slowly loosened to adjust the crosshead positioning to an
appropriate height (notice that the speed of the crosshead position adjustment depends on
how much you turn the “clamp” and “lower” knobs, GO SLOW, BE SAFE. DO NOT
forget to tighten (turn clockwise) the “clamp” and “lower” knobs before doing anything
else, notice though that these knobs do not need to be tightened extremely tight, be snug
but don’t overdo it.

Once the vertical position of the crosshead is determined, the base of the bending
fixture must be adjusted to properly align it with the main bar of the fixture. This is done
by rotating the base of the fixture until two things are possible, one is to lower the
actuator and see that the two linkage points are aligned with the steel balls (from an
overhead view), and there must also be room for the steel bar that attached the base of the

bending fixture to the base of the Instron machine. Once desired alignment is achieved,

177

the base of the bending fixture may be attached to the base of the Instron machine via two
c-clamps and a steel bar on the right side, and a c-clamp by itself on the left side.
Calibrating the axial load cell - The final step prior to mounting the specimen into the
fixture is to calibrate the load cell taking into account the weight of the main bar and two
linkage arms. Load cell should be calibrated with the entire bending arm hanging from it,
but not attached to the fixture (very important step in recording accurate data). If this is
not done then the load cell will consider the weight of the main bar and linkage arms (not
good!) So, attach the main bar, with the two linkage arms, to the actuator and calibrate
the load cell. Load cell calibration is done by pressing the “calibrate” button on the load
panel of the 8500 board, then select “cal”, then “auto”, the screen will ask you to ensure
that the transducer is at zero, double check, then select “go”. The calibrate button will
rapidly blink will it is calibrating itself, you will see it go to its maximum values and then
it will zero itself. The load cell is calibrated to take into account the main bar and linkage
arms when the red light stops blinking. You are now ready to mount the specimen into
the fixture.

Mounting the specimen - The potting can be inserted directly into the bending fixture
cups. If slack is present before a construct is placed into the bending fixture, than check
to make sure it is still present after the specimen is mounted and secured by the cup to
potting locking bolts. Each cup should have two screws on the top only, bottom screws
are not necessary. If a specimen loses its slack when placed into the bending cups, then
the specimen is out of alignment. If this occurs, simply loosen the attachment screws and

rotate the specimen/potting until the actual slack is present again. In these cases, the

178

potting must be reattached to the bending cups via screws that do not have shoulders (any
standard screw).

Setting PID levels - Before or after opening the program you have to adjust the PID
levels. This is done manually on the Instron 8500 by pressing the calibrate button on the
load panel, then loop and then enter the appropriate numbers.

Adjusting actuator position - When running bending tests, the goal is to start at a point
when no load is applied to the specimen. This is done by adjusting the actuator position
until the reading of the axial load cell is zero (or very close). Remember that the weight
of the main bar and linkage arm has already by neglected by calibrating the load cell with
them attached. So, to zero the load on the specimens prior to beginning the test, first go
into high pressure by pressing the “high” button on the 8500 board. You will then
manually adjust the position control on the 8500 board until a load reading close to zero
is achieved. The Instron machine can be placed into load control. It is important to realize
that once the machine is in load control, position is no longer a factor, the machine now
will move in whatever direction it needs to, to achieve a desired load (BE CAREFUL,
load control is dangerous!).

Running the program - The bending test (bendsin.blk) program is located in the
CzMavemaker/Lansdown directory. After the load is set to zero by adjusting the position
set point, double check that you are in “high” pressure, and begin the test.

After the test - After the test is run, be sure to switch the Instron machine back into

position control.

179

COMPRESSION TEST SOP

Test

Sine wave (load control) ------- controlled by 2500 lb axial load cell

—5 N preload to —l76 N. (construct is never in tension)

10 cycles @ 1 second a cycle (1 Hz)

data sampling is 500 Hz for 10.5 seconds

PIDS

Load and position Ple are construct dependent, see Testing Ple

The load vs. time sine wave should be used to determine if proper load PID levels are
being used. Pay attention to the following:

o The most important thing is to analyze the wave to look for areas where it appears
that the Instron machine may go unstable; this is usually characterized by a sharp
jump in load over a very short period of time. (it looks like a spike in the sine
wave, but it may not be very obvious). The Instron machine going unstable is
the most dangerous thing that can happen during testing, if your PID levels are
not carefully managed, severe damage and injury may occur.

0 Check maximum and minimum peak values and make sure that the Instron
machine is not overshooting or undershooting the desired load levels of —5 N and
—176 N. (Do not be satisfied with peak values until you are at least 1 2 N of
your desired load)

0 You must also analyze the peaks and check for flat spots (where a constant load is

held over some length of time)

180

The relationship between the Ple and the characteristics of the load vs. time wave are
based on the construct being tested and the speed of the test. Contributing factors to these
variations are: slack in the system, extreme compliance or stiffness, and overall expected
range of motion.

Data

- Load on the actuator load cell is recorded.

— Actuator displacement measured by the Instron LVDT is recorded.

Materials

cup-to-
potting bolts
and washers

 

 

5/8 In. threaded rods

Figure A2-9. Compression fixture and required materials.

181

Setup

 
   
   

3500 lb axial
load cell

compression cups-to—
potting bolts 18 total)

trocar point

Figure A2-10. View of construct within compression fixture attached to Instron machine.

Procedure

Attaching the compression cups - The first step is to securely attach the top cup to the
load cell. This attachment should consist of two 5/8"-l8 nuts and a 5/8”-18 threaded rod,
with about a l/2” between the two nuts. After the top cup is attached, the bottom cup may
then be attached, but not securely. Attachment of the bottom cup consists of the large
plate (4” dia, '/2” thick), the large washer (3” dia, 1/8” thick), a regular washer (1” dia,
1/16” thick), a 5/8”-18, nut and a 5/8”-18 threaded rod. Remember; do not secure the

bottom cup yet.

182

 

Figure A2-l 1. Order of materials for bottom cup.

Aligning the cups - The next step is to lower the crossheads of the Instron machine and
to lower the actuator position itself so that the two cups are allowed to nearly contact
each other (1 mm apart). Adjustment of the crosshead position is controlled by the 3
black knobs on the silver box under the Instron table on the left side. The “clamp” knob
must be loosened (turn counter clockwise) first, then the “lower” knob must be slowly
loosened to adjust the crosshead positioning to an appropriate height (notice that the
speed of the crosshead position adjustment depends on how much you turn the “clamp”
and “lower” knobs, GO SLOW, BE SAFE). Be careful not to try to get the two
compression cups too close with the crosshead control (knobs), this is just a rough
position adjustment. DO NOT forget to tighten (turn clockwise) the “clamp” and “lower”
knobs before doing anything else, notice though that these knobs do not need to be
tightened extremely tight, be snug but don’t overdo it.

The next step is to lower the actuator position for a more precise adjustment. This
is done by the “jog” buttons on the Instron 8500 board. You will notice that once the
“clamp” knob is loosened, the “actuator” light on the 8500 board turns to off, this a safety
precaution. Before you have any control of the actuator, you must put the machine into

“low” pressure on the 8500 board. While in low pressure, the actuator can be adjusted by

183

the “jog” buttons. This should be done slowly to avoid any accidental contact between
the two cups. When the two cups are within range (approximately 1 mm apart), move the
bottom cup accordingly so that they both line up perfectly axial. ALIGNMENT IS
CRUCIAL to the compression tests!

Securing the cups - Once perfect alignment is reached, the bottom cup can then be
secured. Be sure to take note of the position of the locking holes in the bottom potting
cup, (two holes should be visible; this is the only configuration which allows the
specimen to be placed correctly, with the bolts facing to the right). Don’t worry about the
hole positioning of the top cup, this can be rotated later by turning the actuator shaft
manually.

Once you are satisfied with the cup positioning with respect to each other, the

crossheads can then be raised to an appropriate height to fit the specimen. Raising the
actuator and crosshead is done in the same way as lowering it, except that after the
“clamp” knob is loosened, the “raise” knob is used now. DON’T FORGET to tighten
both of them when the desired position is reached.
Calibrating the load cell - The final step prior to loading the specimen into the fixture is
to calibrate the load cell taking into account the weight of the top compression cup and its
attachments. This is a very important step in recording accurate data. If this is not done
then the load cell will consider the weight of the top cup and its attachments to be a
tensile load, this is not good!

Calibrating the load cell with the top cup is done by pressing the “calibrate”
button on the load panel of the 8500 board, then select “cal”, then “auto”, the screen will

ask you to ensure that the transducer is at zero, then select “go”. The calibrate button will

184

rapidly blink will it is calibrating itself, you will see it go to its maximum values and then
it will zero itself. The load cell is calibrated to take into account the top potting cup when
the red light stops blinking. You are now ready to mount the specimen into the fixture.
Mounting the specimen - It is very important that the compression fixture is properly set
up, if the cups aren’t aligned on there axis, loading the specimen will be nearly
impossible, and erroneous data will result. When the crosshead positioning is done, it is
important that you allow enough room between the two cups for the entire length of the
potted specimen. The specimen is to be consistently mounted with the trocar point of the
nail facing down and the bolts/screws pointing to the right. First slide the bottom potting
partially into the compression cup, then slowly lower the actuator position with the “jog”
buttons until the mounted specimen is restricted to only move vertically on the same axis
as the two compression cups.

Securing the specimen - Once all three (the two cups and the specimen) are aligned
axially, the actuator can be lowered further until the locking bolts 'can be inserted through
the compression cups and into the potting. Take careful notice that you are not loading
you specimen while placing the bolts, loads can go unnoticed do to the minimal
displacement under compressive loads and this can make it difficult to place the locking
bolts, they should go in freely, without much force. Depending on the length of the
locking bolts, washers may or may not be used. Once all eight of the bolts are secured,
the specimen is mounted.

Setting PID levels - Before or after opening the program you have to adjust the PID
levels accordingly. This is done manually on the Instron 8500 by pressing the calibrate

button on the load panel, then loop and then enter the appropriate numbers.

185

Preloading the specimen - The last step before running a compression test is to preload
the specimen. First go into high pressure by pressing the “high” button on the 8500
board. Then manually adjust the position control on the 8500 board until a load reading
close to zero is achieved. Instron machine can now be placed into load control.
Remember that once the machine is in load control, position is no longer a factor, the
machine now will move in whatever direction it needs to, to achieve a desired load (BE
CAREFUL!). Under position control you may now select a desired set point, the screen
will prompt you to give a load value, enter -5 N (hit 5 first, then the — sign), then hit
enter. You should now see a reading close to -5 N on the 8500 board, this is your
compressive preload.

Running the program - The compsine (compsin.blk) program is located in the
:/C/Wavemaker/Lansdown directory. Double check that you are in “high” pressure, then
run the program according to the Wavemaker program instructions.

After the test - After the test is run, be sure to switch the Instron machine back into

position control.

186

TESTING PIDS

Torsion

Specimen type: ILN6s (short & long)

Description: 6 X 185 mm nail, 4-2.7 mm bi—cortical screws, 5 & 12 cm osteotomy
Load PID: -2, 0, 0, 0

Position PID: 13, l, 0, 1

Specimen type: ILN6b (short & long)

Description: 6 X 185 mm nail, 4-2.7 mm bi-cortical bolts, 5 & 12 cm osteotomy
Load PID: -2, 0, 0, 0

Position PID: 13, l, 0, l

Specimen type: ILNSS (short & long)

Descriptionzi8 X .185 mm nail, 4-3.5 mm bi-cortical screws, 5 & 12 cm osteotomy
Load PID: -2, 0, 0, 0

Position PID: l3, 1, 0, 1

Specimen type: ILN8B (short and long)

Description: 8 X 185 mm nail, 4-3.5 mm bi-cortical bolts, 5 & 12 cm osteotomy
Load PID: -2, 0, 0, 0

Position PID: 13, l, 0, l

Specimen type: ILNn (short and long)

Description: 8-6 X 185 mm novel nail, with 4—SCP locking devices, 5 & 12 cm
osteotomy

Load PID: -2, 0, 0, 0

Position PID: 13, l, 0, l

Specimen type: br-DCP (short and long)

Description: broad DCP with 8—3.5 mm bicortical screws, 5 & 12 cm osteotomy
Load PID: -2, 0, 0, 0

Position P11): 13, 1,0, 1

Specimen type: ILN-ESF

Description: 6 X 185 mm nail, with 4 hybrid 2.7 mm ILN bolt/3.2 mm ESF pins and 1
small SKTM titanium connecting bar (6.3 mm [diameter] by 150 mm [length])

Load PID: -2, 0, 0, 0

Position PID: 13, 1,0, 1

.187

 

Mediolateral and Craniocaudal Bending

Specimen type: ILN6s (short & long)

Description: 6 X 185 mm nail, 4-2.7 mm bi-cortical screws, 5 & 12 cm osteotomy
Load PID: 10, 0, 2, 0

Position PID: 13, l, 0, l

Specimen type: ILN6b (short & long)

Description: 6 X 185 mm nail, 4-2.7 mm bi-cortical bolts, 5 & 12 cm osteotomy
Load PID: 10, 0, 2, 0

Position PID: l3, 1, 0, 1

Specimen type: ILN8S (short & long)

Description: 8 X 185 mm nail, 4-3.5 mm bi-cortical screws,5 & 12 cm osteotomy
Load PID: 10, 0, 2, 0 '
Position PID: 13, 1, 0, 1

Specimen type: ILN8B (short & long)

Description: 8 X 185 mm nail, 4-3.5 mm bi-cortical bolts, 5 & 12 cm osteotomy
Load PID: 10,0, 2, 0

Position PID: l3, 1, 0, l

Specimen type: ILNn (short & long)

Description: 8-6 X 185 mm novel nail, with 4-SCP locking devices, 5 & 12 cm
osteotomy

Load PID: 6, 0, 2, 10

Position PID: 13, 1,0, 1

Specimen type: br-DCP (short & long)

Description: broad DCP with 8-3.5 mm bicortical screws, 5 & 12 cm osteotomy
Load PID: 6, 0, 2, 10

Position PID: 13, 1, 0, 1

Specimen type: ILN-ESF

Description: 6 X 185 mm nail, with 4 hybrid 2.7 mm ILN bolt/3.2 mm ESF pins and 1
small SKTM titanium connecting bar (6.3 mm [diameter] by 150 mm [length])

Load PID: 6, 1, 2, 10

Position PID: 13, 1,0, l

188

Compression

Specimen type: ILN6s (short & long)

Description: 6 X 185 mm nail, 4-2.7 mm bi-cortical screws, 5 & 12 cm osteotomy
Load PID: 0, 1, 2, 3

Position PID: l3, 1, 0, 1

Specimen type: ILN6b (short & long)

Description: 6 X 185 mm nail, 4-2.7 mm. bi-cortical bolts, 5 & 12 cm osteotomy
Load PID: 0, l, 2, 3

Position PID: 13, l, 0, 1

Specimen type: ILNSS (short & long)

Description: 8 X 185 mm nail, 4-3.5 mm bi-cortical screws,5 & 12 cm osteotomy
Load PID: 0, 1, 2, 3

Position PID: 13, 1, 0, l

Specimen type: ILN8B (short & long)

Description: 8 X 185 mm nail, 4-3.5 mm bi-cortical bolts, 5 & 12 cm osteotomy
Load PID: 0, 1, 2, 3

Position PID: 13, 1, 0, 1

Specimen type: ILNn (short & long)

Description: 8-6 X 185 mm novel nail, with 4-SCP locking devices, 5 & 12 cm
osteotomy

Load PID: 0, 1, 2, 1

Position PID: l3, 1, 0, 1

Specimen type: br-DCP (short & long)

Description: broad DCP with 8-3.5 mm bicortical screws, 5 & 1.2 cm osteotomy
Load PID: 0, l, 2, 1

Position PID: 13, l, 0, 1

Specimen type: ILN-ESF

Description: 6 X 185 mm nail, with 4 hybrid 2.7 mm ILN bolt/3.2 mm ESF pins and 1
small SKTM titanium connecting bar (6.3 mm [diameter] by 150 mm [length])

Load PID: 0, 1, 2, 1

Position PID: 13, l, 0, 1

189

APPENDIX THREE

THEORETICAL ANALYSIS
CHAPTER ONE

Strain Readout

8s = (Rg)/IFg*(Rg + Rs)] [A34]

Equation A3-l. Shunt calibration strain readout, where:

85 : strain readout with shunt cal.

Rg : strain gage resistance ( 350 Ohms)
Fg : gage factor (2.12 :t 1.0%)

RS : shunt cal. resistance (30060 Ohms)

Longitudinal Strain Derivation/Calculation
8x = - y/p [As-21

Equation A3-2. Longitudinal strain as a function of radius of curvature and distance
from neutral axis, where:

8x : longitudinal strain
y : distance from the neutral axis to the point of interest

p : radius of curvature

Since both plates are bent to the same degree, their curvatures are equal, reducing Eq. [B]

[02

(gap/Yap) = (8LC-DCP/yLC-DCP)

190

Hooke’s Law

a:

Flexural Formula

E*8 [A3-3]

Equation A3-3. Hooke’s Law, where:
6 : elastic stress

E : modulus of elasticity

8 : elastic strain

S = - (M*y)/I [A3-4]

Equation A3—4. Flexural formula, where:
O : bending stress
M : moment at the point of interest
y : distance form the neutral axis to the point of interest

I : area moment of inertia of the cross section at the point of interest

Area Moment of Inertia (AMI)

The AMI is a geometrical property (Equation E) '7, which is related to the amount

of material in a given region, therefore, the AMI is higher in the solid portion of the plate

than it is in the region where a hole is present.

191

2 2
Ix = ly dA Iy = ix dA [AS-5]
Equation A3—5. Area moment of inertia (AMI), where:
Ix, Iy: AMI with respect to x and y-axis, respectively

y, x: equation of the line forming the boundary of the cross section

A: area of the cross section

AMI and Strain
8 = GIE = (M*y*E)/I [A3-6]

Equation A3-6. Showing the inverse relationship between AMI and strain.

192

CHAPTER THREE
Matlab Programming for AMI Calculation

*note: % sign denotes a non—command line and is typically a brief description of the
function performed by following command

Screw, Bolt, and SCP AMI Calculation (ScrewBoltSCP_AMIcalc.m)

clear
clc
global x
%User input Diameter of screw or peg at nail hole level
CD=input('Enter cis core diameterz');
%User input Diameter of screw or peg at nail hole level
TD=input('Enter trans core diameterz');
if CD==TD
AMI=<tpi)*(co/2)‘4)/4
end
if CD>TD
ThreadedSection_AMI=((pi)*(CD/2)‘4)/4
MidTaperedSection_AMI=((pi)*(((CD+TD)/2)/2)”4)/4
SolidSectionAMI=((pi)*(TD/2)*4)/4

end
if CD<TD

'ERROR!!!!! Cis diameter must be greater than Trans diameter'
end

Plate AMI Calculation (Plate_AMlcalc.m)

clear

clc

global x

%User input width of plate at screw hole level
PW=input('Enter width of plate at screw hole level:');
%User input thickness of plate at screw hole level
PT=input('Enter thickness of plate at screw hole levelz');
%User input screw hole diameter at nail hole level
HD=input('Enter diameter of screw holez');

'At nail hole level'
MedioLateral_AMI=((1/12)*(PW)*(PTA3))-((l/12)*(HD)*(PT‘3))
CranioCaudal_AMI=((1/12)*(PT)*(PW‘3))-((1/12)*(PT)*(HD‘3))
'At solid section'

MedioLateral_AMI=((1/12)*(PW)*(PT‘3))
CranioCaudal_AMI=((1/12)*(PT)*(PW‘3))

if PT>PW

'ERRORllll! Plate width must be greater than plate thickness‘
end
if HD>PW

‘ERROR!!!!! Plate width must be greater than screw hole
diameter'
end

193

ILN AMI Calculation (FinalAMlcalc.m)

clear

clc

global x

%User input Diameter of ILN at nail hole level
D=input('Enter ILN diameter at nail hole level ');
%User input Diameter of ILN at solid level
DS=input('Enter Solid ILN Diameterz');

%User input Cis Cortex Hole Diameter
CC=input('Enter Cis Hole diameterz');

%User input Trans Cortex Hole Diamter
TC=input('Enter Trans Hole diameterz');

%Calculate the length of the rectangle and triangle
if CC>TC
StraightLength=[(D*2)/4-(cc‘2)/4]*<1/2)+[(n*2)/4-(Tc‘2)/4]*(1/2);
%Calculate the width of the triangle
TriangleWidth=(CC/2-TC/2);

%Calculate the width of the rectangle
RectangleWidth=(TC/2);

%Ca1culate the height of the top Semi-Circle

ycis=(((D‘2)/4)-((CCA2)/4))All/Z);
%Calculate the height of the bottom Semi-Circle
ytrans=(((DAZl/4)-((TCA21/4llAll/2);

%Calculate the Area of the Top Semi-Circle
TopTheta=2*[asin((CC/2)/(D/2))];
AreaTopSlice=(TopTheta/Z)*(D‘2)/4;
AreaTopTriangle=((D/2)*(cos(TopTheta/2)))*(CC/2);
AreaTopCircle=AreaTopSlice-AreaTopTriangle;

%Calculate the Area of the Bottom Semi-Circle
BotTheta=2*[asin((TC/2)/(D/2))];
AreaBotSlice=(BotTheta/Z)*(D‘2)/4;
AreaBotTriangle=((D/2)*(cos(BotTheta/2)))*(TC/2);
AreaBotCircle=AreaBotSlice-AreaBotTriangle;

%Calculate the Area of the Half Circle
AreaHalfCircle=pi*(D‘2)/8;

%Calculate the Area of the Triangle
AreaTriangle:(StraightLength)*(TriangleWidth)/2;

%Calculate the Area of the Rectangle
AreaRectangle=(StraightLength)*(RectangleWidth);

%Calcu1ate the Total Area of half of the Cross Section
AreaTotal=AreaHalfCircle-
(AreaTopCircle/2+AreaBotCircle/2+AreaTriangle+AreaRectangle);
%Calculate the centroid of the top Semi-Circle (origin is taken as
bottom

%of cross section, not center)
CentroidTopCircle=(D/2)+((((D‘2)*CC)/8)-(((YCis‘2)*CC)/2)-
((CC‘3)/24))/AreaTopCircle;

%Calculate the centroid of the bottom Semi-Circle (origin is bottom of
%cross section, not center)

CentroidBotCircle=(D/2)+(-
(((D”2)*TC)/8)+(((ytrans‘2)*TC)/2)+((TC‘3)/24))/AreaBotCircle;
%Calculate the centroid of the half circle
CentroidHalfCircle=D/2;

%Calcu1ate the centroid of the triangle
CentroidTriangle=(D/2-ytrans)+(StraightLength)*(2/3);

194

%Calculate the centroid of the rectangle
CentroidRectangle=(D/2-ytrans)+(StraightLength/2);

%Calculate the centroid of the entire cross section
SumIndividual:(AreaHalfCircle*CentroidHalfCircle)-
((AreaTopCircle/Z)*(CentroidTopCircle)+(AreaBotCircle/2)*(CentroidBotCi
rcle)+(AreaRectangle*CentroidRectangle)+(AreaTriangle*CentroidTriangle)
);

CentroidTotal:(SumIndividual/AreaTotal);

%Calculate "d" used in parallel axis theorem (distance from centroid to
NA)

d=D/2-(SumIndividual/AreaTotal);

%Calculate the slope of the line in Mediolateral Bending
m_medio=(-ycis-ytrans)/(-TC/2+CC/2);

%Ca1culate the y-intercept of the line in Mediolateral Bending
b_medio=ycis+(m_medio*CC/2);

%Equation of the line in Mediolateral Bending
y=(m_medio)*x+(b_medio);

% Define lower limit of x integral

xmin=-D/2;

% Define upper limit of x integral

xmax=-CC/2;

% Calculate ami
ami1=l/6*xmax*(xmin‘Z—xmax‘Z)*(3/2)+1/4*xmin‘2*xmax*(xmin‘z-
xmax‘Z)A(1/2)+l/4*xmin‘4*atan(xmax/(xmin‘Z-xmax‘2)2(l/2))—
l/8*i*xmin‘4*log(-i*xmin/(-2*xmin)“(l/2))+1/8*i*xmin‘4*log(i*xmin/(-
2*xmin)‘(l/2));

% Define lower limit of x integral

xmin2=-CC/2;

% Define upper limit of x integral

xmax2=—TC/2;

% Calculate ami2
ami2=1/12*m_medio‘3*xmax2‘4+l/3*m_medio‘2*xmax2‘3*b_medio+1/2*m_medio*x
max2‘2*b_medio“2+l/3*xmax2*b_medio”3+l/96*xmax2*(D‘2—
4*xmax2‘2)A(3/2)+1/64*D‘2*xmax2*(D‘Z-
4*xmax2‘2)A(1/2)+1/128*D‘4*atan(2*xmax2/(D‘2—4*xmax2“2)“(l/2))—
1/12*m_medio‘3*xmin2‘4-1/3*m_medio‘2*b_medio*xmin2‘3-
1/2*m_medio*b_medio‘2*xmin2‘2-1/3*b_medio‘3*xmin2-1/96*xmin2*(D‘Z-
4*xmin2‘2)‘(3/2)-1/64*D‘2*xmin2*(D‘2-4*xmin2‘2)*(1/2)-
1/128*D‘4*atan(2*xmin2/(D‘2-4*xmin2*2)“(i/2));

% Calculate final integral

final=amil+ami2;

%Compute Mediolateral AMI with parallel axis theorem taken into
consideration

MedioLateral_AMI=2*(final)—(2*AreaTotal)*(d‘2)

%Ca1culate the slope of the line in CranioCaudal Bending
m_cranio=(TC/2-CC/2)/(ycis+ytrans);

%Calculate the y-intercept of the line in CranioCaudal Bending
b_cranio=CC/2—(m_cranio*-ytrans);
%Equation of the line in Mediolateral Bending

y=(m_cranio)*x-(b_cranio);

ymin=(m_medio)*x+(b_medio);
ymax=((D/2)*2—x*2)*o.s;

195

CranioCaudal_AMI=2*(l/96*ytrans*(DA2-

4*ytrans:2) A (3/2) +1/64*DA2*ytrans*(DA2-

4*ytrans A2) (1/2) )+1/128*DA 4*atan(2*ytrans/( (D 2- -4*ytransA 2) (1/2)
1/12*m_cranioA3*ytransA4—1/3*m_cranioA2*ytransA3*b_cranio—
l/2*m_cranio*ytransA2*b_cranioA2-1/3*ytrans*b_cranioA3+1/96*ycis*(DA2-
4*ycisA2)A(3/2)+1/64*DA2*ycis*(DA2—

4*ycisA2) )A(1/2) )+1/128*DA4*atan(2*ycis/( (DA2—

4*ycis 2) (1/2) )+1/12*m_ cranio A3*ycisA 4-

1/3*m_cranioA 2*ycis A3*b:cranio+1/2*m_cranio*ycisA2*b_cranioA2-
1/3*ycis*b_cranioA3)

Solid_AMI=(pi*(DS/2)A4)/4
% Symbolic Manipulation of Integrals

°\0

Define symbolic variables
syms y x ymin ymax xmin xmax b m .....

0K° 0\° o\°

Assign ymin and ymax
ymin=-((D/2)A2—xA2)AO.5

0\° o\°

o\0

Inisde integration (with respect to y)
z=int(yA2,y,ymin,ymax)

0\° 0\°

Assign xmin and xmax

o\° 0‘" o\°

Outside integral (with resepct to x)
Z=int(z,x,xmin,xmax)

o\° o\o

Substitue true variable names

syms b_medio

%ZZ=subs(ZZ,b,b_medio)

end

if CC==TC

Mediolateral_AMI=2*(-1/96*CC*(D 2- cc 2) (3/2)-1/64*DA2*CC*(DA2-

c6 2) (l/2)-1/64*D 4*atan (cc/( D 2- cc 2) (1/2))-

1/128*i*D 4*log(i*D(1/2))+1/128*1*D 4*log(- i*DA( (1/2)))
Craniocaudal_AMI=2*((1/96*( DA 2- CCA 2) (1/2) (CCA 2) W3/2)+1/64*D 2*(DA 2-
c6 2) Ml/Z (cc 2) (1/2) -1/3*( (D 2- cc 2) (1/2)+1/64*D 4*atan((D 2-

c6 2) (1/2) /( cc 2) (1/2)

2: 2* (1/64*( (D 2) (3/2) /pil H1/2)/(-1/DA2)A(1/2 (2/3*piA( (1/2)*(DA2-

c6 2) 2/D 4- 2*p6 (1/2)*( (D 2-cc‘2) /DA2-1/2*((-2*log(2 -3/2-log(D 2- c6 2)-
log(-1/DA2))*piA(1/2)-1/12*pi W1/2 (D 2- cc 2) 2/D 4*(9/( D 2-
CCA2)A2*DA4-24/(DA2-CCA2)*DA2+8) -1/12*p6 Wl/Z (D 2-
CCA2)A2/DA4*(20/(DA2-CCA2)*DA2-8)*(l-l/(D 2—
CCA2)*DA2)A(l/2)+piA(l/2)*log(1/2+l/2*(l-l/(DAZ-CCAZ)*DA2)A(1/2)1)-
l/24*CCA3*(DA2-CCA2)A(l/2)l

zz =2*(l/64*(DA2)A(3/2) )/piA( 1/2) )/( -1/DA2)A( W1/2 (2/3*p1‘( 1/2)*(DA2-
CCA2)A2/DA4-2*pi Wl/Z (D 2- c6 2) /D 2-1/2*((-2*log(2 -3/2-log(D 2- c6 2)-
log(-l/DA2))*piA( (1/2) -1/12*p6 Wl/Z (D 2- c6 2) :Z/D 4* (9/( D 2-
CCAZ)A2*DA4-24/(DA2-CCA2)*DA2+8) -1/12*p6 (1/2) (D 2-
CCA2)A2/DA4*(20/(DA2-CCA2)*DA2—8)*(l-l/(D 2-
cc‘2)*D‘2)‘(1/2)+pi*(1/2)*1og(1/2+1/2*(1-1/(D‘2-CC‘2)*D‘2)‘(1/2)))-
1/24*cc‘3*(D‘2-CC‘2)‘(1/2))

a2 =-1/96*(l/4*DA2- 1/4*cc*2l (1/2) *(cc‘2)*( (3/2) -l/64*DA2*(1/4*DA2—
1/4*c6 2) (1/2)*( (cc 2) :(1/2)-1/128*D 4*atan( 2*( (1/4*D 2-

1/4*cc 2) (1/2) )/( cc 2) (1/2)) )+1/24*cc 3*(1/4*D 2-1/4*cc 2)‘(1/2)

o\° ok°

196

a3 =1/96*(1/4*DA2—l/4*CCA2)A(1/2)*(CCA2)A(3/2)+l/64*DA2*(l/4*DA2—
l/4*CCA2)A(l/2)*(CCA2)A(l/2)+l/128*DA4*atan(2*(1/4*DA2-
1/4*cc*2)‘(1/2)/(cc‘2)‘(1/2))-1/24*cc*3*(1/4*D*2-1/4*cc‘2)*(1/2);
Craniocauda1_AMI=2*(a3-a2)

Solid_AMI=(pi*(DS/2)A4)/4

End

197

CHAPTER ONE

The following is the compliance data (N/mm) for the HP and LC-DCP constructs which

APPENDIX FOUR

ADDITIONAL DATA

was briefly discussed but not included within the Chapter 1 data:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Specimen 100 N 200 N 300 N
Hybrid 1 0.0036 0.0050 0.0069
Hybrid 2 0.0015 0.0013 0.0013
Hybrid 3 0.0037 0.0046 0.0061
Hybrid 4
Hybrid 5 0.0025 0.0028 0.0034
Hybrid 6 0.0028 0.0031 0.0036
Hybrid 7
Hybrid 8 0,0054 0.0086 0.01 18
Hybrid 9 0.0011 0.0012 0.0012

AVERAGE 0.0031 0.0038 0.0049
STDEV 0.0018 0.0026 0.0037
LC-DCP 1 0.0067 0.0080 0.0101
LC-DCP 2 0.0014 0.0019 0.0020
LC-DCP 3 0.0059 0.0070 0.0091
LC—DCP 4
LC-DCP 5 0.0039 0.0056 0.0074
LC-DCP 6 0.0024 0.0034 0.0047
LC-DCP 7
LC-DCP 8 0.0094 0.0125 0.0146
LC-DCP 9 0.0013 0.0013 0.0013
AVERAGE 0.0044 0.0057 0.0070
STDEV 0.0030 0.0039 0.0047

 

 

 

 

Table A4-1. Compliance data for HP and LC-DCP constructs.

I98

 

The following are the exact p-values for the above compliance data:

HP vs LC-DCP
within 100 N: 0.047
within 200 N: 0.012
within 300 N: 0.007

Load within HP
300 vs 100: 0.009
300 vs 200: 0.051
200 vs 100: 0.189
Load within LC-DCP
300 vs 100: <0.001
300 vs 200: 0.020
200 vs 100: 0.031

The following are the exact p-values for the angular deformation data discussed in
Chapter 1:

HP vs LC-DCP
within 100 N: 0.310
within 200 N: 0.023
within 300 N: 0.002
Load within HP
300 vs 100: <0.001
300 vs 200: 0.010
200 vs 100: 0.047
Load within LC-DCP
300 vs 100: <0.001
300 vs 200: 0.003
200 vs 100: 0.016

The following are the exact p-values for the plate strain data discussed in Chapter 1:

HP vs LC-DCP
within 100 N: 0.226
within 200 N: 0.013
within 300 N: 0.001
Load within HP
300 vs 100: <0.001
300 vs 200: 0.004
200 vs 100: 0.013
Load within LC-DCP
300 vs 100: <0.001
300 vs 200: <0.001
200 vs 100: <0.001

199

Finite Element Analysis
Materials and Methods

To rationalize the hypotheses beyond mathematical calculations, a finite element
analysis was done to simulate the conditions experienced by the plates during pre-
bending. Based on the cross-sections of the plates where the bend occurred (solid
section), a model was created (Figure A) for each cross section in Autodesk Inventor

Professional (AutoCAD v.10).

 

Figure A4-l. Cross-sectional view of the solid section of the HP (left) and LC-DCP (right).

Powered by ANSYS® Design Space, a finite element mesh was created, boundary
conditions were implemented, and equivalent (von Mises) stresses, and resultant
deformations were solved for. Both plates were modeled as Stainless Steel, with
tetrahedral elements. The cross sections were extended to a length of 20 mm and
boundary conditions represented 3-p0int bending with opposite ends constrained to
restrict movement and a point load of 100 N was applied to the dorsal surface of the

plate.

200

Results

Cross sections of the solid section (Figure A4—1) of both plates were created in
Autodesk Inventor Professional and extended to an arbitrary length of 20 mm to give a 3-
dimensional representation of the region where the bend occurred. Applying a 100 N
point load to the dorsal surface of the beam and constraining the ends of the beam

simulated 3-point bending. Results of the simulation are shown in Figure A4-2 and A4—3.

 

 

 

9.03E—04 mm

Figure A4-2. FEA of the HP. Equivalent (von Mises) stresses are shown on the top (5:1 actual
scale), deformation is shown on the bottom (2:1 actual scale)

201

 

 

    
  

8.5-l().5 MPa

 

16.1E—04 mm

Figure A4-3. FEA of the LC-DCP. Equivalent (von Mises) stresses are shown on the top (5:1
actual scale). deformation is shown on the bottom (2:1 actual scale)

 

Under the given load of 100 N, the LC-DCP underwent larger deformations, and
larger magnitudes of stress in the area of interest, which is the dorsal surface of the plate.
The von Mises stresses on the surface of the HP were approximately 6.701 MPa, on the
surface of the LC-DCP they were 10.106 MPa, a difference of 34%. Maximum
deformation in the HP was found to be 9.03E—04 mm in the area of interest, whereas in

the LC-DCP it was 16.1E~04 mm, a difference of approximately 44%.

202

CHAPTER TWO
The following data includes results of the following tests:
Torsion

6 mm ILN with 2.7 mm bolts

6 mm ILN with ESF

6 mm ILN with ESF minus connecting bar

with p-values for pairwise comparisons
ML Bending

6 mm ILN with ESF minus connecting bar
CC Bending

6 mm ILN with ESF minus connecting bar
Compression

6 mm ILN with ESF minus connecting bar

203

Torsion

Tests
_ ,, Term. Comp. (2.5-5
Sample Intl. Comp. (0 1'5 N m) N*m) Slack Def.
+ - avg + - avg
GILT1 NA NA NA 1.17 1.16 1.16 5.43 17.1
63219 GILT2 NA NA NA 1.16 1.15 1.15 6.38 17.95
mm bolts GlLT3 NA NA NA 1.17 1.15 1.16 6.08 17.76
GILT4 NA NA NA 1.16 1.15 1.15 6.58 18.14
Mean NA NA NA 1.17 1.15 1.16 6.12 17.74
SD NA NA NA 0.01 0.01 0.01 0.5 0.45
GXF1T1 13.44

b 2.2 2.28 2.24 0.94 0.95 0.94 NA
6mm“-Ne.x1=1T2 1.18 1.94 1.56 0.83 0.84 0.83 NA 11.53

W, ESF GXF1T3 1.57 2 1.79 0.83 0.84 0.83 NA 11.69

GXF1T5 1.67 2.2 1.94 0.88 0.87 0.88 NA 12.12

Mean NA NA 1.88 0.87 0.87 0.87 NA 12.2

SD NA NA 0.29 0.05 0.05 0.05 NA 0.87

5mm"—“ GXFOT‘ NA NA NA 1.15 1.16 1.16 6.99 ”'58
w/ESF b

(minus GXFOT2 NA NA NA 1.08 1.06 1.07 5.45 16.19

connectin GXFOT3 NA NA NA 1.09 1.08 1.08 5.67 16.53

gbar) GXFOT5 NA NA NA 1.14 1.13 1.13 5.78 17.09

Mean NA NA NA 1.11 1.11 1.11 5.97 17.1

SD NA NA NA 0.04 0.04 0.04 0.69 1.06

Table A4-2. Torsion test data

Torsion
p-values
Terminal Compliance Slack Def.
ILN vs . ILN vs . ILN vs ' .
EXFIX no 0.0254 ”Qfizgfd EXFIX 0.3742 ”151::in EXFIX ”"pfzgfd t'
bar no bar no bar 0.1538
ILN vs . ILN vs .
EXFIX w/ 0.0000 “12:2? EXFIX w/ ““pggfd t‘
bar bar 0.0000
EXFIX EXFIX
w/bar vs 0 0001 paired t- w/bar vs paired t-
EXFIX no ' test EXFIX test
bar no bar 0.0000

Table A4-3. Torsion data p-values

204

 

ML
Bending
Tests

Sample

6 mm ILN w/ GXOml1a

ESF (minus GXOm|2a
connecting GXOm|4a
bar) GXOmISa
Mean
SD
CC
Bending
Tests
Sample

6 mm ILN w/ GXOCma

ESF (minus GXOcha
connecting GXOcc4a
bar) GXOccSa
Mean
SD
Compression
Tests
Sample

GXFOCZ
6 mm ILN w/ ESF GXFOCS

(minus connecting
bar) GXFOC4

GXFOCS

Mean
SD

Table A4-4. Mediolateral bending, Craniocaudal bending, and Compression test data

positive

0.775
0.814
0.713
0.657
0.740
0.069

Terminal Compliance ( 2.5-5 N‘m)

positive negative

1.025
1.100
1.024
1.114
1.066
0.048

0.720
0.852
0.626
0.762
0.740
0.094

1.205
1.280
1.186
1.231
1.226
0.041

6.300
5.500
8.100
6.300
6.550
1.100

Compliance Displacement

(mm/N*E-04) (mm‘E-02)

11.400
9.500
14.600
11.400
11.725
2.116

205

Terminal Compliance ( 2.5-5 N*m)

negative average

0.747
0.833
0.670
0.709
0.740
0.070

average

1.115
1.190
1.105
1.173
1.146
0.042

Slack

5.784
4.969
5.981
5.595
5.582
0.438

Slack

1.897
1.529
2.062
1.521
1.752
0.271

Overall
Def.

10.998
10.764
10.656
10.530
10.737
0.199

Overall
Def.

9.702
9.864
9.828
9.702
9.774
0.084

CHAPTER THREE
The following data are included, pairwise comparisons follow each set of data
Torsion

ILN6s (50 mm fracture gap)

ILN6b (50 mm fracture gap)

ILNSS (50 mm fracture gap)

ILN8B (50 mm fracture gap)

3.5 mm br-DCP (50 mm and 100 mm fracture gap)
ILNn (50 mm fracture gap)

Mediolateral Bending

ILN6s (50mm and 100mm fracture gap)

ILN6b (50mm and 100mm fracture gap)

ILN8S (50mm and 100mm fracture gap)

ILN8B (50mm and lOOmm fracture gap)

3.5 mm br-DCP (50mm and 100mm fracture gap)
ILNn (50mm and 100mm fracture gap)

Craniocaudal Bending

ILN6s (50mm and 100mm fracture gap)

ILN6b (50mm and 100mm fracture gap)

ILN 8S (50mm and 100mm fracture gap)
ILNSB (50mm and 100mm fracture gap)
3.5 mm br-DCP (50mm and 100mm fracture gap)

ILNn (50mm and 100mm fracture gap)

206

 

Torsion

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Moment: 5 TC Range:
Tests Nm :2 to :5
Avg
(+) TC (-) TC (x10- TC(x10-
ILN63 (x10-2) 2) 2) IC (x10-2) Max def Slack (y) Comments
66t1 1.758 1.701 1.729 NA 39.96 22.52
66t2 1.676 1.638 1.657 NA 37.1 20.34
6st3 1.635 1.596 1.616 NA 34.32 18.09
66t4 1.626 1.597 1.612 NA 34.39 18.20
MEAN 1.674 1.633 1.653 36.44 19.79
SD 0.060 0.049 0.055 2.68 2.10
Avg
(+) TC (-) TC (x10- TC(x10-
ILN6b (x10-2) 2) 2) IC (x10-2) Max def Slack (y) Comments
From Scott
GILT2 1.158 1.145 1.152 NA 17.95 6.38 Goet's data
From Scott
GILT4 1.156 1.147 1.152 NA 18.14 6.58 Goet's data
6th 1.094 1.089 1.092 NA 18.15 7.18
66t4 1.092 1.086 1.089 NA 18.2 7.26
MEAN 1.125 1.117 1.121 18.11 6.85
SD 0.037 0.034 0.035 0.11 0.43
Avg
(+) TC (-) TC (x10- TC(x10-
ILN83 (x10-2) 2) 2) IC (x10-2) Max def Slack (y) Comments
88t1 0.717 0.713 0.715 NA 23.90 16.71
88t2 0.719 0.737 0.728 NA 24.47 17.16
EanyJen
data
Computer
file name
an853 0.854 0.742 0.798 NA 23.57 15.58 3555at
Early Jen
data
Computer
file name
an8s4 0.841 0.742 0.792 NA 23.77 15.81 3566at
MEAN 0.783 . 0.734 0.758 23.93 16.31
SD 0.075 0.014 0.043 0.39 0.74

 

 

 

 

 

 

 

 

 

 

207

 

 

Avg

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(+) TC (-) TC (x10- TC(x10-
lLNBb (x10-2) 2) 2) IC (x10-2) Max def Slack (y) Comments
EadyJen
data file
an8b1 0.858 0.699 0.779 NA 13.20 5.45 name 3562ct
EaflyJen
data file
an8b2 0.816 0.764 0.790 NA 13.58 5.77 name 3564ct
88t3 0.586 0.594 0.590 NA 11.44 5.52
88t4 0.567 0.578 0.573 NA 11.41 5.66
MEAN 0.707 0.659 0.683 12.41 5.60
SD 0.152 0.088 0.118 1.14 0.14
Avg
(+) TC (-) TC (x10- TC(x10-
ILNn (x10-2) 2) 2) IC (x10-2) Max def Slack (y) Comments
ant1 0.883 0.883 0.883 NA 8.60 None
ILnt2 0.851 0.837 0.844 NA 8.14 None
ILnt3 0.898 0.844 0.871 NA 8.13 None
ant4 0.890 0.856 0.873 NA 8.30 None
MEAN 0.881 0.855 0.868 8.29
SD 0.021 0.020 0.017 0.22
Avg
(+) TC (-) TC (x10- TC(x10-
3.5erCP (x10-2) 2) 2) IC (x10-2) Max def Slack (y) Comments
Pt1 1.227 1.180 1.204 NA 11.32 None
Pt2 1.358 1.284 1.321 NA 12.18 None
Pt3 1.363 1.337 1.350 NA 12.47 None
Pt4 1.246 1.231 1.239 NA 1 1.74 None
MEAN 1.299 1.258 1.278 11.93
SD 0.072 0.068 0.069 0.50
loss of
Ptsh1 1.793 1.763 1.778 NA 16.74 alignment
loss of
Ptsh2 1.854 1.888 1.871 NA 17.36 None aliment
Ptsh3 1.739 1.715 1.727 NA 16.36 None
Ptsh4 1.861 1.846 1.854 NA 17.25 None
MEAN 1.812 1.803 1.807 16.93 #DlV/O!
SD 0.057 0.078 0.067 0.46 #DIV/O!

 

Table A4-5. Torsion test data

208

 

 

Torsion - Terminal Compliance p-values

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ILN6s ILN6b [LN 88 [LNSB br-DCP ILNn
ILN6s NA p < 0.001 p < 0.001 p < 0.001 p < 0.001 p < 0.001
ILN6b p < 0.001 NA p < 0.001 p < 0.001 p = 0.003 p < 0.001
ILNSS p<0.001 p<0.001 NA p=0.115 p<0.001 p=0.027
ILN8B p<0.001 p<0.001 p=0.115 NA p<0.001 p=0.002
br-DCP p < 0.001 p = 0.003 p < 0.00l p < 0.001 NA p < 0.001
ILNn p < 0.001 p < 0.001 p = 0.027 p = 0.002 p < 0.001 NA
Torsion - Angular Deformation p-values
ILN6s ILN6b ILN8S ILNSB br-DCP ILNn
ILN6s NA p < 0.001 p < 0.001 p < 0.001 p < 0.001 p < 0.001
ILN6b p < 0.001 NA p < 0.001 p < 0.001 p < 0.001 p < 0.001
.[LN8S p < 0.001 p < 0.00] NA p < 0.00] p < 0.001 p < 0.00l
ILN8B p < 0.001 p < 0.001 p < 0.00] NA p = 0.585 p < 0.00]
br-DCP p < 0.001 p < 0.001 p < 0.001 p = 0.585 NA p < 0.001
ILNn p<0.001 p<0.00l p<0.00l p<0.001 p<0.001 NA
Torsion - Slack p-values
ILN6s ILN6b ILN8S ILN8B br-DCP ILNn
ILN6s NA p < 0.001 p < 0.001 p < 0.001 p < 0.001 p < 0.001
ILNob p < 0.001 NA p < 0.001 p = 0.073 p < 0.001 p < 0.001
ILNSS p < 0.00! p < 0.00l NA p < 0.001 p < 0.001 p < 0.001
ILN8B p < 0.001 p = 0.073 p = 0.115 NA p < 0.001 p < 0.001
br-DCP p < 0.001 p < 0.001 p < 0.001 p < 0.001 NA p = 1
ILNn p<0.001 p<0.001 p<0.00l p<0.001 p=l NA
Torsion - Angular Deformation minus Slack p-values
ILN6s ILN6b [LN 88 ILNSB br-DCP ILNn
ILN6s NA p < 0.001 p < 0.001 p < 0.001 p < 0.001 p < 0.001
ILN6b p<0.001 NA p<0.00l p<0.001 p=0.138 p<0.001
ILNSS p < 0.001 p < 0.001 NA p = 0.077 p < 0.001 p = 0.132
ILN8B p < 0.001 p < 0.001 p = 0.077 NA p < 0.001 p = 0.008
br-DCP p<0.001 p =0.l38 p<0.00l p<0.00l NA p<0.001
ILNn p<0.00l p<0.001 p=0.l32 p=0.008 p<0.001 NA

 

 

 

 

 

 

 

 

Tables A4-6 thru A4-9. Torsion data p-values.

209

 

 

 

 

 

 

Bending

Moment: 3.5 Nm

 

50mm gap - 10mm

 

 

120mm gap 14.5mm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Tests TC Range: +/-1.5-3.5 bone diameter ,
bone diameter
ILNGSML (+) TC (x10-2) (-) TC (x10-2) AngC(x10-2) lC(x10-2) Ang def Slack Comments
ML first, CC
68MLb1 0.483 0.636 0.5595 NA 12.62 8.69 second,
interference
CC first, ML
63MLb2 0.533 0.684 0.6085 NA 13.09 8.82 second,
interference
ML first, CC
68MLb3 0.556 0.67 0.613 NA 14.9 10.61 second,
interference
CC first, ML
GSMLb4 0.479 0.624 0.5515 NA 14.86 11.05 second,
interference
MEAN 0.51 0.65 0.58 13.87 9.79
SD 0.04 0.03 0.03 1.18 1.21
- CC first, ML
63M?“ 0.843 1.16 1.0015 NA 21.6 14.53 second, no
interference
ML first, CC
second,
BSMrszs 0.842 1.05 0.946 NA 21.29 14.67 huge slack,
no
interference
CC first, ML
63Mr'1‘b35 0.725 1.253 0.989 NA 25.92 18.93 second, no
interference
68MLb4s no
h 0.753 1.15 0.9515 NA 22.68 16.03 interference
MEAN 0.79 1.15 0.97 22.87 16.04
SD 0.06 0.08 0.03 2.12 2.04
ILNBsCC (+) TC (x10-2) (-) TC (x10-2) Avg TC(x10-2) lC(x10-2) Max def Slack Comments
ML first, CC
68CCb1 0.64 0.88 0.76 NA 12.58 3.07 second,
interference
CC first, ML
BSCCbZ 0.51 0.604 0.557 NA 12.33 3.21 second,
interference
ML first, CC
68CCb3 0.499 0.608 0.5535 NA 12.29 3.25 second,
interference
CC first, ML
SSCCb4 0.559 0.669 0.614 NA 12.55 3.19 second,
interference
MEAN 0.55 0.69 0.62 12.44 3.18

 

 

 

 

 

 

 

 

210

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

so 0.06 0.13 0.10 0.15 0.08
,_ . first.
630$”; 1.603 2.045 1.824 NA 16.94 . 4.2 second, no
. 3 ‘ uterferenc.
_ 1 Lfirst,C-
isscpbzs 2.066 2.789 2.4275 NA 3 18.66 i 2.7 Fecondmo
. 7 g g - , _ _ = _ J interferenc-
15%:th 1.945 2.51 2.2275 NA ‘ 18.61
630$)“ 1.787 2.117 1.952 NA 17.15
MEAN 1.85 2.37 2.11 17.84
so 0.20 0.35 0.27 0.92
ILNGbML (+) TC (x10-2) (-) TC (x10-2) Avg TC(x10-2) lC(x10-2) Max def Slack Comments
GlLBml2a 0.603 0.775 0.689 11.3 6.44
CCfirst, ML
68MLb2b 0.558 0.689 0.6235 NA 11.27 6.96 second
ML first, CC
eBMLba 0.754 0.871 0.8125 NA 12.96 7.3 second
GlLBmISa 0.69 0.75 0.72 11.54 6.53
MEAN 0.65 0.77 0.71 11.77 6.81
so 0.09 0.08 0.08 0.80
55M?” 0.766 0.899 0.8325 NA 12.49
58M?” 0.786 1.013 0.8995 NA 12.37
EBMESS 0.739 0.935 0.837 NA 14.67
63M?“ 0.768 0.887 0.8275 NA 12.38
MEAN 0.76 0.93 0.85 12.98 7.02
so 0.02 0.06 0.03 1.13 1.22
lLN6bCC (+) TC (x10-2 (-) TC (x10-2) Avg TC(x10-2) 1c (x10-2) Max def Slack Comments
GlLBcc2a 1.14 1.24 1.19 10.52 2.24
cc first. ML
6BCCb2b 1.1 1.276 1.188 NA 10.51 2.17 second
ML first, CC
GBCCbS 1.168 1.273 1.2205 NA 10.85 2.38 second
GlLBcc3a 1.13 1.286 1.208 10.57 2.14 ‘

 

211

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MEAN 1.13 1.27 1.20 10.61 2.23
so 0.03 0.02 0.02 0.16 0.11
00 first, ML
63C?” 1.182 1.301 1.2415 NA 11.14 2.44 second, No
interference
7 ML first, CC
GBCEbZS 1.114 1.294 1.204 NA 10.64 2.17 second, No
interference
CC first, ML
GBCEDSS 1.175 1.327 1.251 NA 11.05 2.26 second, No
interference
ML first, 00
6305b“ 1.161 1.335 1.248 NA 10.12 1.38 second, No
interference
MEAN 1.16 1.31 1.24 10.74 2.06
so 0.03 0.02 0.02 0.47 0.47
1LN8sML (+) TC (x10-2) (-) TC (x10-2) Avg TC(x10-2) 10 (x10-2) Max def Slack Comments
ML first, 00
BSMLb1 0.347 0.426 0.3865 NA 12.24 9.61 second
CC first, ML
88ML020 0.446 0.514 0.48 NA 10.84 7.48 second
00 first, ML
88MLb3 0.344 0.447 0.3955 NA 12.58 9.81 second
ML first, 00
BSMLb4 0.453 0.442 0.4475 NA 12.35 9.21 second
MEAN 0.40 0.46 0.43 12.00 9.03
so 0.06 0.04 0.04 0.79 1.06
ML first, 00
83M?“ 0.433 0.575 0.504 NA 17.73 14.18 second, no
interference
file name
88ML020 88MLb2sh,
sh 0.427 0.553 0.49 NA 18.41 14.96 00 first, ML
second
88ML03s ML first, 00
h 0.478 0.553 0.5155 NA 18.31 14.68 second
85MLb4s CC first, ML
h 0.453 0.561 0.507 NA 17.89 14.33 second
MEAN 0.45 0.56 0.50 18.09 14.54
SD 0.02 0.01 0.01 0.33 0.35
ILNBsCC (+) TC (x10-2) (-) TC (x10-2) Avg TC(x10-2) lC(x10-2) Max def Slack Comments
BSCCb1 0.314 0.345 0.3295 NA 7.4 5.09
880002 0.309 0.337 0.323 NA 7.1 1 4.84
88CCb3 0.314 0.345 0.3295 NA 7.34 5.04
88CCb4 0.366 0.4 0.383 NA 7.2 4.52

 

 

 

212

 

 

 

 

 

 

 

 

 

 

 

 

 

MEAN 0.33 0.36 0.34 7.26 4.87
so 0.03 0.03 0.03 0.13 0.26
MLfirst,CC
BSCEMS 0.953 1.07 1.0115 NA 12.96 5.83 second, no
interference
BSCCbZS CC first, ML
h 0.939 1.129 1.034 NA 12.64 5.38 second
88CCb3s ML first, 00
h 0.799 0.99 0.8945 NA 12.15 5.86 second
8SCCb4s CC first, ML
h 0.816 0.93 0.873 NA 11.90 5.73 second
MEAN 0.88 1.03 0.95 12.41 5.70
so 0.08 0.09 0.08 0.48 0.22

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ILN8bML (+) TC (x10-2) (-) TC (x10-2) Avg TC(x10-2) lC(x10-2) Max def Slack Comments
ML first, CC
second,
88MLb1 0.41 0.466 0.438 NA 8.06 5.12 i nteref 6 re n c
e?
ML first, CC
second,
BBMLbe 0.331 0.41 0.3705 NA 6.34 3.73 i nte refe re no
6?
ML first, CC
88MLb3b 0.321 0.469 0.395 NA 6.95 4.18 second
ML first, CC
8BMLb4b 0.348 0.414 0.381 NA 6.91 4.25 second
MEAN 0.35 0.44 0.40 7.07 4.32
SD 0.04 0.03 0.03 0.72 0.58
8BMLb1s ML first, CC
h 0.458 0.535 0.4965 NA 9.31 5.8 second
file name
8BML02b 8BML02sh,
sh 0.479 0.617 0.548 NA 9.54 5.73 CC first, ML
second
file name
BBMLbSb 8BML03sh,
sh 0.487 0.57 0.5285 NA 9.25 5.57 ML first, CC
second
file name
88MLb4b 88MLb4sh,
sh 0.54 0.566 0.553 NA 9 5.16 CC first, ML
second
MEAN 0.49 0.57 0.53 9.28 5.57
SD 0.03 0.03 0.03 0.22 0.29
ILNBbCC (+) TC (x10-2) (-) TC (x10-2) Avg TC(x10-2) IC (x10-2) Max def Slack Comments
ML first, CC
88CCb1 0.305 0.575 0.44 NA 5.99 2.91 second

 

 

 

213

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CC first, ML
880002 0.575 0.416 0.4955 NA 6.5 3.04 second
ML first, CC
8800030 0.452 0.573 0.5125 NA 6.73 3.2 second
CC first, ML
880004 0.638 0.44 0.539 NA 6.21 2.47 second
MEAN 0.49 0.50 0.50 6.36 2.91
so 0.15 0.08 0.04 0.32 0.31
8305b” 0.72 0.798 0.759 NA 7.45 2.12 I‘M-$23150
BBCCbZS CC first, ML
h 0.663 0.692 0.6775 NA 7.18 2.41 second
file name
8800030 880003sh,
sh 0.671 0.76 0.7155 NA 7.52 2.52 MLfirst,CC
second
BBCCb4s CC first,
h 0.617 0.722 0.6695 NA 6.91 2.22 ML second
MEAN 0.67 0.74 0.71 7.27 2.32
so 0.04 0.05 0.04 0.28 0.18
Br-DCP Av IC Max Slac
ML (+) TC (x10-2) (-) TC (x10-2) TC(X1%_2) (X102) def k Comments
PMLb1 0.919 0.954 0.9365 NA 6.23 NS" Beaut
PML02 0.887 0.908 0.8975 NA 6.06 ”2" Beaut
PMLb3 0.901 0.955 0.928 NA 6.14 ”2" Beaut
PMLb4 0.883 0.929 0.906 NA 6.16 ”2" Beaut
MEAN 0.90 0.94 0.92 6.15
so 0.02 0.02 0.02 0.07
PMLb1sh 1.351 1.462 1.4065 NA 9.56 NS”
PMLb2sh 1.375 1.472 1.4235 NA 9.59 NS"
PMLb3sh 1.356 1.456 1.406 NA 9.63 ”2"
PM?“ 1.349 1.456 1.4025 NA 9.54 ”2"
MEAN 1.36 1.46 1.41 9.58
so 0.01 0.01 0.01 0.04
Br-DCP Avg IC Max Slac
CC (+) TC (x10-2) (-) TC (x10—2) TC(x10-2) (x10-2) def k Comments
P0001 0.225 0.229 0.227 NA 1.49 ”3” Beaut
P0002 0.223 0.231 0.227 NA 1.53 ”3" Beaut

 

 

214

 

 

Non

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ILNn ML

(+) TC (x10-2)

(-) TC (x10-2)

Avg
TC(x10-2)

IC
(x10-2)

Max
def

Slac
k

P0003 0.217 0.222 0.2195 NA 1.48 e Beaut
P0004 0.222 0.225 0.2235 NA 1.51 ”2" Beaut
MEAN 0.22 0.23 0.22 1.50

so 0.00 0.00 0.00 0.02
P0001511 0.351 0.363 0.357 NA 2.11 ”1"“

’ ’ i ,
P0002sn 0.345 0.361 0.353 _ NA 2.12; “ml
PCCbash 0.345 0.345 0.345 NA , 2.11
'66 b4“ 0.454 0.464 0.459 NA 2.68 NO”

MEAN 0.37 0.38 0.38 2.26

so 0.05 0.05 0.05 0.28

 

Comments

 

ILnMLb1

0.583

0.553

0.568

NA

3.87

No

3

original file
name
40n1cb,
new file
name has
letter "b"

 

ILnMLb2

0.697

0.657

0.677

NA

4.36

No

3

original file
name
40n3ab,
new file
name has
letter "b"

 

anMLb3

0.602

0.62

0.611

NA

4.07

No

(D
3

original file
name
40mediab,
new file
name has
letter "13"

 

|LnMLb4

0.575

0.544

0.5595

NA

3.8

Non

(‘0

original file
name
40n6cb,
new file
name has
letter "b"

 

MEAN

0.61

0.59

0.60

4.03

 

SD

0.06

0.05

0.05

0.25

 

 

ILnMLb1, ‘
h

 

0.559

 

0.586

 

0.5725

NA

 

 

‘ 3.91

 

Non

1 [1131mm
Mame doeé

. nothave

 

 

215

 

 

 

Instron file

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

anMLsz Non name does
h 0.563 0.589 0.576 NA 3.91 e not h ave
llhll
lnstronfile
ILnMLb35 Non name does
h 0.563 0.597 0.58 NA 3.98 e not have
uh"
file name
anMLb4s Non does not
h 0.55 0.582 0.566 NA 3.89 e have the
”h.
MEAN 0.56 0.59 0.57 3.92
SD 0.01 0.01 0.01 0.04
Avg IC Max Slac
ILNn CC (+) TC (x10-2) (-) TC (x10-2) TC(x10-2) (x10-2) def k Comments
anCCb1 0.821 0.865 0.843 NA 5.38 ”2”
Non
ILnCCbZ 0.756 0.81 0.783 NA 5.06 e
1Ln0003 0.805 0.852 0.8285 NA 5.24 N3"
anCCb4 0.804 0.857 0.8305 NA 5.256 ”2"
MEAN 0.80 0.85 0.82 5.23
SD 0.03 0.02 0.03 0.13
file name
anCCb1s Non does not
h 0.813 0.879 0.846 NA 5.4 e have the
Nb!
file name
anCCsz Non does not
h 0.76 0.805 0.7825 NA 5 e have the
uh“
file name
anCCb3s Non does not
11 0.796 0.849 0.8225 NA 5.2 e have the
Uh.
file name
anCCb4s Non does not
h 0.8 0.86 0.83 NA 5.27 e have the
Rh.
MEAN 0.79 0.85 0.82 5.22
SD 0.02 0.03 0.03 0.17

 

 

 

 

 

 

 

 

Table A4-10. Mediolateral and craniocaudal bending data (long and short specimens).

216

 

Mediolateral Bending - Compliance p-values (50 mm ggp only)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ILN6s ILN6b ILN8S ILN8B br-DCP ILNn
ILN6s NA p = 0.003 p < 0.001 p < 0.001 p < 0.001 p = 0.541
ILN6b p = 0.003 NA p < 0.001 p < 0.001 p < 0.001 p = 0.005
ILN8S p < 0.001 p < 0.001 NA p = 0.36 p < 0.001 p < 0.001
ILN8B p < 0.001 p < 0.001 p = 0.36 NA p < 0.001 p < 0.001
br-DCP p < 0.001 p < 0.001 p < 0.001 p < 0.001 NA p < 0.001
ILNn p = 0.541 p = 0.005 p < 0.001 p < 0.001 p < 0.001 NA
Mediolateral Bending- Angular Deformation p-values (50 mm gap on] )
H.N6s ILN6b ILN8S ILN8B br-DCP ILNn
ILN6s NA p = 0.002 p = 0.002 p < 0.001 p < 0.001 p < 0.001
ILN6b p = 0.002 NA p = 0.657 p < 0.001 p < 0.001 p < 0.001
1LN88 p = 0.002 p = 0.657 NA p < 0.001 p < 0.001 p < 0.001
ILN8B p < 0.001 p < 0.001 p < 0.001 NA p = 0.095 p < 0.001
br-DCP p < 0.001 p < 0.001 p < 0.001 p = 0.095 NA p < 0.001
ILNn p<0.001 p<0.001 p<0.001 p<0.001 p<0.001 NA
Mediolateral Bending - Slack p-values (50 mm ap only)
ILN6s ILN6b ILN 88 ILNBB br-DCP ILNn
ILN6s NA p<0.001 p=0.149 p<0.001 p<0.001 p=1
ILN6b p < 0.001 NA p < 0.001 p < 0.001 p < 0.001 p < 0.001
[LN8S p=0.149 p<0.001 NA p<0.001 p<0.001 p<0.001
ILN8B p < 0.001 p < 0.001 p < 0.001 NA p < 0.001 p < 0.001
br-DCP p<0.001 p<0.001 p<0.001 p<0.001 NA p<0.001
ILNn p<0.001 p<0.001 p<0.001 p<0.001 p=l NA
Mediolateral Bendin - Angular Deformation minus Slack (50 mm gap only)
ILN6s ILN6b ILNSS ILN8B br-DCP ILNn
ILN6s NA p < 0.001 p < 0.001 p < 0.001 p < 0.001 p = 0.821
ILN6b p < 0.001 NA p < 0.001 p < 0.001 p < 0.001 p = 0.001
ILNSS p < 0.001 p < 0.001 NA p = 0.305 p < 0.001 p < 0.001
ILN8B p < 0.001 p < 0.001 p = 0.305 NA p < 0.001 p < 0.001
br-DCP p < 0.001 p < 0.001 p < 0.001 p < 0.001 NA p < 0.001
ILNn p=0.821 p=0.001 p<0.001 p<0.001 p<0.001 NA

 

 

 

 

 

 

 

 

Table A4-ll thru A4-l4. Mediolateral bending data p-values

217

 

 

 

 

Craniocaudal Bending — Compliance p-values (50 mm gap only)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ILN6s ILN6b ILNSS ILN8B br-DCP ILNn
ILN6s NA p<0.001 p<0.001 p=0.001 p<0.001 p<0.001
ILN6b p < 0.00] NA p < 0.001 p < 0.001 p < 0.001 p < 0.001
ILNSS p < 0.001 p < 0.001 NA p < 0.001 p = 0.002 p < 0.001
ILN8B p=0.001 p<0.001 p<0.001 NA p<0.001 p<0.001
br-DCP p < 0.001 p < 0.001 p = 0.002 p < 0.001 NA p < 0.001
ILNn p < 0.001 p < 0.001 p < 0.001 p < 0.001 p < 0.001 NA

 

Craniocaudal Bending - Angular Deformation

-values (50 mm gap only)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ILN6s ILN6b ILNSS ILN8B br-DCP ILNn
ILN6s NA p<0.001 p <0.001 p<0.001 p<0.001 p<0.001
ILNob p < 0.001 NA p < 0.001 p < 0.001 p < 0.001 p < 0.001
ILN8S p<0.001 p<0.001 NA p<0.001 p<0.001 p<0.001
ILN8B p<0.001 p<0.001 p<0.001 NA p<0.001 p<0.001
br-DCP p < 0.001 p < 0.001 p < 0.001 p < 0.001 NA p < 0.001
ILNn p <0.001 p<0.001 p<0.001 p<0.001 p<0.001 NA

Craniocaudal Bending - Slack p-values (50 mm gap only)

ILN6s ILN6b [LN 8S ILN8B br-DCP ILNn
[LN6s NA p < 0.001 p < 0.001 p = 0.039 p < 0.001 p = 0.001
ILN6b p < 0.001 NA p < 0.001 p < 0.001 p < 0.001 p < 0.001
ILN8S p < 0.001 p < 0.001 NA p < 0.001 p < 0.001 p < 0.001
ILN8B p = 0.039 p < 0.001 p < 0.001 NA p < 0.001 p < 0.001
br-DCP p < 0.001 p < 0.001 p < 0.001 p < 0.001 NA p =
ILNn p<0.001 p<0.001 p<0.001 p<0.001 p=1 NA

 

 

 

 

 

 

 

 

Craniocaudal Bending — Ang. Deformation minus Slack p-values (50 mm gap only)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ILN6s ILN6b ILNSS ILN8B br-DCP ILNn
ILN6s NA p<0.001 p<0.001 p<0.001 p<0.001 p<0.001
ILN6b p <0.001 NA p<0.001 p<0.001 p<0.001 p<0.001
ILNSS p<0.001 p<0.001 NA p<0.001 p<0.001 p<0.001
[LNSB p<0.001 p<0.001 p<0.001 NA p<0.001 p<0.001
br-DCP p < 0.001 p < 0.001 p < 0.001 p < 0.001 NA p < 0.001
ILNn p<0.001 p<0.001 p<0.001 p<0.001 p<0.001 NA

 

Table A4-15 thru A4-18. Craniocaudal bending data p-values

218

 

 

 

 

 

__‘_=___r.,