Design and mechanical behavior of the MD series of bone dowels

MISSING IMAGE

Material Information

Title:
Design and mechanical behavior of the MD series of bone dowels
Physical Description:
xii, 139 leaves : ; 29 cm.
Language:
English
Creator:
Bianchi, John R., 1965-
Publication Date:

Subjects

Subjects / Keywords:
Materials science and Engineering thesis, Ph. D   ( lcsh )
Dissertations, Academic -- Materials science and Engineering -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1999.
Bibliography:
Includes bibliographical references (leaves 120-131).
Statement of Responsibility:
by John R. Bianchi.
General Note:
Printout.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 021583927
oclc - 43741979
System ID:
AA00017681:00001


This item is only available as the following downloads:


Full Text











DESIGN AND MECHANICAL BEHAVIOR OF THE MD SERIES OF BONE
DOWELS

















By

JOHN R. BIANCHI


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1999


























This dissertation is dedicated to Dominick and Jean Bianchi, who raised me to fill the
space between my ears; to my brother, Michael, who has somehow managed a great
excuse in NOT going to another graduation; and to my wife, Laurie, who has managed to
find herself in another. One of the most unselfish acts one human may do for the other is
to share the gift of organ and tissue donation. This research is dedicated to the memory
of the donors and their families. For if it had not been for them, this research would
never have been possible.














ACKNOWLEDGMENTS

I wish to thank Dr. Chester E. Sutterlin III, MD, for teaching me about spinal

surgery and how to apply my knowledge of engineering to its practice. I thank Jamie

Grooms and Regeneration Technologies Inc for financial support of this project and

having the stamina to ensure that it was carried out. I am grateful to John J. Mecholsky,

Jr. for his patience and judgement, and Christopher Batich for encouraging me to begin

my studies at UF. I would like to thank my Doctoral committee, Anthony Brennan,

Edward Walsh, David Greenspan, and Andrew Rapoff for ensuring that this research was

carried out to the highest of standards. I am thankful to C.Randal Mills for his intellectual

conversations, and James E. Keesling Jr, an Engineer who crossed the line and became a

Statistician, assisting with many of the statistical analysis. I would to thank Chris

Glymph for setting this research to practice, Kevin Carter for his most valuable technical

assistance, Kevin Ross for helping out with the mechanical testing, and Sanna Martinez-

Saare for her artistic sketches. But most of all I would like to thank and acknowledge my

wife for her patience, pep talks, comic relief, and love.

I would like to thank many of my family and friends who insured that I had a

good time including the founding members of the Thundering Amoebea Investment

Club: Matt Corrigan, Mike Rickey, Doug Mullin, Dan Garlington, Will Lyons, and Mike

Lyons.














TABLE OF CONTENTS
page


ACKNOW LEDGM ENTS .............................................................................................iii

LIST OF TABLES ............................................................................ ......................... vii

LIST OF FIGURES ..................................................................................................... ix

INTRODUCTION ........................................................................................................ 1

REVIEW OF THE LITERATURE ................................................................................. 5

The Function and Anatomy of Bone .......................................................................... 5
M echanics of Bone .................................................................................................. 7
Historical Development of Bone as a Transplantable M aterial........................... ... 14
Indications for Spine Surgery .................................................... ......................... 19
Surgery of the Spine...................................................................... ...................... 20
Spinal M echanics ................................................................................................... 20
M echanics of Spine Fusion...................................................... ........................... 21
Biology of Spine Fusion ....................................................................................... 21
Synthetic Implants Used for Spinal Fusion ......................................... ........... ...... ... 22
Bone Grafts Utilized in Spine Surgery ................................................ ............. 23
Interbody Autografts or Allografts for Spine Surgery.......................................... 23
Structural Properties of Allograft Interbodies for Spinal Fusion...................... 38
Alternative M materials to Bone.............................................................. ........ .... 39
Alternatives to Spinal Fusion................................................... ........................... 40
Summary ..................................................................................................................... 40

LOCAL STRESS FIELD............................................................................................ 42

Introduction........................................................................................................... 42
M materials and M ethods.......................................................................................... 42
Results............................... ................................................................................. 46
Hertzian Contact Between Concentric Cylinders .............................................. 46
In-Vivo Loading Conditions............................................. ............................. 46
Discussion............................................. ........................ ........................................ 49
Conclusions....................................................................................................... 51

NORM ALIZED STRENGTH .................................................................................... 52

Structural Characteristics of Bone ............................................................................ 52









Structural Characteristics of Bone Dowels........................................ .......... ... 54
Theoretical Stress Analysis................................................................................... 55
M odes of Loading ................................................................................................ 55
M materials and M ethods... ........................................................................................ 56
Specimen Preparation ..................................................................................... ... 56
M echanical Testing ........................................................................................... 56
Physical M easurements......................................................................... ............. 58
Data Analysis ........................................................................................................ 58
Normalized Strength ......................................................................... ............... 59
Results................................................................................................................... 60
Failure Load of M D-II Bone Dowels............................. .............. .............. 60
Failure Loads of M D-II Dowels as a Function of Diameter.............................. 61
Failure Load as a Function of Dimensions ...................................... .......... ... 61
Failure Loads as a Function of Density, Hardness, or Ultrasonic Velocity.......... 67
Analyzing the Data by M ultivariable Regression Analysis................................ 70
Applicability to the Bone Dowels......................................................................... 75
Conclusions........................................................................................................... 75

BIOMECHANICAL QUALITY CONTROL PROCEDURE.......................................... 77

Introduction........................................................................................................... 77
M materials and M ethods............................................................................................... 77
Normalized Strength of Bone Dowels ........................................................... ... 77
Acceptable Quality Criteria ...................................................................................... 78
Design Loads on the Dowels ........................................................... .......... 78
Probability of Survival....................................................................................... 78
W eibull Approach Applied to the Bone Dowels ........................................................ 79
Validation.............................................................................................................. 80
Specimen Preparation ...................................................................................... 81
Sample Size..................................................................................................... 81
Nondestructive and Destructive Analysis....................................... ............ .. 81
Results................................................................................................ ........... ..... 82
W eibull M odulus, Intercept and Threshold Load............................................ 82
Validation Results.............................................................................................. 85
Results in Manufacturing Since Implementation of NDE at RTI..................... 85
Conclusions........................................................................................................... 88

PROCESSING AND DONOR DEMOGRAPHICS OF CORTICAL BONE................ 89

Introduction........................................................................ ...................................... 89
M materials and M ethods.......................................................................................... 90
Specimen Preparation ...................................................................................... 90
Experimental Design.......................................................................................... 91
Treatment Groups ............................................................................................ 92
Physical and M echanical Testing.......................................... ........................ 94
Results................................................................................................................... 97
Discussion ........................................................................................................... 100


v









Conclusions.................................................................................... ..................... 106

DONOR DEMOGRAPHICS, PROCESSING, AND DESIGN OF THE MD SERIES OF
BONE DOW ELS ...................................................................................................... 108

Introduction......................................................................................................... 108
M materials and M ethods.............................................................................................. 108
Specim en Preparation .................................................................................... 108
Donor Attributes ............................................................................................ 109
Processing Param eters ................................................................................... 109
Dowel Design................................................................................................ 110
Experim ental Design............................................................. ...................... 110
M echanical Testing....................................................................................... 111
Norm alized Strength ..................................................................................... ... 11
Statistical Analysis............................................................................................. 111
Results and Discussion ........................................................................................ 112
Donor Dem ographics....................................................................................... 112
Processing Param eters .................................................................................. 114
Design .............................................................................................................. 115
Conclusions..................................................................................................... 116

SUM M ARY ................................................................................................................... 117

REFERENCES ................................................. ...................................................... 120

APPENDIX RAW DATA, REGRESSION, AND WEIBULL ANALYSIS ................. 132

BIOGRAPHICAL SKETCH ..................................................................................... 139
























vi















LIST OF TABLES



Table page

2-1 Ultimate strength of human cortical and cancellous bone............................................9

2-2 Elastic constants for human cortical bone .................................................................9

2-3 Axial load bearing capacity of different vertebral levels .............................................20

2-4 In vivo loading on the human spine........................................................................ 21

2-5 Load bearing capacity of various interbodies............................................................38

4-1 Best regression subsets ............................................................................................ 72

4-2 Regression coefficients for various size dowels..........................................................72

5-1 Minimum sample size for validation...................................................................... 81

5-2 Possible Outcomes for validation...........................................................................82

5-3 Weibull analysis results for various dowels ...................................................................83

5-4 V alidation actual outcome es ....................................................................................... 85

6-1 Axial compression treatment groups ...................................................................... 93

6-2 Shear treatm ent groups.............................................................................................93

6-3 Diametral compression treatment groups....................................................................93

6-4 Density of each compression group prior to treatment...........................................97

6-5 Ultimate compressive strength of each treatment group ...................................... ..97

6-6 Density of each shear group prior to treatment .........................................................98

6-7 Ultimate shear strength of each treatment group....................................................98

6-8 Density of each transverse tension group prior to treatment........................................98








6-9 Ultimate transverse tensile strength of each treatment group................................ ...99

7-1 Bone dow el treatm ent groups...................................................................................... 110

7-2 Variation of strength of 16mm MD-2 due to anatomic origin ...................................... 112

7-3 Variation of strength of 18mm MD-2 due to anatomic origin ..................................... 112

7-4 Variation of strength of 16mm MD-III as a function of donor age.............................. 13

7-5 Variation of strength of 16mm MD-II as a function of donor age ...............................13

7-6 Variation of strength of 18mm MD-II as a function of donor age ............................... 114

7-7 Variation of strength of 20mm MD-II as a function of donor age ................................114

7-8 Variation of strength of 16mm MD-II due to processing.............................................. 115

7-9 Variation in strength between 16mm MD-II and MD-III dowels ................................116

7-10 Variation in strength between 18mm MD-II and MD-IV dowels..............................116















LIST OF FIGURES



Figure pag

1-1 Sketch of typical M D-II bone dowel ........................................................ ...............4

2-1 Cloward dowel...........................................................................................................27

2-2 Crock Dowel ..............................................................................................................27

2-3 Vich threaded cancellous dowel...................................... ............... .........................28

2-4 M D-II and M D-III cortical bone dowels .......................................................................28

2-5 Smith-Robinson iliac crest graft ....................................... .............. .........................32

2-6 Bailey-Bagley iliac crest graft ................................................................................... 32

2-7 TangentTM cortical wedge...........................................................................................33

2-8 Cortical ring from long bone ..................................................................................... 35

2-9 SRTM graft......... ... ......................................................................................................35

2-10 VertigraftTM .............................................................................................................. 36

2-11 FRA-SpacerTM .... ................................................................................................... 36

2-12 ALIFTM .....................................................................................................................37

3-1 Test Jig for M oir6 Loading ........................................................................................ 45

3-2 Hertzian contact stress V-field ........................................ ............... .........................47

3-3 Hertzian contact stress U-field ........................................ ............... .........................47

3-4 In-vivo loading V-field ............................................ ............................................... 48

3-5 In-vivo loading U-field ..............................................................................................48

3-6 Typical fracture patterns observed in bone dowels under axial compression .................50


ix








4-1 Cross-section of hum an fem ur................................................................................... 53

4-2 Typical production dow els ........................................................................................54

4-3 Dowel under axial compression .......................................................................... 57

4-4 H istogram of failure loads .........................................................................................63

4-5 Average failure loads of different size MD-II dowels.............................................64

4-6 Histogram of failure loads of 16 and 20mm MD-2 dowels.............................................65

4-7 Failure load as a function of endwall thickness.................... ....................................66

4-8 Failure load of 16mm MD-II as a function of density.............................................68

4-9 Failure load of 18mm MD-II as a function of density.............................................68

4-10 Failure load of 16mm MD-II as a function of ultrasonic velocity in cortical ring........69

4-11 Failure load of 16mm MD-II as a function of hardness of cortical bone ....................69

4-12 Multiple regression analysis of 16mm MD-II bone dowels........................................73

4-13 Multiple regression analysis of 18mm MD-II bone dowels........................................73

4-14 Multiple regression analysis of 20mm MD-II bone dowels ........................................74

5-1 Probability plots of 16 and 18mm MD-2 dowels ....................................................84

6-1 Shear Test Fixture........................................................................ ............................ 95

6-2 Test configuration for axial and diametral compression ....................................... ..96

A-1 Determination of Weibull modulus for 16mm MD-2...................................................138















Abstract of Dissertation Presented to the Graduate School
Of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



DESIGN AND MECHANICAL BEHAVIOR OF THE MD SERIES OF BONE
DOWELS

By

John R. Bianchi

December 1999

Chairman: John J. Mecholsky, Jr.
Major Department: Materials Science and Engineering

Allograft bone dowels, developed at the University of Florida Tissue Bank, Inc

(UFTB) and Regeneration Technologies, Inc (RTI), offer an alternative to the more

conventional metallic and other synthetic dowels for spinal fusions. These dowels are

machined from the long bone of human donor tissue. They are an advance over current

implants because they possess the precise dimensional characteristics that are typical of

metallic or other synthetic implants, are composed of mostly cortical bone, do not cause

additional donor site morbidity associated with autografts, yet they retain the

advantageous osteogenic characteristics of allografts and autografts.

Allograft and autograft tissues have a well-established history in spinal fusions.

However, postoperative failures are commonly reported. These failures are due to the

variations in geometric, material, and mechanical properties of the implants. In addition,








little research effort has been placed on insuring that these types of implants have at least

a minimum level of load bearing capacity.

The results of this research developed a novel method, based on statistical

procedures and fracture mechanisms, that defines the strength of the MD-series of bone

dowels and uses this technique to establish a nondestructive mechanical quality control

procedure. In addition, the influence of donor characteristics such as age, and sex on the

strength of the dowels was established. The role of different tissue banking processing

steps on the strength of machined tissue was identified, as well as differences in strength

among different dowel designs determined.














CHAPTER 1
INTRODUCTION

Allograft bone dowels, developed at the University of Florida Tissue Bank, Inc

(UFTB) and Regeneration Technologies, Inc. (RTI), offer an alternative to the more

conventional metallic and other synthetic dowels for spinal fusions. The MD-series

(machined dowel) of dowels were developed as an intervertebral spacer for spinal fusion.

The dowels are machined from the long bone of human donor tissue. They are an

advance over current implants because they possess the precise dimensional

characteristics that are typical of metallic or other synthetic implants, are composed of

mostly cortical bone, do not cause additional donor site morbidity associated with

autografts, yet they retain the advantageous osteogenic characteristics of allografts and

autografts.

Bone dowels are manufactured from the long bones of humans. Generally, a

dowel may be machined from a male or female donor between the ages of 15-70 years.

Typically, the dowels are obtained from the femur, tibia, and humerus. In addition, if the

anatomy of the bone will allow for machining of the implant, dowels may be obtained

from the radius and ulna.

Dowels are cut perpendicular to the long axis of the bone within the diaphysis.

Generally, the angle of the cut is made which will yield the dowel with the best

dimensional characteristics. Typically, this is an anterior-to-posterior cut. The last

machining step is to place threads on the outside surfaces of the dowel. The threads serve





2


two primary purposes. They allow for ease of insertion and prevent the graft from

backing out or dislodging after it is implanted.

A sketch of a typical MD-II is shown in Figure 1-1. Various regions have been

defined for the dowels. The endcaps are two cylindrical disks that are at each free end.

The endcaps may be distinguished from one another by the presence or absence of a

driver slot and slotted driver hole. The function of the driver slot is to mesh with a

surgical instrument in order to rotate the dowel into the intervertebral space. The slotted

driver hole is utilized to secure the dowel to the instrument. The instrument is detached,

after the dowel is implanted. The two endcaps may be described as the unslotted endcap

and slotted endcap. The endcaps are connected to one another via the sidewall. An MD-

II dowel contains two sidewalls. The open space within the two endcaps and the

sidewall(s) is an anatomic artifact of the medullar canal. All of these regions are

identified in Figure 1-1.

Allograft and autograft tissues have a well-established history in spinal fusions.

However, postoperative failures are commonly reported. This is due to the variations in

geometric, material, and mechanical properties of the implants. The properties of neither

bone nor the products that are derived from bone are homogeneous. The intrinsic material

properties of bone vary along the long axis, the circumferential and radial directions. The

dimensional features of bone are not uniform. The cross-sectional area of a long bone

may be tubular, oval-like, or triangular. Also, two types of bone, both cortical and

cancellous, may be present in the bone stock as well as the final product. In addition

there are donor to donor variations as well as variations in the processing of the tissue.








These facts create a wide range of geometric, material, and mechanical properties

variations in the final MD series product line.

There are many disadvantages associated with these biological implants because

of the variability. Some clinical studies discuss failures that occur after implantation of

allograft or autograft interbodies for spinal fusions. Biomechanical complications

associated with the interbody include extrusion (25, 152) and collapse (25, 133, 152,

159). It is reported that between 6% and 44% of failed cervical fusions are due to

collapse of the interbody (25, 31, 80, 133, 152, 159). The potential complications to the

patient associated with this type of malfunction may range from the mild discomfort,

pain, paralysis, or even death. Collapse of the graft occurs because the applied loads

exerted via the spinal column exceed the load carrying capacity of the interbody. Little

research effort has been placed on insuring that these types of implants have at least a

minimum level of load bearing capacity.

The purpose of this research is to propose a method based on statistical methods

and fracture mechanics to define the strength of the MD-series of bone dowels and use

this technique to establish a nondestructive mechanical quality control procedure. In

addition, the influence of donor characteristics such as age, and sex on the strength of the

dowels will be explored. The role of different tissue banking processing steps on the

strength of machined tissue will be identified. Differences in strength among different

dowel designs will be determined.













Medullar Canal


------ f


Slotted driver


Figure 1-1. Sketch of typical MD-II bone dowel














CHAPTER 2
REVIEW OF THE LITERATURE

The Function and Anatomy of Bone

Bone performs many different functions, such as, support and protection, mineral

reserve, calcium and phosphate homeostatisis, blood and immunocompetent cell

formation, and storage of large number of growth factors.

Soft tissue of the periosteum and endosteum cover the outer and inner surfaces of

bone respectively. Bone may be grossly classified as either primary or secondary.

Primary bone is present during embryonic development or fracture repair. One of the

exclusive features of this type of bone is that the fibers are randomly distributed. It is

also temporary and replaced by secondary bone. Secondary bone consists of cancellous

and cortical bone. Cancellous bone is porous and weak compared to cortical bone. In

long bones the epiphyses, the bulbous end, contains cancellous bone that is surrounded

by a thin cortical layer. At the diaphysis, the midportion, it is almost exclusively cortical

bone.

Bone is composed of a three dimensional polymeric organic component (Type I

collagen) and inorganic hydroxapatite crystals (CAIo(PO4)6(OH)2). These crystals,

approximately 40x25x3nm, lie along the collagen fibrils and are surrounded by ground

substance and a hydration shell. The long bones contain rings called lamellae that

compose the outer and inner surfaces. Between the inner circumferential lamellae and

outer circumferential lamellae are the interstitial lamellae. Other microstructural features

of long bones include the Haversian canal, which is in turn formed by cellular activity of








osteoclasts and osteoblasts. These canals traverse the bone in a longitudinal direction and

are connected by numerous lacunae and even smaller canals called canaliculi.

Volksmans canals run transverse to the Haversian canals. These microscopic features

result in anisotropic mechanical properties in that the properties vary in different

directions.

Bone is in an equilibrium state of bone resorption by osteoclasts and bone

formation by osteoblasts. It is estimated that as much as 10% of the human skeleton is

remodeled annually (72). There are many local and systemic factors that control this

process such as mechanical stress, electrical signals, the condition of the orthotopic site,

the presence of agents that suppress bone remodeling such as non-steroidal anti-

inflammatory agents, or malnutrition (59).

When bone fracture occurs, a local hemorrhage and blood clot appear forming a

callus which is eventually removed during repair by macrophages. The endosteum and

periosteum produce osteoprogenitor cells and cellular tissue surrounds the fracture site.

This is followed by trabeculae of primary bone connecting the fractured ends and

forming a callus. Physiologic applied loads eventually remodel the callus to the anatomic

features prior to fracture and, secondary bone replaces the primary bone.

Transplanted bone dies but is remodeled in a similar process termed "creeping

substitution". The host reorganizes the transplant by the invasion of vessels and cells,

endosteum, marrow and other tissue. The surgeon (72, 116) relies on the resorption,

revascularization, osteoconduction, and osteoinduction as well as the fracture repair (74)

process to create a solid union in order to fulfill the goal of the surgery.








The ability of a material to enhance the bone healing process has been described

by three distinct properties (41). These properties are described as osteogenic potential,

osteoconductivity, and osteoinductivity. Osteogenic potential is described as the "ability

of the graft to make bone" (72). In order for a graft to possess this property certain cells

must be present. Most implant materials including autografts do not have osteogenic

potential. Osteoconductivity is described as the ability of a graft to support the healing

process similar to a scaffold (72). The implant provides a three dimensional structure for

invasion by the host. All autografts and allografts as well as certain synthetic porous-like

structures are osteoconductive. Cancellous bone has excellent osteoconductive

properties. Osteoinductivity is the ability of a graft to induce certain cells into becoming

bone (72). Only autografts and certain allografts have the ability to be osteoinductive. In

addition to the osteogenic properties, bone, especially cortical bone, has advantageous

mechanical properties.

Mechanics of Bone

Human bone has been broadly classified into two types: cancellous bone and

cortical bone. Cortical bone covers the outside surface of long bone however the

thickness varies along the longitudinal and circumferential axis. The thickness is largest

at the diaphysis and gradually tapers to a thin shell at the epiphyses. The humerus and

femur may be generally described as tubular while the tibia has a thick knob of cortical

bone on the anterior side that runs along the longitudinal axis. Cortical bone is

characterized as a densely packed ceramic in a collagenous matrix. Cancellous bone is

characterized by its porosity and intricate lattice network. However, the transition from

cortical bone to cancellous bone is difficult to identify by visual observation. Other

means of distinguishing between these two types of bone have been reported. The








density of cortical bone has been reported to be 1.85 g/cm3 while cancellous bone has

been found to be 0.307g/cm3 (135). Similarly, cortical bone has four times the mass of

cancellous bone. The mechanical property differences between cortical and cancellous

bone are large. Table 2-1 and 2-2 summarizes some of the mechanical properties

regarding these two types of bone. The volume and quality of cortical bone present will

govern the failure strength of products made from bone. The challenge exists in

quantifying these values.

In addition to variations in the dimensional features of the cortical bone, the

elastic properties of cortical bone vary not only between different individuals but also

between different bones of the same individual (73). Furthermore, the elastic properties

vary around its circumference. To complicate the analysis more, the elastic properties of

cortical bone have been classified as anisotropic. Some report that the properties are

orthotropic while others define the properties as transversely isotropic (44, 64).

The mechanical and material properties of bone have been investigated by a wide

variety of researchers. The research may be roughly divided between two separate

groups. The first group may be classified as focused on material property measurements

of bone. In these tests, bone is evaluated under a generally accepted test method that has

been established for polymers, metals or ceramics. However, it is interesting to note that

an ASTM (The American Society for Testing and Materials) test method to describe the

mechanical testing of bone or products derived from bone does not exist. Yet recently,

ASTM has established a committee on test methods for biologic products. The second

group may be classified into the structural property measurements of allograft implants.










Table 2-1 Ultimate strength of human cortical and cancellous bone

Direction (44, 135) Tensile Compressive Shear
Strength Strength Strength
(MPa) (MPa) (MPa)
Cortical 133 195 69
Longitudinal
Cortical Transverse 51 133
Cortical Radial 51 133
Cancellous 8 0.1-100


Table 2-2 Elastic constants for human cortical bone

ELASTIC CONSTANT Cortical Bone
(44)
El (GPa) 12
E2 (GPa) 13.4
E3 (GPa) 20
G12 (GPa) 4.53
G13 (GPa) 5.61
G23 (GPa) 6.23
v12 0.376
v13 0.222
v23 0.235
v21 0.422
v31 0.371
v32 0.350


Cancellous Bone
(73)
0.614
0.544
1.202


There is a large volume of literature that describes the test results of cortical or

cancellous bone. In fact, Cowin (44) has authored a wonderful book that describes many

of the material and mechanical properties of bone. In addition, the current literature

contains volumes of material that covers material and mechanical property measurements

of bone. A lot of this effort has focused on elastic property measurements. This includes

methods of determining elastic modulus (14, 119), elastic modulus or strength variations

in anatomic location or bone type (15, 49, 54, 86), or differences in Young's modulus








between cancellous and cortical bone (121). These techniques have been investigated for

clinical use such as determining the elastic property changes after hip arthroplasty (81),

and correlation between mechanical properties and computerized tomography values

(123). Other areas include fracture mechanics of bone (98), fractal analysis (122), and

fatigue (110). Several authors have reported the changes associated with cortical bone as

a function of age such as toughness (46), presence of microcracks (42, 104), and bone

loss (93). Martin (94, 95) reported on several factors such as density, collagen fiber

orientation, and percent mineral content that influence the ultimate failure stress of

cortical bone.

Some work has been devoted to predicting the ultimate strength of bone in

anatomically specific areas. Usually these regions have specific clinical relevance in

which the motivation has been in identifying those patients that have a substantial risk of

fracture (36, 68, 102). There are numerous failure theories available for predicting the

load bearing capacity of structures. These range from the relatively simple Maximum-

Shear-Stress Theory or von Mises-Hencky theory (132) for isotropic materials to the

complex anisotropic formula such as Tsai-Wu (140). In addition, Cowin (43) developed

a failure theory specifically for bone. However, in order to utilize these failure theories,

the principal and/or shear stresses must be determined as well as the failure stresses in

these directions. In order to determine the applied stress field in the cortical bone, the

elastic constants, and dimensions of cortical bone must be known. In addition, the mode

of loading must be properly described mathematically.

In the field of tissue banking there are a variety of conditions including material,

processing, and design that may lead to material as well as structural variation in the final








product. Material may include broad descriptors of the bone stock such as donor age,

sex, height, race, geographic origin, and lifestyle. Processing includes the typical

industry practices such as chemical exposure, freezing, freeze-drying ethylene oxide, and

irradiation. Design refers to the shape and state of the final product. Although some

research effort has been devoted to exploring these effects, many of these biomechanical

issues remain to be explored especially if allograft tissue is to compete with the synthetic

implant market.

Mechanical testing of bone after certain treatments that are specific to the tissue

banking industry have consisted mainly of compressive, tensile, or torsion testing. Many

have drawn conclusions regarding other modes of loading based on the results of this

testing. Torsional strength (4) (measured as a torque load not a torsional stress) was

found to decrease after autoclaving however this study has many limitations for example

the authors fail to report the specimen size or mass. Specimens (113) after various

reservations were tested in a similar manner in torque, compression, and bending. The

preservation treatments included different freezing cycles and freeze-drying while the

specimens were rat femurs (torque) or rat vertebral bodies (compression). The authors

concluded that freezing treatments did not affect the mechanical properties and that

freeze-drying decreased the torsional strength but not compression. The limitations of

this study were that the torque was analyzed as a torque load not torsional stress and that

once again neither masses nor dimensions of the specimens were reported. Compression

was analyzed by determining a compression stress assuming a constant cross-sectional

area of 10mm2. A similar study (78) was performed on rat femur (bending) and vertebra

(compression) after freezing or freeze-drying and loading under compression or bending.








Once again the authors conclude a reduction of strength (expressed as load not stress)

without accounting for cross-section, mass, or density variations. The sample sizes for

each group are small (n=10) and differences may exist between rat and human bone.

Some authors (61) reported the material property changes that occur as freeze-dried tissue

is rehydrated as well as the changes that occur after irradiation. The specimens were

human tissue and machined into dogbone-like specimens (rectangular parallelepipeds)

along the long axis of either the tibia or femur and tested under tension. They report that

the material properties were decreased by freeze-drying but recovered completely after 8

hours ofrehydration. Irradiation to 3.5Mrads in the frozen state was reported only to

increase the plastic modulus. However, irradiation in the freeze-dried state and

rehydration for 24 hours was found to reduce the yield stress, and ultimate strain. These

authors do not report the magnitude of any of the changes in any of the properties that

they measured. In addition, they do not report the associated density of the specimen and

only have tested the specimens in one mode of loading. More recently Hamer (66)

proposed a method for standardizing the test methodology by testing rings under a three-

point bending apparatus. However, this method assumes a constant cross-sectional area,

which, especially for biological materials, is usually not the case. This technique was

utilized to determine differences in strength (again expressed as load), and stiffness

among specimens after various treatments (67). These treatments included gamma

irradiation at different levels (0.7-6.0 Mrads), freeze-thawing, and freezing and

irradiation. Freezing was identified to not have an effect, while irradiation decreased the

strength to a maximum of 45% of controls at 6.0Mrads. In addition, a separate study (65)

was performed to assess the difference between irradiation in the frozen state versus








irradiation at room temperature. The irradiation dosage was 3.0 Mrads. The strength

measured, as maximum load at failure, was found to be significantly decreased in the

room temperature group. However, the authors do not mention how much time the

specimens were at room temperature prior to testing. Speirs (138) described their test

procedure on human malleuses and incuses that have been machined into cylinder

(diameter=1.5mm, length=4.0mm), treated differently, and tested under compression.

They reported the changes in ultimate compressive stress, yield stress, and elastic

modulus. They identified the ultimate compressive stress was significantly decreased do

to autoclaving and 1 N NaOH. Conversely, the LpH (a penolic solution) group did not

significantly affect the failure stress when compared to the controls. However they do

not account for differences in density in the specimens.

Several papers (111, 112, 128) are available which summarize the current

knowledge regarding the biomechanical differences in cortical bone due to a variety of

factors. Many of these authors relied on the work of Bright. There exists a need to

quantify how these industrial practices influence the material properties of bone under

each of the orthogonal directions. In addition, a methodology to quantify the influence of

extraneous factors such as density, and donor properties must be established.

Some authors have discussed the structural properties of allograft products. Most

of these products are used exclusively for spinal surgery. Although a lot of mechanical

problems, especially with collapse of the bone graft, have been reported for these

implants, little effort has been given in assuring at least a minimum level of residual load

bearing capacity. The results of this field of research will be discussed later in the section

on structural properties ofallograft interbodies for spinal fusion.








Historical Development of Bone as a Transplantable Material

The early pioneers in bone transplantation successfully demonstrated that bone

could be utilized for a wide range of applications in osseous repair in many different

anatomic locations. The types of bone that may be transplanted depend on the origin of

the transplanted graft. Autografts are taken from the same subject in whom the graft is

implanted. Allografts are from a different subject but the donor is the same species.

Xenografts are transplants from one species to another. Currently, autograft has been

designated "the gold standard." Its osteogenic capabilities have been shown to be

superior to all other materials. However, autografts have a variety of limitations (12, 87),

including additional surgical time, risk of infection, pain, and/or other complications at

the site. In addition, the bone stock in the patient may be inadequate for the procedure.

The retrieved graft may weaken a load bearing tissue such as grafts taken from the tibia.

For these reasons, allografts have become quite popular. It is estimated that over 250,000

allograft procedures occur every year. The major disadvantages of allografts are that they

carry the risk of disease transmission including infection and HIV (128). In addition, the

mechanical property variations may lead to mechanical complications in the host.

However, if the donor and products are properly screened and the processing carried out

under appropriate conditions, this risk becomes extremely small.

Several authors reference Macewen (92) in 1878 as performing the first

successful allograft transplant. In 1889, Senn (131) published his findings of utilizing

demineralized zenograft tissue (tibia of an ox) in order to fill osseous defects in the tibia

and femur. He stated that "antiseptic decalcified bone is the best substitute for living

bone grafts in the restoration of a loss of substance in bone." Senn stated that the

material was antiseptic by storing in a "solution of sublimate in alcohol 1:500 in a








hermetically sealed bottle." In 1912, Carrel (33) published his finding on preservation of

tissue in a condition of potential life. Tissues in this condition may be preserved outside

the body for an indefinite time. A variety of tissues were deposited into different

mediums (isotonic sodium, chloride solution, Locke's solution, Ringer's solution,

defibrinated blood, confined human air, and petrolatum) and stored between -1 C and

+7C. In addition, skin and bones procured from an infant were preserved in petrolatum

and ringer's solution and transplanted to several different patients. Carrel also expressed

the need for the availability of variety of allograft tissues that are packaged conveniently

for many different operative procedures. In 1923, Hass (70) demonstrated that

osteoblasts may survive for as long as nineteen hours after removal from the host. In

1937, Orell (106) reported the clinical success of allografts, xenografts, and autograft that

had been boiled. The grafts were utilized for filling defects, correcting deformities,

implanted in the hand, foot, and spine. He also utilized boiled bone for osteomyelitis and

tumors. In addition, the illustrations in his paper show several different types of

machined allografts and xenografts from hard tissues. However, he did not discuss his

methods for shaping these tissues. Later other researchers have utilized autoclaving as a

means of sterilizing the tissue. However, heat not only eliminates the osteogenic

capability of bone but also may diminish the strength (4).

Many of the early bone transplant procedures utilized tissues from amputated

limbs or procured the tissues from donors in the operating room with the patient.

However, these tissues were often of inferior quality for the recipient. Later research

focused on methods of preserving and sterilizing tissue in more convenient forms for the








people responsible for distributing the tissue as well as the users. In addition, many

investigators proposed methods of screening the donors for potential infectious diseases.

In 1939, O'Conner (105) reported his experience with merthiolate as a

preservative. The tissue was preserved in sealed jars containing an aqueous merthiolate

solution (1:1000) and saline, and stored in a refrigerator. He reported that he utilized this

method to preserve cartilage and transplanted the grafts in 375 procedures. In 1949,

Reynolds (120) advocated a similar preservation technique for bone. He stated that the

healing process of "merthiolated-preserved bone was essentially the same as that of the

fresh autogenous grafts." In addition, the container with the tissue may be stored at room

temperature in the operating room. He reported that this tissue had been successfully

utilized in over seventy patients without any patients experiencing "sensitivity" to the

implant. Although, these authors provide scientific evidence that these implants may be

used successfully, merthiolate contains mercury that is a well-established toxin. In

addition, others report the failure rate between 13%-40% (88).

In 1942, Inclan (76) explained his results of utilizing donor bone that had been

procured aseptically and submerged in citrated blood of the donor or patient and

refrigerated between 37-40F and bacteriologically controlled during preservation

without explanation. In 1947, Bush (30) proposed donor acceptance criteria as well as

describing methods for properly storing the tissue. The donor's history, including the

following tests should be kept: source of bone, Kline or Kahn reaction (for syphilis),

history of jaundice, malaria or infections, date of storage, and weight of bone. Bush also

stated that blood typing or Rh factor of grafts was unnecessary. The reservations

described were regular refrigeration or deep freezing. Regular refrigeration (between +20








and +5C) was recommended for three weeks with the bone stored in a glass screw top

container and inserted into a larger bottle. In addition, rubber sheeting and gauze capped

the top of the inner vessel. The bone may be preserved by deep freezing (<25C)

indefinitely in the previously described container. In 1947, Wilson (153) published his

finding on utilizing bone retrieved from amputations and other surgical procedures where

it would not be harmful to the donor. The tissue was stored between 10-200F. Wilson

also performed cultures and ruled out donors based on Kline, malaria, hepatitis, and other

recent acute infections. Wilson stated that he believed that creeping substitution

progressed easier in bone that was dried or boiled. However, he offered no scientific

evidence as proof. Wilson also professed the need for tissue banks to provide tissues to

large regions that were composed of many different hospitals. In 1949, Weaver (148)

described his experience using banked bone. He followed Bush's suggestions but did not

perform any culturing. In addition, he stated that he had shipped the tissue to other

locations. The tissue was packaged in glass vials and placed in insulated containers

packed with dry ice and transplanted successfully. In 1952, Cloward (39) described his

experiences using banked bone for spinal fusions. His technique was to procure the

tissue under aseptic conditions, place the graftss in glass jars in a solution of blood or

plasma, and stored either between 0-20F or 360-42F. He stated that the blood is "an

excellent culture medium, the chances of detecting the slightest contamination are greatly

enhanced."

Several researchers associated with the Navy (82) provided a detailed article of

their experience with lyophilized or freeze-dried bone that is stored at room temperature.

They concluded that freeze-dried bone "incorporated in the same manner as fresh








autogenous bone grafts but at a slower rate." Freeze-drying offers many advantages,

including more convenient storage and ease of handling. However, the lyophilization

process has been reported to degrade the structural characteristics of the graft.

More recently the American Association of Tissue Banks (AATB) held their first

scientific meeting in 1977 (28). The AATB (1) provides guidelines for Tissue Banks

that covers a variety of issues. These include organization, records, donor suitability,

tissue procurement, processing, labeling, and storage. This standard provides detailed

information regarding the donor screening for pathogens, however, it does not provide

any guidelines for ensuring the structural reliability of any of these products. These

standards list several methods for the sterilization, preservation, and storage of tissue.

This includes cryopreservation, freezing (< 40C), irradiation (>1.5 Mrads), chemical

(ethylene oxide, ethylene chlorohydrin, or ethylene glycol), and lyophilization (ambient

or less). The guideline specifically state "the following should not be used: mercurials,

quaternary compounds, formaldehyde, beta propiolactone, glutaraldehyde, and

chloroform."

Irradiation of tissue in order to reduce the bioburden has been discussed in the

literature (57). However, irradiated tissue has been reported to reduce the osteogenic

capability and compromise the biomechanical properties by destroying the collagen.

Large doses of irradiation greater than 3.0Mrads are required in order to reduce the HIV

DNA to below detectable levels. As discussed earlier, the magnitude of the

biomechanical effects of irradiation may depend on whether the bone is frozen, at room

temperature, freeze-dried, or irradiation dosage (57). These experiments were carried out

under bending loads on variable specimen geometry and were not normalized by cross-








sectional area or density. In addition, the anisotropic directions were not accounted.

Cancellous bone (11) appears to be affected in a similar manner.

Cloward first explored the use of ETO (ethylene oxide) sterilization for allografts

(144). However there is a lot of controversy regarding the residuals of not only ETO but

also some of the by-products such as ethylene glycol and ethylene chlorhydrin that may

remain in the tissue. Several reports have described the adverse clinical consequences

involved in utilizing this chemical sterilent (16, 143). Surprisingly, it remains on the list

of acceptable terminal sterilization procedures in the AATB Standards for Tissue

Banking (1).

Indications for Spine Surgery

Over eighty percent (51, 146) of adults experience back pain. It is one of the

primary reasons (129) for health care advice, hospitalization, and disability

compensation. The cost associated with this condition is estimated to be 50 billion

dollars annually. Fortunately, for most patients, after two weeks (51, 52), the pain

subsides. However, for the rest, additional treatment may be required. This may range

from conservative therapy such as bed rest or physical therapy to aggressive surgical

treatment. Approximately 0.5%-2% (51, 62) of back pain patients require surgical

intervention. There are approximately 100,000-450,000 lumbar spine fusions every year

in the United States (34, 146). There are myriad conditions that may be indications for

spinal surgery. These include but are not limited to herniated disks, spinal stenosis,

trauma, infection, spondylolisthesis, degenerative conditions. A long list of indications

could be constructed but Sutterlin (26, 141) has narrowed this list to five surgical goals:

correction of deformity, decompression of neural elements, debridement of pathology,

stabilization of the spinal column, and pain control.








Surgery of the Spine

Spinal surgery may range from the relatively simple discectomy to the complex

multilevel spinal reconstruction and fusion. These techniques range from the occiput to

the sacrum, and either anterior, posterior, or lateral. In addition, they may be either open

or endoscopic techniques. There are numerous references (19, 21, 58, 62, 75) the reader

may consult in order to review the details of these intricate procedures. Often, some type

of implant is usually used in conjunction with any of these procedures in order to

stabilize the spine and achieve fusion (96).

Spinal Mechanics

Several authors have studied the loading characteristics of the spinal column.

Many studies have focused on the axial load carrying capacity of different vertebral

levels on human cadaveric spines. Table 2-3 presents a summary of these findings.


Table 2-3 Axial load bearing capacity of different vertebral levels

Level Mean Stdev # Loading rate Reference
Load (kN) (kN) Specimens
Lumbar 6.1 3.4 32 0.64mm/s (23)
L5 6.7 2.2 4 Not Reported (89)
L4 5.8 2.6 5 Not Reported (89)
L3 5.5 2.0 7 Not Reported (89)
L2 5.5 2.2 6 Not Reported (89)
L1 5.0 1.9 4 Not Reported (89)
T12 4.6 1.8 6 Not Reported (89)
T11 4.7 2.5 4 Not Reported (89)
L4-males 5.4 1.8 7 0.1 mm/sec (17)
L4-females 4.1 1.9 13 0.1 mm/sec (17)


In addition, other researchers have explored the in vivo loading on the human

spine. Table 2-4 summarizes the results of these findings.









Table 2-4 In vivo loading on the human spine

Type of Maximum Load Reference
Activity (kN)
Standing 0.5 (101)
Sitting 1.0 (101)
Weight Lifting 3.0 (101)
Brick Laying 4.5 (48)
Heavy Lifting 7.2 (130)


Mechanics of Spine Fusion

In order to create the fusion, the motion segment must be mechanically stable and

appropriate osteogenic media must be present. Benzel (21) describes the events that lead

to bony fusion as a "proverbial race between the failure of the instrumentation construct

and the acquisition of bony union. During fusion the instrumentation construct and its

interface with bone become progressively weaker and the bony union becomes stronger."

Mechanically, the implant system must provide rigid stabilization. However, if

the implant is too rigid stress induced osteopenia of the vertebral bodies and spinal

degeneration of the juxtaposed disc may result. Alternatively, if the system is too supple,

pseudoarthrosis may result.

Biology of Spine Fusion

There are complex interactions (41) during the fusion process with many systemic

and local factor contributing. Some systemic factors include age, state of nutrition,

health, and presence of bone growth factors. Local factors that may have an influence

include mechanical stability, mechanical loading, and the presence of other local growth

factors, and infection. Ray (118) reported the difference in bone grafts that had been

prepared under different conditions. One graft was simply frozen homogenous bone, the

other was inorganic bone salt, and the last was demineralized bone. They concluded that








demineralized bone was the best. The authors indicate that Senn came to the same

conclusion in 1889. It was left to Urist (145) to identify one of the major factors present

in demineralized bone that provides its osteoinductive properties. Urist named it bone

morphologic protein (BMP). BMP is an inductive non-collagenous protein.

The other major focus has been devoted to the biology of fusion and many clinical

studies examined the difference between autograft with allograft (9, 155). In some of

these studies the allograft has been combined with other substances such as autograft or

demineralized bone matrix. The clinical findings in the autograft group were better.

Synthetic Implants Used for Spinal Fusion

Berthold Hadra used the first implants described in the literature in 1891 (41). He

utilized a wire around adjacent spinous processes. In 1915, Albee (6, 7) published fusion

techniques of using autogenous bone from the tibia. Currently, numerous implant

systems are available in the market that are utilized to promote arthrodesis. These may

consist ofintervertebral spacers such as allografts, autografts, zenografts or some other

synthetic material (metallic, ceramic, or polymeric). For example, the BAKTM (8) or

Moss cage are examples of metallic (Ti 6A1-4V) intervertebral spacers. The BAKTM is

manufactured and distributed by SpineTech, Inc. The design loads for these implants is

9600N under axial compression (109). Both of these devices are cylindrical and allow

either autogenous or allograft bone to be packed inside the cylinder. ProOsteonTM

(Interpore, Irvine, CA), a coral based hydroxyapatite material, has been implanted in a

goat model in the cervical spine. However, it was not recommended for human use

because of the "significant rates of collapse" (158). A carbon fiber composite is currently

being implanted that has mechanical performance superior to tricortical autografts and








allografts (53). In addition to the synthetic interbody, screw and/or plate or rod systems

may be used to supplement the intervertebral spacers.

Bone Grafts Utilized in Spine Surgery

Albee (5) advocated some of the first techniques that were developed for spinal

fusions that utilized bone grafts. He described a tibial bone graft along the spinous

processes that were secured with suture. This type of procedure successfully arrests the

spondylolisthiesis and fuses the spine; however, the graft will not resist physiologic

loading. Other procedures propose that the graft be placed between the bodies or as an

interbody graft. This type of procedure subjects the graft to large compressive loads.

The function (25, 31, 156, 157) of the interbody fusion is to assist in maintaining disc

height, protecting nerve roots, restoring weight bearing to anterior structures, restoring

the annulus fibrosis to tension, and immobilizing the unstable degenerated intervertebral

disc.

Interbody Autografts or Allografts for Spine Surgery

Capener (32) proposed a method of placing a bone graft through lumbosacral

bodies but concluded "the technical difficulties... precluded their trial." In 1938 Speed

(137) successfully demonstrated that this type of procedure could be performed. In 1936,

Mercer (99) demonstrated one of the first interbody grafting techniques. These load-

bearing implants are placed between the vertebral bodies in order to achieve intercorporal

fusion. In 1944 Briggs and Milligan (27) describe a fusion using autogenous chips made

from excised spinous processes. They report that they were unsuccessful in obtaining an

intercorpus fusion when the chips were packed between the bodies. However, they also

report that in later cases they "fashion a round peg" which is driven between the two

bodies. However, they do not mention the bone origin or size of this round peg. Ovens








and Williams (108) report that "pain...could be prevented by building up the joint space

between the adjacent bodies of the vertebra and by fusing this space to prevent motion at

this joint." Their technique was to place a 4x3x10mm size piece of bone from the

spinous process between the vertebral bodies. They report clinical success in two

patients. Jaslow (77) reported a similar technique in 1946. In this time period both

Smith- Robinson and Cloward popularized two independent techniques for anterior

cervical spinal fusions. Cloward advocated the use of unicortical cancellous dowels

obtained from either a bone bank or autogenous iliac crest. Smith and Robinson

proposed a rectangular keystone graft retrieved from the patient's own iliac crest.

Dowels

In 1953, Wiltberger (154) suggested a simplified method of vertebral body fusion

using a prefit dowel. Wiltberger stated that he first used this procedure for

"subastragaloid gallie fusion." First the size of the dowel (either a V2 inch or 5/8 inch) is

determined from preoperative x-rays. The dowel is obtained from either "bone bank

bone or autogenously through the posterior superior iliac spine." The dowel usually has a

thin cortical wall and the rest is cancellous bone. Three or four dowels that are one inch

long may be obtained. A mating hole is made posteriorly in the lumbar spines to be fused

and either one or two dowels are placed. Wilberger stated that the double dowel

technique provides a "faster and surer fusion." Wilberger did not report any

complications associated with this technique. Sacks (127) and Harmon (69) reported a

similar technique but they utilized "coin" dowels that were stacked and placed anteriorly.

In 1958, Cloward (37) popularized the use of a single dowels for anterior cervical

spinal fusion by advocating the use of bone bank dowels that were V2 inch in diameter

that were inserted into 14mm (38) diameter holes located on the anterior vertebral bodies.








All dowels were obtained from the ilium in an anterior-superior direction. A sketch of

this dowel is shown in Figure 2-1. However, he did report that one patient was operated

on through a posterior approach and another patient received an autogenous dowel and

several patients received multiple level fusions. Cloward reported that 42 of the 47

patients operated on by this technique were completely relieved of pain while the other 5

improved. However, the bone graft resorbed in three patients. Cloward reported that

50% of the patients developed an anterior angulation. He attributed this to "compression

force applied to the anterior half of the vertebral bodies disrupts the cancellous bone

resulting in a weakening of its ability to support weight in the vertical plane."

The demand for allografts has usually outstripped the supply. Alternatives such

as bovine (20) bone have been investigated to determine their suitability. Several authors

(100, 142), discussed their clinical results of utilizing a Cloward dowel machined from

Kiel (Calf) bone. These authors reported satisfactory clinical outcomes with no unusual

complications associated with the zenograft.

In 1982, Crock (45, 63), described his technique of utilizing autogenous tricortical

dowels retrieved from the iliac crest for anterior lumbar interbody fusion. These were

unthreaded dowels and between 25 and 26mm in length. This dowel is shown in Figure

2-2.

In 1985, Otero Vich (107) published his findings on a threaded cylindrical dowel

that was between 12-16mm in diameter. These dowels were machined from both

autogenous and allograft bone. Threads are placed on a standard Cloward type dowels by

inserting the core through a die. An example is shown in Figure 2-3. The site is

untapped except for sizes larger than 14mm. The graft is screwed into place. The








advantages of this technique are that the graft may be removed and reinserted more

precisely and with ease. It may be inserted in a more controlled manner without the risk

of fracturing by impact. The loads required for extrusion of the graft are larger. A series

of 37 patients treated with this technique did not have intraoperative or postoperative

complications. Vich reported that the smaller size dowels appeared to subside more than

the larger size and attributed this to the load bearing capacity of the smaller sizes.

Currently, Cloward Instruments, Inc. (3) supplies a complete surgical set in order

to machine and insert unthreaded interbody bone graft dowels. In addition, this particular

graft is a primary product line for many tissue banks and is supplied to surgeons around

the country.

Complications associated with this type of graft are extrusion or expulsion,

collapse and subsidence. This is due to the grafts' reliance on a large volume of

cancellous bone in order to resist the applied loads. This has led to mechanical problems

with the graft and resulted in complications as high as 47.5% (80). In fact, the authors of

this paper advocated the use of the Smith-Robinson type interbody graft over the

Cloward type dowel. Others offer that this type of procedure is not rigid enough, the

surface area is much less for a dowel than a rectangular Smith-Robinson type graft, or the

angle of flexion required to extrude the dowel is less than a Smith-Robinson graft. In

addition, the clinical results were better for a keystone graft than a dowel graft (134).

In order to address many of the limitations of the cancellous type dowels, the

MD-series of bone dowels were developed at the University of Florida Tissue Bank, Inc.,

in 1996. Sketches of the MD-II and MD-III bone dowels are shown in Figure 2-4.



























Figure 2-1. Cloward dowel


Figure 2-2. Crock Dowel































Figure 2-3. Vich threaded cancellous dowel


Figure 2-4. MD-II and MD-III cortical bone dowels








Cubes and Rectangles

In 1936, Mercer (99) demonstrated one of the first interbody grafting techniques.

These load bearing implants are placed between the vertebral bodies in order to achieve

intercorporal fusion. Mercer's technique was to utilize autogenous iliac crest grafts that

were placed anteriorly between the lumbosacral joint and secured with a screw. In

addition, the surgery is performed on the patient who was contained in a plaster shell.

This brace remains on the patient for 4 months.

In 1952, Cloward (40) published a similar technique for posterior lumbar

interbody fusion (PLIF). He utilized autogenous bone from the iliac crest but preferred

allograft bone. The grafts were 1.5 x 3 cm and the full thickness of the ileum. The

cortical crests of the graft are located posteriorly and up to three grafts are positioned into

the disc space. These were placed with bone chips onto the cleaned surfaces of the

endplate. The endplates are also notched to help stabilize the grafts. Cloward also

pointed out that he did not notice if the cortical surfaces of the graft subsided into the soft

cancellous bone of the vertebral bodies. However, he did note that if a graft is used that

is composed only of cancellous bone without any cortical bone, it crushes under

physiologic load.

Smith and Robinson (125, 136), proposed shaping a 20x20x20mm piece of

autogenous bone retrieved from the iliac crest to fit the anterior cervical spine. This piece

of bone is horseshoe shaped and surrounded on three sides by cortical bone. The ends

and one side are composed of cancellous bone. The graft is inserted by a tamper such

that the cortical exterior is in a vertical position and on the anterior part of the bodies.

The superior and inferior endplates have been previously countersunk to receive the graft.

Care must be taken not to remove too much "hard bone" from the endplates or the graft








may sink into the soft cancellous bone of the bodies (126). A sketch of this graft is

shown in Figure 2-5.

In 1981, a variation (24) of this procedure was proposed. It was recommended to

reverse the placement of the graft and locate the cortical crest on the posterior side of the

vertebral body and closest to the spinal canal. These authors advocated that if the graft

must be trimmed to size the cancellous end is removed rather then the cortical end. This

reduces the chances that the graft will extrude or collapse. Clinical results following this

procedure were better than the original results (152).

In 1960, Bailey and Badgley (18) reported their findings on their technique of

fusion of the cervical spine by an anterior cervical approach with an autogenous bone

graft. Their technique consisted of creating a trough in the anterior bodies that is

approximately V2 inch wide and 3/16 of an inch deep and the height of the vertebra. A

small ledge is left on the inferior and superior bodies to hold the graft. The disc spaces

are cleaned and all cartilage removed from the endplates. The bone graft is obtained from

the iliac crest and spans more than one motion segment. The graft is held in place by the

suturing the prevertebral fascia and by compression under physiologic loading. A sketch

of this graft is shown in Figure 2-6.

The iliac crest has become the choice graft for spinal fusions. It contains the

osteoconductive and osteoinductive properties of cancellous bone as well as the

advantageous structural properties of the cortical wall. Currently, this graft is

supplemented by internal fixation and has shown excellent results with successful fusions

obtained in up to 98% of patients (79).








Complications of this type of grafts are similar to the Cloward dowels. There

have been reports that the graft may extrude, collapse, and/or resorb. The findings in

literature (50, 90, 133, 159) report between 6%-34% incidence of collapse and/or

resorption but no significant differences between dowels or tricortical blocks.

In order to address some of these limitations, the Tangent cortical wedge was

developed at Regeneration Technologies, Inc. in 1998. This implant is composed entirely

of cortical bone with ridges on both the superior and inferior surface to prevent migration

after implantation. A sketch of this graft is shown in Figure 2-7.




























Figure 2-5. Smith-Robinson iliac crest graft


Figure 2-6. Bailey-Bagley iliac crest graft

































Figure 2-7. TangentTM cortical wedge








Rings and Other Shapes

Stauffer (139) in 1972 reported that he utilized autogenous iliac crest graft for

anterior lumbar fusion and these crumbled. In later cases he reinforced this iliac crest

with a fibula or tibia graft. However, he did not describe the details of this assembly.

In order to confront the structural issues associated with predominantly cancellous

grafts such as Cloward type dowels or Smith-Robinson keystone graphs, in 1990 Salib

and Brown conceived of utilizing allograft femoral rings packed with autogenous

corticocancellous chips. These grafts were maintained in place by metallic interference

screws. A sketch of a typical ring is shown in Figure 2-8. Kumar et al reported on a

series of 32 patients with a follow-up time between 2-4 years. Although they document

that in 41% of the patients the implant was observed to subside, none of the implants

were noted to fail or crack (84).

In 1997, the Cornerstone-SR graft was developed at Regeneration Technologies,

Inc. This graft is ring-like and composed of cortical bone. It contains ridges on both the

superior and inferior surfaces to prevent migration after implantation. It is has a small

cross-section in order to be utilized in the cervical spine. The graft is shown in Figure 2-

9.

Other precisely machined allografts that are currently being machined by other

tissue banks include the Vertigraft (Figure 2-10) machined by Lifenet and the FRA

spacer that is machined by MTF (Figure 2-11). In addition, Regeneration Technologies

also manufactures a large ring for the lumbar spine named the ALIF (Figure 2-12).

































Figure 2-8. Cortical ring from long bone


Figure 2-9. SRTM graft




























Figure 2-10. VertigraftTM


Figure 2-11. FRA-SpacerTM






































Figure 2-12. ALIFTM








Structural Properties of Allograft Interbodies for Spinal Fusion

Some of the early mechanical research focused on comparing the load bearing

capacity of these three different types of grafts (Cloward, Smith-Robinson, and Bailey

and Bagley). They report the average loads to be 3373N (Smith-Robinson), 1912N

(Bailey and Bagley), and 1507N (Cloward Dowel). Tissues retrieved from younger

donors are implied to be stronger than older donors. However, the sample size was

inadequate to draw any definitive conclusions regarding donor age (150, 151). Later

research identified the absolute load-bearing capacity under compression as shown in

Table 2-5.


Table 2-5 Load bearing capacity of various interbodies

Type of Spacer Material Maximum Reference
Load (kN)
Cancellous Bone Human Allograft 0.8 (25)
Tricortical Bone Iliac Human Allograft 2.4 (25)
Crest
Tricortical Bone Iliac Human Allograft 5.7 (156)
Crest
Femoral Ring Human Allograft 45.0 (117)


More recent results in the general literature have focused on predicting the

biomechanical properties of this type of graft. Wolfinbarger and Zhang (156, 157, 160,

161) published a series of articles in which they attempted to identify correlation

between certain physical parameters of the allograft specimens with the failure loads.

These included physical, donor, or processing parameters such as freeze-drying and

gamma irradiation. However, the correlation coefficients were low (r2 < 0.35). In

addition, they attempted to provide some guidelines to the clinician as to which graft

would be most appropriate for the clinician to use. It is interesting that these authors,








who work in the tissue banking industry, recommended that the customer calculate which

graft to choose rather then setting their own "in-house" threshold. Others have

investigated if strength may be a function of radiographic density or anatomical location

on the ileum (10, 85). However, they report that "no useful correlation was observed" (r2

not reported) (25). Some research has focused on identifying the location within the

ilium that yields the highest strength. However, none of these results is particularly useful

nor have they made an impact in the tissue banking industry.

A mechanical problem is clearly evident with many of these allograft products for

spinal applications. In vivo compressive loads as high as 7.2 kN may be applied to the

interbody while many conventional allograft materials fail below this level. It is not

surprising that many clinical studies have identified the incidence of collapse as high as

44%.

Alternative Materials to Bone

There are a wide variety of implant materials. Each material exhibits unique

biologic behavior that depends on a wide range of factors. These include surface

roughness, porosity, surface area, surface charge, hydrophilicity, chemical composition,

extent of chemical leeching, implant size, presence of sharp covers, modulus mismatch,

and wear particles (71). Typically, a fibrous capsule surrounds the implant. For metals

the thickness of this capsule depends on the relative reactivity of the particular material.

For example titanium and titanium alloys form thin capsules. Alternatively, some Co-Cr

and stainless steels created thicker capsules. However, this is not necessarily true for

polymers. Polymeric encapsulation has been attributed to the toxic chemicals that leech

into the tissue from the polymers. In addition, foreign-body reactions have been reported

to occur in some of the plastic systems (149). Ceramics such as A1203 are biologically








inert and exhibit extremely small fibrous capsules. Certain calcium phospate ceramics

have been considered osteoconductive. Beta tricalcium phosphate (TCP) and

hydroxapatite (HA) are often used. In addition, other ceramics such as BioglassTM (2)

exhibit controlled reactions between body fluids such as fibrin, collagen fibers, or growth

proteins and the materials surface to regenerate tissue. However, most ceramics are

brittle and exhibit poor ductility and their use as structural implants, especially for the

spine, have been limited.

Alternatives to Spinal Fusion

Disk replacement devices have been explored in a variety of different forms.

Obvious mechanical advantages of that such a device would restore the spine's original

mobility. Allograft motion segments have been explored in a dog model (96). Disc

prosthesis have been developed such as a ball bearing, elastomeric cushions, and

hydrogel cylinders. Although none of these devices have any current long-term clinical

or commercial success, it is clear that a lot of resources are being devoted to the

development of such a device (22).

Summary

Allografts have a rich history of being utilized for spinal fusions. However,

postoperative mechanical complications have been widely reported. This is due to the

applied loads exerted on the implant via the spine exceed the load bearing capacity of the

graft. The tissue banking industry has focused little effort to ensure at least a minimum

level of load bearing capacity of these structural implants.

The purpose of this research is to present a method based on statistical methods

and fracture mechanisms to define the strength of the MD-series of bone dowels and use

this technique to establish a nondestructive mechanical quality control procedure. In





41


addition, the influence of donor characteristics such as age, and sex on the strength of the

dowels will be explored. The role of different tissue banking processing steps on the

strength of machined tissue will be identified. Differences in strength among different

dowel designs will be determined.














CHAPTER 3
LOCAL STRESS FIELD

Introduction

Allograft tissue, especially bone, is often implanted as structural material in many

surgical procedures. These tissues may be machined into specific shapes in order to

optimize their structural characteristics and add convenience to the end user. One such

allograft that is commonly utilized for spinal fusions is the MD-series of bone dowels

(Regeneration Technologies, Alachua, FL). These implants are primarily under two

different modes of loading. Torsional loads are applied during insertion, while in-vivo

they are primarily under axial compression. The purpose of this investigation was to

examine the local stress field under simulated in-vivo loading and compare this to loading

with a Hertzian contact load in order to identify similarities and differences between

these different loading regimes.

Materials and Methods

An unthreaded 14mm MD-3 dowel without a driver slot and hole was

manufactured from human cadaveric tissue. A sketch of a typical dowel is shown in

Figure 1-1. An MD-III dowel contains only 1 sidewall as shown in Figure 2-4. Various

regions have been defined for the dowels. The endcaps are two cylindrical disks that are

at each free end. The endcaps may be distinguished from one another by the presence or

absence of a driver slot and slotted driver hole. The function of the driver slot is to mesh

with a surgical instrument in order to position the dowel into the intervertebral space.

The slotted driver hole is utilized to secure the dowel to the instrument. The instrument








is detached after the dowel is implanted. The two endcaps may be described as the

unslotted endcap and slotted endcap. The endcaps are connected to one another via the

sidewall. A MD-II dowel contains two sidewalls however the MD-III dowel contains a

single sidewall. The open space within the two endcaps and the sidewall is an anatomic

artifact of the medullary canal.

One technique for observing the full stress field is Moir6 interferometry. This

well established technique (47, 115) allows direct observations of the stress gradients

while under load by visually observing diffracted light. A diffraction grating is applied to

the surface to be observed and the reflected light analyzed. The changes that occur in the

reflection pattern are directly correlated to surface displacements. Typically, two

different orthogonal in-plane directions are observed. The displacements in the

horizontal or x-direction are termed the U-field while displacements in the y-direction are

termed the V-field. Usually, some knowledge of the local stress field must be

hypothesized, i.e. tensile or compressive, in order to identify whether or not the specimen

is under an expansive or contractile displacement (47, 115).

A diffraction grating was applied to the free edge of one side of the dowel as

described previously (103). A 450 Newton (100 lbf) load was applied to the specimen

with an aluminum test jig as shown in Figure 3-1. The test jig consisted of two different

loading platens. The radius of the first loading platen differed from the radius of the

dowel by less than 0. 1mm. This loading was typical of the surgical procedure and the in-

vivo loading conditions. The radius of the second platen differed from the radius of the

dowel by more than 0.5mm. The second condition is typical of the well-established

Hertzian contact stress between concentric cylinders (124, 132).





44

The Moir6 fringe pattern was observed under the compressive loading scheme as

described in Figure 3-1. This fringe was observed under no load and load conditions.

These patterns were observed both parallel (V-field) and perpendicular (U-Field) to the

loading directions for both the in-vivo loading and Hertzian contact stress conditions.










Load cell


Force balance
device


Bone


fixture


Figure 3-1 Test Jig for Moir6 Loading








Results

Hertzian Contact Between Concentric Cylinders

The fringe patterns that were observed under Hertzian conditions are shown in

Figures 3-2 and 3-3. The displacement fields that were observed under Hertzian contact

stress show patterns typical of this type of loading (47). The V-Field (Figure 3-2)

indicates a compressive loading pattern in which the contact surface is at the 12 o'clock

and 6 o'clock positions. Both the 3 and 9 o'clock positions are under the smallest

localized stress.

The U fields (Figure 3-3) indicates a large tensile stress below the contact

surfaces (12 and 6 o'clock positions) and through the center of the specimen. Both of

these loading patterns appear asymmetric and is attributed to the fact that this is an MD-

III dowel and contains only one sidewall. A similar pattern observed on an MD-II with

sidewalls of the same dimension would be expected to be symmetric.

In- Vivo Loading Conditions

The fringe patterns that were observed are shown in Figures 3-4 and 3-5. These

patterns are different from the Hertzian patterns. The V-field, Figure 3-4, indicates that

the contact surface has moved to the 2 and 10 o'clock positions on the upper loading

platen and the 5 and 7 o'clock positions on the lower loading platen. The low strained

regions appear to be the 12 and 6 o'clock positions as well as the 3 and 9 o'clock

positions. The U-field (Figure 3-5) is similar to the Hertzian loading in that a larger

tensile region in observed through the center of the specimen. However, this tensile

region has shifted laterally. Once again these patterns appear to be asymmetric and are

most likely due to the asymmetric specimen dimensions or asymmetric loading.






























Figure 3-2. Hertzian contact stress V-field


Figure 3-3. Hertzian contact stress U-field






























Figure 3-4. In-vivo loading V-field


Figure 3-5. In-vivo loading U-field










Discussion

Other than our submitted publication, a review of the literature did not reveal any

reports of applying Moir6 Interferometry to bone. In addition, no other reports of

applying Moir6 Interferometry to Interbody implants derived from synthetic materials has

been reported. However, several authors have reported finite element analysis of a

cylindrical interbody under a similar loading scheme. The localized high stress fields for

the in-vivo loading conditions were identified to be in the same location for the BAK

fusion cages (109).

Ultimate failure testing of MD dowels under similar loading conditions revealed

two failure modes shown in Figure 3-6. The major failure modes are visible cracks

across both the top and bottom surfaces of the dowel parallel to the long axis of the dowel

at this same location. In addition, another less prevalent failure mechanism was a crack

parallel to the loading direction that traversed vertically through the center of the dowel

on the free edge of the dowel. The first mode of failure is due to the localized stresses

exhibited in the U-Field and the second mode of failure is due to the localized tensile

stresses exhibited in the V-field.

There are several limitations to the current study. Simulated in-vivo loading was

produced by utilizing aluminum blocks. This is very different from the vertebral bodies.

The vertebral bodies are not homogeneous but are a functionally gradient material. The

endplates are hard cortical bone while any distal movement away from the endplates the

bone becomes cancellous. Therefore, depending on the placement of the interbody in

either the superior-inferior or lateral-medial direction, the localized loading may be






50

































,,*i 1f -, a 4- -.. .'

..4, .''
[ ^- !.i < _" -' "-.
-
.9 p r ~ -'


Figure 3-6 Typical fracture patterns observed in bone dowels under axial compression








different. In addition, the loads applied to the spine are complex. Multiaxial loads that

involve flexion, extension, tension, compression, and/or lateral bending may be present.

However, the major mode of loading is primarily axial compression.

Conclusions

This test successfully illustrated the application of Moir6 interferometry on

compliant cortical bone in order to identify regions of high stress. In-vivo loading

identified similarities and differences between the in-vivo loading and the Hertzian

loading conditions. A large tensile region was observed in both of these loading

conditions. However, the compressive regions were different. The Hertzian conditions

exhibited two contact surfaces while the in-vivo conditions exhibited four contact

surfaces in which the location differed from the Hertzian conditions. These high stressed

regions correlated well with the typical modes of failure observed for dowels tested to

failure under axial compression.














CHAPTER 4
NORMALIZED STRENGTH

Structural Characteristics of Bone

Human bone has been broadly classified into two types: cancellous bone and

cortical bone. Cortical bone covers the outside surface of long bone however the

thickness varies along the longitudinal and circumferential axis. The thickness is largest

at the diaphysis and gradually tapers to a thin shell at the epiphyses. A cross-sectional

cut of a human femur is shown if Figure 4-1. The humerus and femur may be generally

described as tubular while the tibia has a thick knob of cortical bone on the anterior side

that runs along the longitudinal axis. Cortical bone is characterized as a densely packed

ceramic in a collagenous matrix while notable features of cancellous bone are its porosity

and intricate lattice network. However, the transition from cortical bone to cancellous

bone may be difficult to identify by visual observation. Other means of distinguishing

between these two types of bone have been reported. The density of cortical bone has

been reported to be 1.85 g/cm3 while cancellous bone has been found to be 0.307g/ cm3.

Similarly, cortical bone has four times the mass of cancellous bone. The mechanical

property differences between cortical and cancellous bone are large and have been

identified in Chapter 1. The volume and quality of cortical bone present will govern the

failure strength of products made from bone. The challenge exists in quantifying these

values.

In addition to variations in the dimensional features of the cortical bone, the

elastic properties of cortical bone vary not only between different individuals but also








between different bones of the same individual. Furthermore, the elastic properties vary

around its circumference. To complicate the analysis even more, the elastic properties of

cortical bone have been classified as anisotropic. Some report that the properties are

orthotropic while others define the properties as transversely isotropic (13, 44, 64).


Figure 4-1 Cross-section of human femur








Structural Characteristics of Bone Dowels

Figure 1-1 shows uniform and symmetric dimensional attributes of the dowels.

Typically, this attribute is not present in most of the dowels that are produced. A

photograph of typical production dowels is shown in Figure 4-2. Notice the variation in

curvature inside and outside of the medullar canal. In addition, notice the absence and

presence of cancellous bone between these two dowels. These attributes lead to

difficulties in accurately determining dimensional features of the cortical bone. In

summary, every dowel is unique. Any technique that evaluates the strength must account

for the unique dimensional variation of the cortical bone within each dowel produced.


Figure 4-2 Typical production dowels








Theoretical Stress Analysis

There are numerous failure theories available for predicting the load bearing

capacity of structures. These range from the relatively simple Maximum-Shear-Stress

Theory or von Mises-Hencky strain energy theory for isotropic materials to the complex

anisotropic formula such as Tsai-Wu (140). In addition, Cowin developed a failure

theory specifically for bone (43). However, in order to utilize these failure theories, the

principal and/or shear stresses must be determined as well as the failure stresses in these

directions. In order to determine the applied stress field in the cortical bone, the elastic

constants, and dimensions of cortical bone must be known. In addition, the mode of

loading must be understood. Partly because of the difficulties involved in determining

the quantities of all of these values and partly due to the initial success of empirical

methods, the theoretical stress analysis has not been vigorously pursued.

Modes of Loading

The dowels are placed under a variety of service loads during its lifetime. During

implantation, the dowels are primarily under torsional loading. However, in-vivo, the

dowels are primarily under axial compressive loading. It should be recognized that if the

spine undergoes flexion/extension and/or torsional loads, the dowels are under multiaxial

loading. This dissertation seeks to define the strength of the dowels under axial

compressive loading. All other modes of loading, if needed, will be dealt with

independently in a separate document.

Specific regions of the dowel are responsible for the load bearing capacity during

the different modes of loading. The sidewalls are responsible for bearing the torsional

loads, while the endcaps are responsible for bearing the majority of the in-vivo axial

compressive loads. The objective of this portion of the study was to identify the








important structural variables that correlate to the axial load bearing capacity of the bone

dowels.

Materials and Methods

Specimen Preparation

The compression strength of over 1000 dowels were mechanically tested from

over 900 different donors and the data compiled. All tissue was manufactured according

to proprietary Regeneration Technologies Inc (RTI) processing techniques. This includes

ultrasonic cleaning in chemical baths including hydrogen peroxide, and isopropyl alcohol

for less than 30 minutes, packaging and frozen storage at (-200C or -800C). The dowels

were retrieved from the diaphysis of the femur, tibia, radius, or ulna. All specimens were

thawed prior to mechanical testing.

Inclusion/exclusion criteria were applied to the set of dowels in order to analyze a

consistent set of specimens. These criteria included: only dowels with full and complete

threads along the length of both the top and bottom surfaces; dowels with driver slot

perpendicular to medullar canal and containing center hole; diameter mismatch less than

0.4mm (/D1-D2/ < 0.4mm see Figure 1); diameter of dowel within specifications (D

+0.0-0.3mm where D=16,18, 20mm).

Mechanical Testing

The dowels were mechanically tested to failure utilizing an MTS Bionix 858

servohydraulic mechanical test system. The maximum compressive load of this machine

was 25000N. The dowels were placed in stainless steel platens that have machined

grooves to match the thread profile and diameter on the dowels. The dowel was located

in the middle of the fixture with the medullar canal parallel to the direction of loading.

The loading ram was positioned in a machined recess in the top of the upper platen and









an initial axial compressive preload of 100N applied. Load was applied at a rate of

25mm/min and failure occurred when a change in load 1000N was detected. The

orientation and loading simulated in-vivo conditions. Figure 4-3 illustrates the loading

apparatus.


Figure 4-3. Dowel under axial compression








Physical Measurements

A variety of data was captured from the donor, donor tissue, or the dowel. Donor

data included age and sex. This data was determined from the donor charts. Donor tissue

data included ultrasonic velocity and anatomic location. The ultrasonic velocity may be

determined by the methodologies described by Ashman (13-15). This was determined by

calculating the speed of an ultrasonic signal by dividing the thickness of the specimen (to

the nearest tenth of a millimeter) by the time it took the signal to traverse the specimen.

Ultrasonic velocity was identified by averaging four different values 90 apart on a ring

of a long bone from which the dowel was obtained. The individual responsible for

manufacturing the dowel identified the anatomic location. Dowel data included size

(defined by diameter), design (MD-II, MD-III, or MD-IV) endcap thickness, mass,

durometry, density, length, diameter, and diameter mismatch. All dimensional

measurements were determined with digital calipers. Mass was determined on a digital

scale (Ohaus model LS200). Density was determined by dividing the mass by the

volume (via Archimedes principal).

Donor demographics, anatomic location and the different designs will be

discussed in separate chapters.

Data Analysis

The data was analyzed by a variety of different methods that are sequentially

more complex. The first method analyzed the failure loads of the dowel regardless of

size or design. The MD-2 and MD-3 series of dowels were separated by size from the

master database. Regression analysis was utilized to identify any variable that

potentially could correlate to the load bearing capacity. Lastly, a multiple regression









analysis was performed on the remaining variables to maximize the correlation

coefficient (r2).

Normalized Strength

One of the most elementary approaches to predicting failure of an object under

tensile or compressive loading is to utilize the failure criterion as shown in equation (4-

1).


A Equation 4-1
=--
CS


Where a is an empirically determined failure stress (usually yield or ultimate) that

is unique to a material, A is the actual failure load, CS is the cross-sectional area of the

material under load. The cross-sectional area depends on the geometry and is calculated

from dimensional properties of the specimen. The failure is predicted to occur when



A _2> o Equation 4-2
CS


Rearranging this equation to



A = CS Equation 4-3



This equation predicts the actual failure load (A) to occur when this value is equal

to the predicted failure load (o*CS). In fact, this equation may be rearranged to the

following:

A 1.0 Equation 4-4
r* CS











Ideally, the ratio described in equation 4-4 will be 1.0. However, deviations from

ideality occur for a variety of reasons. These may include loading variations, specimen

variation, processing variation, environmental variation, and mode of failure variation.

Some data may fall above the line while others may fall below it. If data falls above the

line, it fails above prediction and is interpreted to be strong. Conversely, if data falls

below the line, it fails below prediction and is interpreted to be weak. Thus, the ratio of

the actual Stress, A/CS, to the predicted theoretical stress, C, should be 1.0. If we assume

that the cross-sectional area of loading is the same, then this may also be expressed

quantitatively as the ratio of actual stress to predicted stress. If this ratio is larger than

1.0, the specimen is strong. If it is less than 1.0, the specimen is weak. If it is similar to

1.0, the specimen has normal strength.

Results

Failure Load of MD-II Bone Dowels

The simplest approach was to ignore all attributes of the dowels and to examine a

histogram of the failure loads of MD-II dowels. This is illustrated in Figure 4-4. Note in

this Figure how the failure loads can vary between from below 10000 N to a high of

25000N. An artificial upper limit is created by the 25000 N capacity of the mechanical

test equipment. Clearly, the fundamental question to answer is "why do certain dowels

fail at low loads (< 10000 N) while others exceed the capacity of the mechanical testing

equipment (>25000N). In order to answer such a question the dowels were separated by

size and only MD-2s were considered.








Failure Loads of MD-II Dowels as a Function of Diameter

Dowels were separated by type and size in order to isolate the propensity for

certain dowels to fail at high or low loads. The size was identified by the outer thread

diameter. Examining the data from only MD-II dowels of Figure 4-4 but separating the

data based on the size of the dowel reveals the histograms shown in Figure 4-6. The

histogram for the 16mm dowels falls within the limitations of the testing apparatus.

However, the 18mm dowels are artificially cut off at 25000N. Clearly, the size of the

dowel influences the load bearing capacity. An alternate method of examining this

phenomenon is to examine the average failure load as a function of the size and design of

the dowel. This is shown in Figure 4-5. Statistical analysis utilizing single factor

ANOVA and Fisher's multiple comparison test revealed these values to be significantly

different (p<0.05). The question asked in the previous section may be revised to the

following: "What determines the load bearing capacity of a 16mm (or 18mm or 20mm)

dowel (MD-II or MD-III)?"

Failure Load as a Function of Dimensions

One of the first empirical approaches to predict the load bearing capacity was

based on the total thickness of the endcaps. The hypothesis is, if the total thickness of the

endcaps is large, the load carrying bearing capacity will be large. In theory, this

technique uses a simple approach with a simple measurement technique (e.g. calipers).

The correlation between these two values is shown in Figure 4-7. The correlation was

poorer than expected and decreased as the size of the dowel increased. The reason for

this discrepancy is that difficulties arise in quantifying not only the volume but also the

quality of cortical bone in the endcaps with calipers. In practice as shown in Figure 4-2,

the dimensions of the endcap vary. Typically, it is a minimum along the centerline of the





62


cylinder and may increase (or decrease depending on the anatomy) to a maximum at the

sidewall. In addition, the presence of any cancellous bone will exaggerate the readings

on the calipers. Lastly, the dimensional measurements from the calipers reveal nothing

about the quality of the cortical bone. Information about the cortical bone at the endcaps,

such as the presence of a corticocancellous interface or porosity or the cortical bone

density is not considered when measuring the dimensions of the endcaps.


























Histogram of Failure Loads of MD-series of Bone Dowels
14mm MD-1, 16mm MD-II & MD-III, 18mm MD-II, 20mm MD-II

50
45
40
* 35
c 30
S25
0 20
S15
10
5
0
0 5000 10000 15000 20000 25000 30000
Failure Load Newtonn)


Figure 4-4. Histogram of failure loads


























Mean Failure Loads of MD Dowels
Error Bars indicate standard deviation

25000 -


20000
0

c 15000

0
-J
* 10000

L

5000


0
16rnn MD-2 (n=139) 18mm MD-2 (n=310) 20mn MD-2 (n=122)

Type of Dowel


Figure 4-5. Average failure loads of different size MD-II dowels























30

25

,020
o 16mm
615 .. -
o 20mm
010

5


0 5000 10000 15000 20000 25000 30000


Failure Load Newtonn)


Figure 4-6. Histogram of failure loads of 16 and 20mm MD-2 dowels
































Failure Prediction as a function of endwall Thickness
Size: 16mm, Product: MD-2, n=101


y = 1256.9x + 4685.9
R2 = 0.4101


0[
D 0 -


ESP 0


i I i I i i


0 2 4 6 8 10
Endwall thickness (mm)


12 14 16


Figure 4-7. Failure load as a function of endwall thickness


30000


25000


20000


15000


10000


5000








Failure Loads as a Function of Density. Hardness, or Ultrasonic Velocity

Another empirical approach was to hypothesize that the load bearing capacity of

the dowels is related to the quality of the cortical bone. The quality of the cortical bone

was identified by a variety of different methods. These included durometry (Shore D-

scale), ultrasonic velocity, and density of the dowel. A specimen that is composed of

high quality cortical bone may be strong while a specimen composed of low-quality bone

specimen may be weak. Density was calculated by dividing the mass by the volume,

which was determined by Archimedes principle. The failure load for each of these

dowels as a function of the quality measuring technique is shown in Figures 4-8 to 4-11.

The correlation coefficient (r2) for these curves are very poor. In summary, by

accounting for only the quality of the bone, and ignoring details such as how much bone

is available for carrying the load, will result in poor correlation of failure load with

quality. A similar analogy is to examine two different metal rods. If the rods are exactly

the same materials but only differ in diameter, the densities of the two are the same.

However, there will be differences in compressive load carrying ability between these

two rods. The magnitude of this difference depends ONLY on the difference in diameter

(volume of material).












30000

25000 R2 = 0.0145
25000

1 20000

15000 *
,L
10000

0 5000

0
0.0 0.5 1.0 1.5 2.0 2.5
Density (gramlvolume)

16mm MD-2 Linear (16mm MD-2)

Figure 4-8. Failure load of 16mm MD-II as a function of density


30000
2 I = 0.2141
S25000 -

2 20000

15000 a.
U *
S10000

9 5000

0 -
0.0 0.5 1.0 1.5 2.0 2.5
Density (gram/volume)

18mm MD-2 Linear (18mm MD-2)

Figure 4-9. Failure load of 18mm MD-II as a function of density












30000 -
R2 0.0175
25000

i 20000


5 15000

10000

5000

0
0 1000 2000 3000 4000 5000
Ultrasonic Veloicity (m/s)

Figure 4-10. Failure load of 16mm MD-II as a function of ultrasonic velocity in cortical
ring


25000


S 20000
0
o

15000


U- 10000

0o
-1 5000


R2 = 0.0366 *




-*


82 84


88
Hardness


Figure 4-11. Failure load of 16mm MD-II as a function of hardness of cortical bone








Analyzing the Data by Multivariable Regression Analysis

The technique that correlates with the load bearing capacity that most successfully

deals with the heterogeneous nature of the bone dowels is multivariable regression

analysis. It is successful because it is able to quantitatively assess the volume and quality

of cortical bone. In addition, it also accounts for loading variations due to machining

differences. Several factors were considered in order to optimize the regression analysis.

These included dimensional measurements (slotted and unslotted endwall thickness,

diameter mismatch, and length) and mass. The other variables were not considered due

to the poor correlations previously described.

A multivariable regression analysis was performed on these measurements (see

appendix 1). Table 4-1 indicates the best subsets and corresponding r2 value for three

types of dowels similar results were identified for the other dowels. This table indicates

that mass is the most significant predictor. The addition of the other variables contributes

little to the prediction. However, in order to ensure that the specific dowel of interest

adheres to RTI dimensional criteria the endcap dimensions and diameter mismatch were

maintained in the equation. In summary, four variables were identified that met these

criteria. These variables are as follows: Slotted endcap thickness, unslotted endcap

thickness, diameter mismatch, and mass.

All of these variables may be measured directly from the dowel of interest. A

regression equation takes the form of (1) below

Predicted Load = o +Pslotted*X+Punslotted*X2+Pmass*X3+mismatchX4

where p values are the coefficients that were determined by the multiple

regression analysis while the x values are determined from the dowel that is being

evaluated where:








x,- slotted thickness in mm (00.0 mm)

x2- unslotted thickness in mm (00.0 mm)

x3- mass of dowel in grams (0.000 g)

x4- diameter mismatch in mm (0.0mm)

and

p values are coefficients determined from the regression analysis

x values are determined directly from the dowel of interest

The result of this regression is shown in Figures 4-12 through 4-14. The

coefficients as well as the correlation coefficients are illustrated in Table 4-2. This is a

large improvement in the correlation coefficients when compared to the techniques

discussed previously. This technique realized r2 values as high as 0.74. However, there

appears to be a size dependence on the prediction capabilities of this technique. The

prediction capability increases with decreasing size of the dowel. This phenomenon has

repercussions in the Weibull analysis that will be discussed in the next chapter.









Table 4-1 Best regression subsets


16mm MD-2 16mm MD-3 18mm MD-2
Variables R Variables R Variables R

mass 0.740 mass 0.620 mass 0.455
unslotted 0.317 unslotted 0.155 slotted 0.238
mass+mismatch 0.742 mass+slotted 0.664 mass+slotted 0.470
mass+slotted 0.741 mass+length 0.630 mass+mismatch 0.456
mass+mismatch 0.743 mass+slotted 0.671 mass+unslotted+s 0.472
+slotted +length lotted
mass+mismatch 0.744 mass+slotted 0.672 mass+slotted+uns 0.474
+slotted+length +mismatch loted+mismatch
+length

Table 4-2 Regression coefficients for various size dowels

Coefficient Po (N) Islotted Punslotted Pmismatch Pmass (r)
(N/mm) (N/mm) (N/mm) (N/gram)
16mm MD-2 -2588 -161 -23 1500 5078 0.74
16mm MD-3 -2769 -913 15 -2743 5874 0.67
18mm MD-2 1337 586 188 -1495 2573 0.47
20mm MD-2 4644 -415 110 1583 2587 0.42






73


Multivariable Regression analysis of MD dowels
Size: 16mm, Product: MD-2, n=101


30000
25000 -- =+60
20000 R2 = 0.7429 E 3

o 15000
10000
5000


0 5000 10000 15000 20000 25000 30000
Predicted Load (N)

Figure 4-12. Multiple regression analysis of 16mm MD-II bone dowels


Failure Prediction of MD dowels using multvariable
regression
Size: 18mm, Product: MD-2, n=259

30000

25000
y = x + 2E-09 a
E 20000- R2= 0.4735

5 15000- P



5000--

0
0 5000 10000 15000 20000 25000 30000
Predicted Load (N)

Figure 4-13. Multiple regression analysis of 18mm MD-II bone dowels



























Failure Prediction of MD dowels using multivariable
regression
Size: 20mm, Product: MD-2, n=99
30000


25000 -- yx + 1E-09
R2 = 0.418

S20000- CeIpn


S15000 -


10000


5000-



0 5000 10000 15000 20000 25000 30000
Predicted Load (N)

Figure 4-14. Multiple regression analysis of 20mm MD-II bone dowels








Applicability to the bone dowels

The qualitative and quantitative descriptions of strength from the section on

Normalized Strength are applicable. Therefore, the strength ratio (equation 4-4) of the

bone dowels is defined to be the ratio of the actual load to the predicted load.

A variety of strength behaviors are shown in nature. Generally metals have

predictable strength, or simply are well behaved, and specimens derived from them

would be expected to have a strength ratio of 1.0. Conversely, the strength of ceramics is

controlled by the presence and the size of flaws. Because the flaw size is a Gaussian

distribution, the strength is also Gaussian. The behavior of dowels is similar to ceramics;

however, the cause of this behavior has not been identified. Figures 4-5 and 4-7

illustrates the Gausssian distribution for the strength of the various sizes and types of

bone dowels. When designing for the strength of bone dowels, well-established

techniques from the ceramics field are appropriate. In particular, the Weibull modulus

technique for identifying a probability of survival or failure at a particular strength level

is appropriate.

Conclusions

A systematic method of assessing the strength of the MD-series of bone dowels

has been defined. This technique may be utilized for a variety of purposes. It may be

utilized to assess differences in strength of the MD-series of bone dowels such as:

1. design changes to the dowels

2. immersion in solutions such as isopropyl alcohol, hydrogen peroxide

3. donor characteristics such as age, or sex

4. certain disease processes such as osteoporosis, steroid use

5. Storage variables such as freeze-drying, and/or reconstitution, freezing





76


6. Sterilization techniques such as irradiation, ethylene oxide, heat

In addition, this definition may be combined with a statistical treatment of

fracture such as Weibull analysis to establish a biomechanical quality control procedure

in order to identify any structurally deficient dowels.














CHAPTER 5
BIOMECHANICAL QUALITY CONTROL PROCEDURE

Introduction

The dowels may be placed under a range of axial compressive loads that

are dictated by the age, activity level, and general health of the recipient of the dowel. In

addition, the applied loads will vary depending on such factors as the spinal level where

the dowel is placed, the surgical technique as well as if any auxiliary plating or rod

systems are utilized. In addition, as can be confirmed in Figure 4-6, the larger diameter

dowels carry more load. However, it is impractical to know which dowel will be going

into which patient. Therefore all dowels must adhere to an acceptable quality criterion.

This criterion is partially defined by a minimum load bearing capacity. Concomitantly, a

systematic method of removing structurally deficient dowels must be explored.

The purpose of this chapter is to illustrate how the methods developed in the

previous chapter combined with a quality criterion for the bone dowels may be applied to

the well established Weibull technique. The Weibull analysis is often utilized to prevent

failure when failure is controlled by statistical phenomena (56, 147).

Materials and Methods

Normalized Strength of Bone Dowels

Typically, strength is defined by a failure stress (either yield or ultimate). This

involves normalizing the load carrying ability by dividing by the specimens cross

sectional area. Unfortunately, bone dowels are not amenable to this traditional

engineering approach to failure. However, the previous chapter identified a method or








normalizing the strength by multivariable regression analysis. Thus the normalized

strength may be defined to be the ratio of the actual load to predicted load.

Acceptable Quality Criteria

By applying our definition for strength, identifying an acceptable quality criteria,

and combining this with the statistical treatment of strength based on Weibull statistics as

outlined by Wachtman or Felbeck and Atkins (56, 147), a minimum predicted load may

be isolated. Any dowel that is found to have a predicted load above this value has

sufficient strength to survive in-vivo axial compressive loads and is appropriate for

implantation. Conversely, any dowel that is found to have a predicted load below this

value is deemed to have insufficient strength to survive in-vivo axial compressive loads

and is rejected for implantation.

Design Loads on the Dowels

The vertebral disk has been reported to fail at 10 000 N, while the maximum

applied load on the spine has been reported to be 7200N. Based on the maximum applied

load, a two-dowel construct, and a factor of safety of 3.0, a design load may be specified.

This design load was determined to be 10800N (7200*(3.0/2)). This design load

represents maximum static axial compressive load that potentially could be applied to the

dowel in-vivo. It is the goal of the nondestructive evaluation (NDE) procedure to

eliminate any dowel that potentially may not sustain a load of 10800 N.

Probability of Survival

Any probability may be chosen; however the lower the probability of survival the

greater the risk of premature failure. The current literature advocates that when a human

life is at stake, the probability of survival should be defined as 99.9999%.








Weibull Approach Applied to the Bone Dowels

A convenient approach to applying this technique to bone dowels was to separate

the dowels by design (i.e. MD-II, MD-III, and MD-IV). These were then sorted by

diameter. The multiple regression analysis was performed and the normalized strength

(defined as the ratio of actual to predicted) was determined.

The normalized strength was sorted from smallest to largest and the natural log of

the value computed. These values were numbered from 1 to n where n is the sample size.

One estimate of the probability of survival can be described as noted by Wachtman (147)

is


P i 0.5 Equation 5-1
n


where i is the assigned value in the numerically ranked list from 1 to n. The

inverse of the probability of survival is computed and the natural log of this value is taken

twice. A plot was constructed of Ln(Actual/Predicted) versus Ln(Ln(1/Ps) and the slope

(m) and intercept (b) determined. The slope represents the Weibull modulus. Once again

Wachtman presents the following expression for Weibull:


o- = e-b Equation 5-2



Finally, the probability of survival may be defined by the following expression



S exp- ected) lowlimit ) Equation 5-3
P=exp\ L [ c J /










Lowlimit represents the lowest normalized strength ratio that would be expected

to be observed. This value was defined to be 70% of the lowest observed normalized

strength or 0.44. The details for calculating the Weibull modulus for 16mm MD-2

dowels are described in the appendix.

The goal is to isolate the threshold load in order for the dowel to survive our

design criteria (i.e. 10800N load @ 99.9999%). This threshold load is compared to the

predicted load (via the regression equation). If the predicted load is equal to or below

this threshold load the dowel is rejected for implantation. If the computed predicted load

is greater than the threshold load, the dowel is accepted for implantation. If equation 5-3

is solved for the predicted load the following expression results:


-10800N
Predicted threshold = 1 N Equation 5-4
lowlimit (- ln(P, ))q '



Where the variables have been defined previously.

Validation

A validation was performed to ensure that the nondestructive technique

effectively evaluates the axial compressive strength of the MD-Series of dowels.

Comparing nondestructive analysis to destructive analysis accomplished this validation.

Once validated, the nondestructive analysis can then be used as a mechanical quality

control procedure to verify the strength and therefore safety of the dowels.








Specimen Preparation

Dowels were manufactured in clean rooms by typical dowel manufacturing

techniques and transferred to the Biomechanics lab for testing. Two or more dowels

were provided from each donor. Typically, processing personnel chose these dowels,

because of their perceived sub-optimal structural characteristics. However, all dowels

included in the validation study met the following criteria: There were full and complete

threads across the entire length of the top and bottom surfaces of the dowel. Dowel

diameter was within dowel specifications. Dowel diameter mismatch was less than or

equal to 0.3mm. Diameter, and endcap dimensions as well as mass were recorded by the

typical production techniques.

Sample Size

Table 5-1 illustrates the minimum sample sizes required to complete the

validation study.


Table 5-1 Minimum sample size for validation

Dowel Type Dowel Size # of specimens # of donors
MD-2 16 40 >3
MD-2 18 40 >3
MD-2 20 40 >3
MD-3 16 40 > 3


Nondestructive and Destructive Analysis

Predicted load analysis was performed utilizing the regression equations

identified in chapter 3. This value was compared to the threshold to identify a pass or

fail. Destructive testing was completed using the 858 Bionix MTS Servohydraulic

mechanical test machine per the methods described earlier. All data from nondestructive

evaluation and destructive test were compared for final analysis.








Possible Outcomes

Comparison between nondestructive and destructive results may produce four

possible outcomes as described in Table 5-2. In an ideal situation, all evaluated dowels

will be correctly predicted to sustain loads above or below the design load. However, it

is recognized that errors may occur. Acceptable errors mistakenly reduce the number of

dowels available for transplantation. Although undesirable, acceptable errors provide a

buffer between dowels with correctly predicted outcomes and those with unacceptable

errors. Unacceptable errors may result in poor clinical outcomes.


Table 5-2 Possible Outcomes for validation

Nondestructive Evaluation Destructive Evaluation Possible Outcome
Predicted to survive 10800N Sustain load above 10800 N Correctly predicted
Predicted to survive 10800 N Sustains load below 10800 N Unacceptable error
Predicted to fail 10800 N Sustains load above 10800 N Acceptable error
Predicted to fail 10800 N Sustains load below 10800 N Correctly predicted


Acceptable Outcomes

In order for the NDE procedure to be validated, the following criteria were met.

The number of unacceptable errors for each size and type were set to zero. If any

unacceptable errors result, the nondestructive evaluation procedure must be revised.

Results

Weibull Modulus, Intercept and Threshold Load

Table 5-3 identifies the m, b, lowlimit, design load, and ao for each of the dowels

that have been analyzed. The sample size (n) used to compute these parameters is also

given.









Table 5-3 Weibull analysis results for various dowels


Design Diameter n m b co Lowlimit Design load Threshold
(kN) load (kN)
MD-II 16 101 14 -0.5 1.0 0.44 10.8 13.0
MD-II 18 259 10 -0.5 1.1 0.44 10.8 15.3
MD-II 20 99 12 -0.5 1.0 0.44 10.8 13.8
MD-II 22 36 11 -0.5 1.0 0.44 10.8 15.8
MD-II 24 31 7 -0.5 1.1 0.44 10.8 15.9
MD-III 16 47 9 -0.5 1.1 0.44 10.8 14.5
MD-IV 18 44 9 -0.5 1.1 0.44 10.8 18.3


Plots of Ps versus Actual/Predicted may be constructed. These curves are typical

of survival curves that are well documented in the ceramics literature. Plots of the 16 and


18mm MD-2 dowels are shown in Figure 5-1.






84










Probability of Survival as a function of strength
MD-2 bone Dowels



D
00

0*


[]
0
0





*0
0
--
D


0.50


0.75


1.00


1.25


1.50


Strength (Actual:Predicted)
*16mm MD-2 18mm MD-2


Figure 5-1. Probability plots of 16 and 18mm MD-2 dowels


100%


80%


60%


40%


20%


0%








It is an interesting observation of Table 5-3 and Figures 4-9 and 4-10 if the 16 and

18 mm MD-2 dowels are compared, the 18mm dowels have a average failure strength

greater than the 16mm dowels. Surprisingly, the threshold load is greater for the 18mm

than the 16mm MD-2. The reason for this is that the regression analysis for the 18mm

has a lower correlation coefficient, which translates into a lower Weibull modulus (10

versus 14). One could deduce from Table 5-3 that the 24mm MD-2 has the poorest

correlation coefficient; however the Weibull analysis allows these inequities to be

satisfied by increasing the threshold.

Validation Results

Table 5-4 illustrates the results of the validation procedure. The actual number of

specimens that were tested exceeded the minimum sample size. All dowels met the

requirements described in the study design. Out of the 606 dowels tested, 590 dowels

(97.4%) were correctly predicted to sustain loads above the design load and 16 dowels

(2.6%) were predicted to fail at the design load but actually sustained loads above the

design load (an acceptable error). As required, zero unacceptable errors resulted.


Table 5-4 Validation actual outcomes

Type Dowel # of # of Correctly Correctly Acceptable Unacceptable
Size Dowels Donors Predicted to Predicted Error Error
Pass to Fail
MD-2 16 104 89 101 0 3 0
MD-2 18 215 201 203 0 12 0
MD-2 20 188 171 188 0 0 0
MD-3 16 99 86 98 0 1 0


Results in Manufacturing Since Implementation of NDE at RTI

In July of 1998, RTI evaluated 7,027 dowels by this procedure. Of these, 6,869

(97.8%) of the dowels met or exceeded both dimensional and structural loading








requirements. 133 (1.9%) dowels failed for dimensional reasons while 25 (0.3%) failed

for structural reasons. The average predicted load for the MD-Series dowels is 1.44 times

stronger than the minimum threshold.

In addition, 13 dowels that were rejected by the NDE procedure were retrieved

and destructively tested. 2 of these 13 failed below 10,800N. These rejected dowels had

an average failure load of 12,725 N. MD-II and MD-III dowels sent to the Biomechanics

lab for destructive evaluation fail at an average of 17,067 N (16mm MD-II) and 18,297

(16mm MD-III).

It should be noted that these 13 dowels met all dimensional requirements and

could have been released as product if the nondestructive evaluation procedure was not in

place.

In October of 1999, a histogram of the predicted loads of the 16mm MD-2 dowels

(n=8134) that have been evaluated since the process was adopted by RTI was plotted and

is shown in Figure 5-2. Plots of 18 and 20mm MD-2 dowels are similar. The threshold

predicted load for this size and design (13029N) is also shown on the plot. It is clear that

many of the dowels that are manufactured by RTI are well above the acceptable quality

standard that has been established for the dowels.






















700



600



500



>, 400


I-I
c*
LL 300



200



100




0 5000 10000 15000 20000 25000
Predicted Load


Figure 5-2. Typical results when applied to manufacturing environment for 16mm MD-2


30000








Conclusions

An acceptable quality criterion can be identified for allograft materials.

An empirical treatment of failure identified by regression analysis may be

successfully combined with a quality criteria and Weibull analysis in order to establish a

nondestructive quality control procedure.

The nondestructive evaluation procedure was successfully validated. This process

accurately predicts whether or not a MD-Series of dowels will survive axial compressive

design loads.




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EKUOL5NJ7_1OMJPN INGEST_TIME 2013-10-09T23:54:27Z PACKAGE AA00017681_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES