UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Mechanical characteristics in impaction allografting - the role of graft density and cement penetration… Albert, Carolyne Izabel 2009

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2010_spring_albert_carolyne.pdf [ 3.63MB ]
Metadata
JSON: 24-1.0070873.json
JSON-LD: 24-1.0070873-ld.json
RDF/XML (Pretty): 24-1.0070873-rdf.xml
RDF/JSON: 24-1.0070873-rdf.json
Turtle: 24-1.0070873-turtle.txt
N-Triples: 24-1.0070873-rdf-ntriples.txt
Original Record: 24-1.0070873-source.json
Full Text
24-1.0070873-fulltext.txt
Citation
24-1.0070873.ris

Full Text

MECHANICAL CHARACTERISTICS IN IMPACTION ALLOGRAFTING – THE ROLE OF GRAFT DENSITY AND CEMENT PENETRATION PROFILE by CAROLYNE IZABEL ALBERT B.A.Sc., Université de Sherbrooke, 1997 M.A.Sc., University of British Columbia, 2001  A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Materials Engineering)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) October 2009 © Carolyne Izabel Albert, 2009  Abstract Revision total hip arthroplasty with femoral impaction allografting has an attractive potential for restoring bone stock in femurs with bone loss caused by the failure of hip implants. However, problematic implant subsidence is often reported after this procedure. A lack of understanding remains over the mechanisms that cause subsidence. The objectives of this study were to: a) explore the relationships between subsidence and morphometric features of the graft and bone cement regions after femoral impaction allografting in a cadaveric femur model; b) characterize mechanical properties of the graft bed as a function of impaction force, and explore new alternative graft compaction methods; and c) develop a finite element model to investigate the key mechanisms that contribute to initial implant subsidence. High levels of cement penetration into the graft bed were observed, resulting in extensive regions of cement contact with the host bone in a cadaveric femur model. The implant subsidence correlated negatively with the amount of cement-endosteum contact. The density, compression stiffness and shear strength of the graft were proportional to the impaction force. A slower alternative graft compaction method resulted in higher graft stiffness and shear strength than traditional graft impaction, but the benefit of this new compaction method was small compared to the effect of increasing the impaction force. In a finite element model, the relationship between graft density and subsidence was dependant on cement penetration profile. Without cement-endosteum contact, subsidence decreased with increasing graft density; however, graft density did not affect subsidence in constructs with cement-endosteum contact. Initial subsidence was primarily attributed to slippage at the stemcement and endosteum interfaces, and the latter mechanism was greatly affected by changes in graft density and cement penetration profile.  ii  This study demonstrated that extensive cement penetration can occur in femoral impaction allografting, which may compromise the potential for new bone formation but may be important in preventing excessive subsidence. The endosteum interface was identified as a key factor in the development of subsidence. Finally, our results indicate that the potential benefit of achieving a denser graft bed depends upon the cement penetration profile.  iii  Table of Contents Abstract ....................................................................................................................................... ii Table of Contents....................................................................................................................... iv List of Tables .............................................................................................................................. vi List of Figures ........................................................................................................................... vii Acknowledgements ..................................................................................................................... x Co-authorship Statement ......................................................................................................... xii Chapter 1. Introduction ........................................................................................................... 13 1.1 Total Hip Arthroplasty........................................................................................................ 2 1.2 Failure and Revision Options ............................................................................................. 3 1.3 Femoral Impaction Allografting ......................................................................................... 5 1.3.1 Technique .................................................................................................................... 5 1.3.2 Indications and use of the technique............................................................................ 7 1.3.3 Femoral fracture........................................................................................................... 9 1.3.4 Implant subsidence .................................................................................................... 10 1.3.5 Cement penetration into the graft bed ....................................................................... 15 1.4 Summary and Project Motivation..................................................................................... 16 1.5 Objectives and Scope........................................................................................................ 17 1.6 References ........................................................................................................................ 18 Chapter 2. Cement Penetration and Primary Stability of the Femoral Prosthesis - a Biomechanical Study in the Cadaveric Femur ...................................................................... 24 2.1 Introduction ...................................................................................................................... 25 2.2 Methods ............................................................................................................................ 26 2.3 Results .............................................................................................................................. 31 2.4 Discussion......................................................................................................................... 38 2.5 References ........................................................................................................................ 43 Chapter 3. The Effect of Impaction Force and Alternative Compaction Methods on the Mechanical Characteristics of Morsellized Cancellous Graft .............................................. 47 3.1 Introduction ...................................................................................................................... 48 3.2 Methods ............................................................................................................................ 50 3.3 Results .............................................................................................................................. 56 3.4 Discussion......................................................................................................................... 62 3.5 References ........................................................................................................................ 71  iv  Chapter 4. Influence of Cement Profile and Graft Properties on Stem Migration and Micromotion - a Finite Element Study ................................................................................... 75 4.1 Introduction ...................................................................................................................... 76 4.2 Methods ............................................................................................................................ 77 4.3 Results .............................................................................................................................. 81 4.4 Discussion......................................................................................................................... 86 4.5 References ........................................................................................................................ 91 Chapter 5. General Discussion and Conclusions ................................................................... 97 5.1 Discussion......................................................................................................................... 98 5.1.1 Cement penetration profile ........................................................................................ 98 5.1.2 Effects of cement profile and graft density on initial implant subsidence ................ 99 5.1.3 Subsidence mechanisms .......................................................................................... 101 5.1.4 Postoperative biological changes............................................................................. 101 5.1.5 Relevance of current work with respect to other in vitro studies ............................ 110 5.1.6 Trends in impaction allografting research ............................................................... 111 5.2 Contributions .................................................................................................................. 115 5.3 Conclusions .................................................................................................................... 116 5.4 Future Work.................................................................................................................... 117 5.5 References ...................................................................................................................... 118 Appendix 1. Mechanical Testing Set-ups ............................................................................. 127 Appendix 2. Histomorphometric Analysis Methods ........................................................... 139 Appendix 3. Graft Shear and Compression Testing ........................................................... 143 Appendix 4. Finite Element Analysis Mesh and Stem Motion Results.............................. 146 Appendix 5. UBC Research Ethics Board Certificates of Approval.................................. 160 Appendix 6. Glossary ............................................................................................................. 164  v  List of Tables Table 1.1 Published clinical articles on femoral impaction allografting. ...................................8 Table 2.1 Description of specimens...........................................................................................27 Table 2.2 Cement contact with endosteum, cement area and graft porosity for the pressure and no-pressure group at eleven matched-level cross-sections.........................................................33 Table 2.3 Migration components for the pressure and no-pressure groups. ..............................33 Table 2.4 Micromotion components for the pressure and no-pressure groups..........................34 Table 2.5 Pearson’s correlation coefficients (R) and slopes for the linear regressions between the motion components and the cement morphological parameters (average cement contact and average cement area). .................................................................................................................36 Table 3.1 Test matrix. Graft compaction techniques and number of specimens used for compression, shear, and density tests. ........................................................................................51 Table 3.2 Results of the creep compression test. .......................................................................57 Table 3.3 Mohr-Coulomb parameters versus strength definition for three impaction forces: 300 N (I300N), 600 N (I600N) and 900 N (I900N), and for the creep compaction method (C300N). Also shown is the shear strength, τf, under a normal stress, σn, of 375 kPa...............................59 Table 3.4 Graft density and percentage bone as a function of impaction force.........................61 Table 4.1 Graft properties used in the finite element model, representing low, moderate or high graft density. ...............................................................................................................................80 Table 4.2 Contribution of each material and interface to the distal implant motion relative to the bone.......................................................................................................................................86  vi  List of Figures Figure 1.1 Anatomy of the hip (left) and primary total hip arthroplasty components (right). ....3 Figure 1.2 Endo-Klinik classification system for femoral bone stock deficiency.......................4 Figure 1.3 Surgical options for revision THA. (a) Girdlestone arthroplasty. (b) Structural allograft. (c) Cemented revision. (d) Cementless revision with a larger prosthesis. (e) Replacement of the proximal femur with a megaprosthesis.........................................................5 Figure 1.4 Impaction allografting technique. (a) Insertion of distal plug, with guide wire attached. (b) Distal impaction of the morsellized graft. (c) Proximal impaction of the graft. (d) Injection and pressurization of bone cement into the stem-shaped neo-medullary canal formed by the impacted graft. (e) Stem insertion......................................................................................6 Figure 1.5 CPT impaction allografting tools (Zimmer Inc., Warsaw, Indiana). .........................7 Figure 1.6 Settling of the stem within the cement mantle. A polyethylene centralizer (insert) is usually implanted at the tip of smooth tapered cemented stems to provide a space into which the stem tip can subside. Settling of the stem into the cement is expected to generate compressive radial stresses within the surrounding cement and graft regions...........................11 Figure 1.7 Stem subsidence relative to the cement due to longitudinal cement fracture. .........12 Figure 1. 8 Subsidence due to compression deformation of the graft bed (left), shear failure in the graft bed (center), and slippage at the endosteum interface (right). .....................................13 Figure 1.9 Subsidence caused by expansion of the femoral canal.............................................14 Figure 1.10 Extensive cement penetration observed after impaction allografting in a cadaveric femur model. (a) Cross-section taken from the metaphysis region, i.e., proximal femur. (b) Cross section from the diaphysis region, around mid-stem........................................................16 Figure 2.1 Diagram showing the loading set-up. The potted femur was mounted on a linear guide to avoid undesired horizontal reaction forces in the frontal plane. A craniocaudal load (Fcc) was applied with the linear actuator, generating the mediolateral (Fml) and proximodistal (Fpd) components. The anteroposterior load (Fap) was applied by controlling the moment, M, applied with the rotary actuator (M = Fap × offset). ...................................................................28 Figure 2.2 Photograph of the motion measurement system. Six linear variable differential transformers were mounted on an aluminum frame that was rigidly attached to the femur. They measured the motion of a triangle that was fixed rigidly to the lateral side of the implant through a hole in the femur. The relative motion between implant and bone was determined from that of the triangle, relative to the frame............................................................................29 Figure 2.3 Graph showing cement contact with endosteum for the pressure and no-pressure groups, at eleven matched-level cross-sections. .........................................................................32 Figure 2.4 Graphs showing distal migration (left), valgus/varus rotation migration (center) and distal micromotion (right) for the two groups. ...........................................................................34 Figure 2.5 Graph showing distal micromotion (top) and migration (bottom) versus number of walking cycles for all specimens tested......................................................................................35  vii  Figure 2.6 Scattergraph showing distal migration as a function of: (top) the average cement contact with endosteum (averaged for each specimen over all 11 levels); (bottom) the average cement area. ................................................................................................................................37 Figure 3.1 Particle size distribution for the morsellized graft. ..................................................50 Figure 3.2 Compaction techniques. Impaction (left): Peak compressive forces of 300, 600 and 900 N (1, 2, and 3 MPa) were used. Creep (center): a compressive force of 300 N was held for 90 s. Cyclic relaxation (right): a force of 300 N was applied ten times, holding displacement constant for 9 seconds after each loading. ..................................................................................51 Figure 3.3 Compression tests (top) consisted of one hour creep tests. Shear tests (bottom) were performed at three normal stress levels, σn.. ..............................................................................52 Figure 3.4 Definition of initial compression stiffness, Ei, and stiffness after one hour creep, E1hr, under an applied stress of 1.1 MPa.....................................................................................53 Figure 3.5 Strain during creep test. The specimen was compressed initially from 0.034 to 1.1 MPa during a 1 s loading ramp. Subsequently, the strain was measured during 1 hour of creep at 1.1 MPa. Initial strain, εi, and creep constant, Cε, are defined as shown...............................54 Figure 3.6 Histological measurement of graft density, quantified as the percentage of the specimen area occupied by bone. ...............................................................................................56 Figure 3.7 Shear stress (black) versus shear displacement curve for a typical specimen. In grey is the same curve when the shear stress is derived with the initial specimen area, i.e., without accounting for the changing cross-sectional area. ......................................................................58 Figure 3.8 Results of the shear tests - shear strength, τf, versus normal stress, σn, for all impaction forces: 300 N (I300N), 600 N (I600N) and 900 N (I900N), and for the creep compaction method (C300N). ....................................................................................................58 Figure 3.9 Post-impaction stiffness, Ei (top) and E1hr (bottom) versus graft density for each force group..................................................................................................................................60 Figure 3.10 Creep constant during post-impaction creep compression versus graft density for each force group. ........................................................................................................................61 Figure 3.11 Median stiffness of the graft during impaction, Eimp, and during subsequent creep test for all impaction forces. .......................................................................................................62 Figure 4.1 Geometries of the nine cement penetration profiles modeled. Profiles A-C: idealized constructs with uniform cement penetration that reached depths of 25%, 50%, and 75% of the graft bed width, respectively. Profiles D-I: constructs with 11%, 24%, 36%, 49%, 61% and 80% cement contact with the endosteal surface, respectively.....................................78 Figure 4.2 Shear stress at endosteum interface for profiles A-C (top) and D-I (bottom). For profiles D-I, simulations in which the cement-endosteum contact region was bonded are shown in grey, and those in which the entire endosteum interface was sliding are shown in black. ....82 Figure 4.3 Implant migration (top) and micromotion (bottom) for cement penetration profiles A-C. The grey data points represent constructs in which the stem-cement interface was bonded, while the black points represent those in which this interface was sliding, i.e., debonded........84  viii  Figure 4.4 Implant migration (top) and micromotion (bottom) for profiles D-I. Simulation results are shown in black and grey, and experimental data are shown as white triangles (after 10 cycles of walking loads) and white squares (after 5000 cycles)............................................85 Figure 5.1 Relationship between distal migration (left) and micromotion (right), and graft density after impaction allografting in cadaveric femurs (Chapter 2)......................................100 Figure 5.2 Major mechanisms of subsidence for femoral impaction allografting constructs without (left) and with (right) cement-endosteum contact. ......................................................104 Figure 5.3 Comparison of migration (top) and micromotion (bottom) finite element results when the graft shear failure criterion was included in the model versus when it was not, for cement profile A, with a low graft density and when the stem-cement interface was defined as sliding. ......................................................................................................................................105  ix  Acknowledgements I would like to thank Dr. Göran Fernlund and Dr. Thomas Oxland for their guidance. This project has been a tremendous learning opportunity, and it would not have been possible without them. Thanks also to Dr. Bassam Masri and Dr. Clive Duncan for their clinical input which has been helpful in understanding the context of this research. I also wish to thank Dr. Sanjeev Patil who, in spite of his busy schedule during his fellowship at Vancouver General Hospital, kindly agreed to perform all the surgical procedures. Thanks to Dr. Hanspeter Frei for sharing his knowledge and inspiration, and to Dr. Youngbae Park for the wonderful collaborations. Special thanks to Jesse Chen for his assistance in harvesting the graft and to both him and Caron Fournier for their skillful preparation of the histological slides. To Dr. Qing-An Zhu, Dr. Juay Seng Tan, Dr. Danmei Liu, Tim Schwab, and to the other members of the Division of Orthopaedic Engineering Research, I am grateful for all the insightful discussions. And a very special thanks to Emily McWalter for her support in the lab. Thanks to Deetria Egeli for her assistance with the pQCT measurements, and to Dr. Saija Kontulainen for teaching me how to analyze these scans. Thanks to Kevin Siggers for transforming my sketches of the shear box into a functional testing apparatus and to Ross McLeod and Carl Ng from the Materials Engineering Department machine shop for machining some of the parts. Thanks also to Robert Courdji whose help was instrumental in getting the finite element model up and running. I gratefully acknowledge the Michael Smith Foundation for Health Research, the Canadian Institutes of Health Research, and Cy and Emerald Keyes for their financial contributions. Finally, my deepest gratitude goes to Dr. Meress and the staff at the Fox Valley Wellness Center, for my completion of this thesis would not have been possible without their help. x  This thesis is dedicated to my husband Jeff, for his love and immeasurable support.  xi  Co-authorship Statement Sections of this thesis have been submitted or published as multi-authored papers in refereed journals. Details of the authors’ contributions are provided. Chapter 2 Albert C, Patil S, Frei H, Masri B, Duncan C, Oxland T, Fernlund G (2007). Cement penetration and primary stability of the femoral prosthesis after impaction allografting: A biomechanical study in the cadaver femur. Journal of Bone and Joint Surgery British Edition. 89 (7): 962-70. Authors’ contributions: The origin of this project was a grant proposal written by Thomas Oxland, Goran Fernlund, Bassam Masri and Clive Duncan. Carolyne Albert was responsible for the design of this study based on ideas from the grant proposal, conduction of the experiments, analysis and presentation of the findings, and writing and editing of the original paper. Göran Fernlund and Thomas Oxland supervised the project and were the key editors of this paper. Sanjeev Patil conducted the surgical procedures on the cadaveric femora and provided editorial assistance. Hanspeter Frei stimulated discussions and provided editorial assistance. Bassam Masri and Clive Duncan provided clinical input and editorial assistance. Chapter 3 Albert C, Masri B, Duncan C, Oxland T, Fernlund G (2008). Impaction allografting – The effect of impaction force and alternative compaction methods on the mechanical characteristics of the graft. Journal of Biomedical Materials Research Part B: Applied Biomaterials. 87 (2): 395-405. Authors’ contributions: Carolyne Albert was responsible for the original ideas behind the paper, conduction of the experiments, analysis and presentation of the findings, and writing and editing of the original paper. Thomas Oxland provided guidance as well as editorial assistance. Bassam Masri and Clive Duncan provided editorial assistance. Göran Fernlund was the key editor on this paper. Chapter 4 Albert C, Masri B, Duncan C, Oxland T, Fernlund G (2009). Influence of cement penetration and graft density on stem stability in impaction allografting: A finite element study. Clinical Biomechanics (in press). Authors’ contributions: Carolyne Albert was responsible for the original ideas behind the paper, construction of the finite element model, analysis and presentation of the findings, and writing and editing of the original paper. Göran Fernlund was the key editor on this paper. Thomas Oxland provided guidance as well as editorial assistance. Bassam Masri and Clive Duncan provided editorial assistance.  xii  CHAPTER 1 INTRODUCTION  xiii  1.1 Total Hip Arthroplasty Musculoskeletal conditions are the leading cause of long-term pain and physical disability, affecting hundreds of millions of people worldwide (Woolf and Pfleger, 2003). Approximately 4.4 million Canadians suffer from arthritis (Statistics Canada, 2008), which affects many aspects of their lives including daily and social activities, and their ability to work. Musculoskeletal conditions constitute the second most costly diagnostic category in Canada in 1998, these conditions were responsible for an estimated CDN$16.4 billion in direct (medical expenses) and indirect (loss of productivity) healthcare costs in Canada (Health Canada, 1998). In the United States, the total annual healthcare costs of musculoskeletal conditions in 2004 was US$849 billion – almost 8% of the national gross domestic product (American Academy of Orthopaedic Surgeons, 2008). Moreover, the prevalence of musculoskeletal problems is rising due to the ageing population and sedentary lifestyles (American Academy of Orthopaedic Surgeons, 2008). The number of people who suffer from arthritis is expected to rise by more than 50% between the years 2001 and 2026 (Health Canada, 2003). Total hip arthroplasty (THA) can be an effective treatment for musculoskeletal conditions affecting the hip, such as severe osteoarthritis, rheumatoid arthritis, congenital deformities, hip fractures, and femoral head necrosis (Karrholm et al., 2007). In THA, the diseased or injured hip joint is replaced with implants: an acetabular cup with an ultra-high molecular weight polyethylene liner, and a femoral stem which may be press-fit or cemented into the femur using polymethyl methacrylate bone cement (Figure 1.1).  2  Pelvis Acetabulum Acetabular cup and liner  Femoral head  Femoral head Femur  Stem  Cancellous (trabecular) bone Cortical bone  Cement  Endosteum  Figure 1.1 Anatomy of the hip (left) and primary total hip arthroplasty components (right). Adapted from American Academy of Orthopaedic Surgeons, 2007. In general, THA surgeries have very good outcome. Prior to undergoing THA, most patients experience continuous pain that limits their ability to perform common daily activities such as walking or kneeling. After surgery, most of them become pain free and are able to resume jobs as well as some physical activities. For this reason, THA surgeries have become popular around the world. Over 25,000 such procedures are performed annually in Canada (Canadian Institutes for Health Information, 2005). Moreover, the demand for THA is growing Canadian hospitals have observed a 101% increase in hospitalizations for hip and knee replacements between 1996 and 2006 - and these procedures are performed increasingly in younger patients (Canadian Institute for Health Information, 2008).  1.2 Failure and Revision Options Hip implants, however, have a finite life span. The survival rate of primary THA is 80 to 98% at 10 years post surgery (Karrholm et al., 2007). Failure of the femoral and/or acetabular component can require revision THA surgery to replace the failed component(s). Revisions represent 10-20% of all THA procedures (Karrholm et al., 2007; Kurtz et al., 2005), and they are becoming more and more common. In the United States, the number of revision THA surgeries performed annually increased by as much as 79% between 1990 and 2002 (Kurtz et al., 2005). The leading causes for revision THA are aseptic loosening, instability or 3  dislocation, and infection (Karrholm et al., 2007; Bozic et al., 2009). The most common indication for revision of the femoral component is aseptic loosening; i.e., a mechanical loosening of the implant from the surrounding bone (Bozic et al., 2009). Loosening of the femoral component is often accompanied by bone loss, which can make revision procedures much more challenging than primary THA. The severity of the bone loss is variable; it can be classified, for example, using the Endo-Klinik classification system (Figure 1.2).  I  II  III  IV  Figure 1.2 Endo-Klinik classification system for femoral bone stock deficiency. Type I: radiolucent zone confined to the upper half of the cement mantle with clinical signs of loosening. Type II: generalized radiolucent zones and endosteal erosion of the proximal part of the femur, leading to widening of the medullary cavity. Type III: widening of the medullary cavity by expansion of the proximal part of the femur with possible perforation. Type IV: gross destruction of the proximal and middle third of the femur. Adapted from Engelbrecht E. and Heinert K., 1987.  In the presence of significant bone deficiency in the proximal femur, a number of revision options are available to the surgeon (Figure 1.3). These include: Girdlestone arthroplasty, i.e., removal of the implants and allowing the formation of scar tissue without rebuilding the joint surfaces; the use of a longer/larger stem with or without cement; replacement of the proximal femur with a structural allograft or a megaprosthesis; and impaction allografting (Duncan et al., 1998).  4  a.  b.  c.  d.  e.  Figure 1.3 Surgical options for revision THA. (a) Girdlestone arthroplasty. (b) Structural allograft. (c) Cemented revision. (d) Cementless revision with a larger prosthesis. (e) Replacement of the proximal femur with a megaprosthesis.  1.3 Femoral Impaction Allografting 1.3.1 Technique In femoral impaction allografting, the bone deficient proximal femur is reconstructed with particles of cancellous bone graft, obtained from donors, before a new implant is cemented in place. The graft, usually harvested from femoral heads but occasionally from distal femurs or proximal tibiae, is morsellized with a bone mill. The failed femoral stem is first removed along with the bone cement if present, and the canal is cleared of any debris. A tight-fitting polyethylene plug (distal plug) is inserted in the femoral canal to a distance of approximately 2 cm distal of the most distal bone defect (Figure 1.4 a). The femoral cavity is gradually filled with the morsellized cancellous bone graft (MCB), impacted first using flat-headed distal impactors (Figure 1.4 b), then using proximal impactors which bear the oversized shape of the new stem (Figure 1.4 c). Once the medullary cavity is filled with impacted MCB, the proximal impactor is left in place temporarily while further MCB is packed around the impactor using proximal packers (Figure 1.5). After graft impaction, the neo-medullary canal (newly formed stem-shaped cavity) is filled with bone cement. The bone cement consists of a liquid (methyl methacrylate monomer) and a powder (polymethyl methacrylate) which, when mixed together, polymerize in an exothermic reaction within approximately 10 minutes after they are mixed  5  together. The cement is injected into the canal early in the polymerization process, and is pressurized using a cement gun (Figure 1.4 d). Finally, when the cement has reached the desired ‘doughy’ consistency, the new stem is slowly and carefully inserted in place (Figure 1.4 e). A set of surgical instruments used for femoral impaction allografting (CPT, Zimmer Inc., Warsaw, Indiana) is shown in Figure 1.5.  stem cement mantle  graft distal plug  a.  b.  c.  d.  e.  Figure 1.4 Impaction allografting technique. (a) Insertion of distal plug, with guide wire attached. (b) Distal impaction of the morsellized graft. (c) Proximal impaction of the graft. (d) Injection and pressurization of bone cement into the stem-shaped neomedullary canal formed by the impacted graft. (e) Stem insertion.  6  Distal impactors  Central guide wire  Proximal packers Proximal impactors  Figure 1.5 CPT impaction allografting tools (Zimmer Inc., Warsaw, Indiana).  1.3.2 Indications and use of the technique The femoral impaction allografting technique was introduced in the early 1990s as a means to reverse problematic bone loss that is often encountered after THA (Ling et al., 1993; Gie et al., 1993; Nelissen et al. 1995). Good clinical results have been reported, including alleviation of pain and improvement of function (Gie et al., 1993; Leopold et al., 1999), and this procedure is now performed in patients with varying degrees of bone loss, e.g., types I to IV according to the Endo-Klinik classification system (Sierra et al., 2008). Impaction allografting is also performed in patients of various ages, e.g., 31-89 years old (Gore, 2002; Piccaluga et al., 2002; Robinson et al., 2002; Wraighte and Howard, 2008). This procedure, however, is technically challenging and time-intensive (Morgan et al., 2004), and a high incidence of complications have been reported (Eldridge et al., 1997a; Pekkarinen et al., 2000). For these reasons, some surgeons reserve this procedure for a specific group of patients for whom reconstitution of bone stock is deemed important: young (<50 years old) and/or active patients with compromised bone stock (Duncan et al., 1998; Leopold et al., 1999; Morgan et al., 2004). Some surgeons have expressed concerns about performing such a lengthy operation in elderly  7  patients after an 88-year-old woman died due to a cerebrovascular accident six months after surgery (de Thomasson et al., 2005). Nonetheless, femoral impaction allografting has become a popular option for revision THA at many hospitals, particularly in Europe. This procedure accounted for most of the revisions performed between 1991-2003 at the Royal Devon and Exeter Hospital in the United Kingdom (Sierra et al., 2008), and many clinical studies of femoral impaction allografting have been published recently by British centres (e.g., Wraighte and Howard, 2008; Sierra et al., 2008; Aulakh et al., 2009; Hassaballa et al., 2009). To this day, over 90 clinical articles about femoral impaction allografting can be found in English through Pubmed (Table 1.1).  Table 1.1 Published clinical articles on femoral impaction allografting. Year  UK  NL  1993  3  *  DN  FR  SP  BE  SW  *  *  NO  FN  US  CA  JP  SK  TL  IS  AR  AU  3  1994  0  1995  1  1996  1  1997  1  2 1  1 1  2  2  5 1  1  2 3  3 2  1999 2000  *  1  1998  2001  Total  1  3  5  1  1  1  7  4  9  1  2002  2  3  2  2003  1  1  3  2004  1  2  1  1  6 1  1  * *  1  13 5 5  2005  1  1  *  2  4  2006  3  1  2  2  8  2007  3  2008  3  1  2009  2  1  Total  23  9  1  1  1  1 1 2  6  1 3  1 1  14  1  1 1  1  25  7 6  5  1  1  1  1  2  90  Note: UK: United Kingdom, NL: the Netherlands, DN: Denmark, FR: France, SP: Spain, BE: Belgium, SW: Sweden, NO: Norway, FN: Finland, US: United States, CA: Canada, JP: Japan, SK: South Korea, TL: Thailand, IS: Israel, AR: Argentina, AU: Australia. Asterisk (*) denotes co-author participation from countries.  Nonetheless, there are considerable problems associated with femoral impaction allografting. Bone graft sources are limited, and two to four donor femoral heads are required to create the allograft bed. There is also a risk of disease transfer from the graft (Sommerville et al., 2000). Moreover, incomplete graft remodeling (Linder, 2000), a high incidence of femoral fracture (Meding et al., 1997; Leopold et al., 1999; Pekkarinen et al., 2000; Ornstein et al., 2002), and 8  high levels of implant migration relative to the femur have been reported (Meding et al., 1997; Eldridge et al., 1997a; Masterson et al., 1997a; Masterson et al., 1997b; Nelissen et al., 2002).  1.3.3 Femoral fracture Femoral fractures are common with impaction allografting and can occur intraoperatively (Gie et al., 1993; Meding et al., 1997; Masterson et al., 1997a; Pekkarinen et al., 2000; Knight and Helming, 2000; Ullmark et al., 2002; Robinson et al., 2002; Piccaluga et al., 2002; Ornstein et al., 2002) or postoperatively (Gie et al., 1993; Meding et al., 1997; Fetzer et al., 2001; Lind et al., 2002; Ornstein et al., 2002). Postoperative fractures occur typically around the stem tip or more proximally, and usually lead to re-operation, i.e., either fracture repair or another revision THA (Leopold et al., 1999). In one study, a postoperative fracture rate of 9% was reported within five months of the surgery; most of these were alleged to have originated at a cortical window or perforation that was present at the time of the surgery (Ornstein et al., 2002). Intraoperative fractures are especially common (Pekkarinen et al., 2000; Ornstein et al., 2002; Piccaluga et al., 2002; Frances et al., 2007) – these have been reported at rates up to 27% (Ornstein et al., 2002). Intraoperative fractures can occur during the removal of the cement from the primary THA, during graft impaction, or during trial reduction (Ornstein et al., 2002). Intraoperative fractures require subsequent cortical reinforcement using either cortical strut graft, synthetic mesh and/or cerclage wire (Masterson et al., 1997a; Leopold et al., 1999). Intraoperative perforation of the femoral shaft is also common, occurring typically during removal of the cement from the previous THA, and it is often treated with a cortical strut graft (Leopold et al., 1999). It has been found that greater bone stock deficiency and intraoperative perforation of the femur are risk factors for intra- and postoperative femoral fractures (Ornstein et al., 2002). To minimize the risk of fracture, several surgeons advocate prophylactic reinforcement of the femur with cables, meshes and/or plates, particularly in zones of extensive defect, prior to graft impaction (Duncan et al., 1998; Pekkarinen et al., 2000; Schreurs et al., 2005; Sierra et al., 2008).  9  1.3.4 Implant subsidence Gradual distal migration of the stem within the femur, often referred to as subsidence, is not uncommon after impaction allografting (Franzen et al., 1995; Meding et al., 1997; Masterson et al., 1997a; Eldridge et al., 1997a; Leopold et al., 1999; van Biezen et al., 2000; Pekkarinen et al., 2000; Ornstein et al., 2000; Boldt et al., 2001; Robinson et al., 2002). While it is often less than 2 mm (Halliday et al., 2003), much higher levels of subsidence are often observed (Meding et al., 1997; Masterson et al., 1997a; Eldridge et al., 1997a; Pekkarinen et al., 2000; Gore, 2002). Subsidence up to 8 cm has been reported (Sierra et al., 2008). Some implant subsidence within the cement mantle is expected to occur (Figure 1.6). The most common stems used in this procedure are double-tapered polished stems. Subsidence of these stems does not necessarily reflect a failure in fixation, as the implant may settle into a stable position (Gie et al., 1993; Elting et al., 1995; Nelissen et al., 2002; Ornstein et al., 2004). It is believed that the resulting radial compression generated as the stem subsides within the cement is beneficial in enhancing the rotational fixation of the implant (Gie et al., 1993). These tapered stems are often implanted with a centralizer, i.e., a polyethylene cap intended to leave a space into which the tip of the implant can migrate (Figure 1.6). It has been speculated that this process generates compression on the surrounding graft which may also be beneficial to the remodeling process (Gie et al., 1993). Nonetheless, a correlation between implant migration and the radiographic appearance of the graft has not been found (Ornstein et al., 2004). Furthermore, it is not clear whether the stem subsides mainly within the cement, or whether the whole construct subsides within the femur (van Biezen et al., 2000; Ornstein et al., 2001).  10  Figure 1.6 Settling of the stem within the cement mantle. A polyethylene centralizer (insert) is usually implanted at the tip of smooth tapered cemented stems to provide a space into which the stem tip can subside. Settling of the stem into the cement is expected to generate compressive radial stresses within the surrounding cement and graft regions. In some patients, however, stem migration is progressive (Ornstein et al., 2001; Nelissen et al., 2002), and can cause further complications (Meding et al., 1997; Masterson et al., 1997a; Eldridge et al., 1997a; Masterson et al., 1997b; Robinson et al., 2002; Piccaluga et al., 2002). ‘Massive’ subsidence (>10 mm) has been associated with a risk of aseptic loosening of the stem (Meding et al., 1997), hip dislocation (Masterson et al., 1997a) and thigh pain (Eldridge et al., 1997a; Robinson et al., 2002; Piccaluga et al., 2002; Hassaballa et al., 2009), and is therefore regarded often as a clinical failure (personal communication with Dr. Bassam Masri, 2009). The mechanisms through which subsidence develops after impaction allografting are not fully understood. Many mechanisms have been suggested to contribute to subsidence: settling of the stem within the cement (Gie et al., 1993; Elting et al., 1995; Karrholm et al., 1999; Knight and Helming, 2000; Ornstein et al., 2001; Schreurs et al., 2005; Yim et al., 2007; Wraighte and Howard, 2008), shear failure within the graft (Ornstein et al., 2001; Halliday et al., 2003), compression of the graft (Karrholm et al., 1999; Ornstein et al., 2000; Ornstein et al., 2001; Nelissen et al., 2002; Gore, 2002; Ornstein, 2002; Halliday et al., 2003), sliding at the endosteum interface (Ornstein, 2002; Frei et al., 2005a; Frei et al., 2005b), cement 11  fatigue/failure (Masterson et al., 1997a; Masterson et al., 1997a; Masterson et al., 1997b; Ullmark et al., 2002; Ornstein, 2002), allograft resorption (Karrholm et al., 1999; Ornstein, 2002), sliding at the graft/cement interface (Ornstein et al., 2001; Yim et al., 2007), and expansion of the femoral canal (Karrholm et al., 1999; Ornstein, 2002; Frei et al., 2005b). Longitudinal fracture of the cement mantle can cause the stem to subside within the cement (Figure 1.7). Some stems with massive subsidence showed radiological evidence of cement mantle fracture (Masterson et al., 1997a). It was suggested that cement mantle defects (defined as a cement layer less than 2 mm thick) may be a risk factor for massive subsidence, as this would predispose the construct to fracture of the cement (Masterson and Duncan, 1997; Masterson et al., 1997a; Masterson et al., 1997b). Since these reports were published, impaction tools have been updated to accommodate a minimum cement mantle thickness of 2 mm (personal communication with Dr. Clive Duncan, 2009). Without cement fracture, however, it is not clear whether cement mantle defects influences subsidence. There appeared to be a relationship between subsidence and cement mantle defects in one study (Nelissen et al., 2002) while none was observed in another study (Ornstein et al., 2004).  Figure 1.7 Stem subsidence relative to the cement due to longitudinal cement fracture.  Postoperative compression of the graft has been speculated to cause implant subsidence (Figure 1.8 a), (Karrholm et al., 1999; Ornstein et al., 2000; Ornstein et al., 2001; Nelissen et al., 2002; Ornstein, 2002; Gore, 2002; Halliday et al., 2003). This is presumed because the 12  graft region has the lowest stiffness of all the regions in a femoral impaction allografting construct and it exhibits some creep deformation under compressive loading (Giesen et al., 1999; Voor et al., 2000; Voor et al., 2004).  Figure 1. 8 Subsidence due to compression deformation of the graft bed (left), shear failure in the graft bed (center), and slippage at the endosteum interface (right).  Shear failure, or slippage in the graft bed, is also believed to contribute to implant subsidence (Ornstein et al., 2001; Halliday et al., 2003; Ornstein et al., 2004). Particulate materials are well known to be susceptible to shear stresses. Under certain loading conditions, particulate materials form shear failure planes that essentially slide until the normal loads on those planes become so high that the shear strength exceeds the shear stress. This process may contribute to implant subsidence in impaction allografting, as the graft region surrounding the stem is subjected to shear loading (Figure 1.8 b). It has also been suggested that subsidence may be associated with slippage at the endosteal interface (Figure 1.8 c), (Ornstein, 2002; Frei et al., 2005a; Frei et al., 2005b). The importance of this mechanism is implied by radiographic observations that the stem and cement appear to migrate together relative to the femur in some patients (Masterson et al., 1997a; Ullmark et al., 2002; Halliday et al., 2003; Yim et al., 2007).  13  Expansion of the femoral canal has been proposed as another possible contributor to subsidence (Figure 1.9), (Karrholm et al., 1999; Ornstein, 2002; Frei et al., 2005b). This mechanism could occur due either to the natural expansion of the femoral canal due to ageing (Ornstein, 2002; Ornstein et al., 2004), or to resorption or cancellization of the cortical bone at the endosteal (inner) surface caused by an interrupted vascular supply (Karrholm et al., 1999; Frei, 2003).  Figure 1.9 Subsidence caused by expansion of the femoral canal, adapted from Frei, 2003.  Finally, the graft incorporation process has been speculated to be the source of some subsidence (Karrholm et al., 1999; Ullmark et al., 2002; Ornstein, 2002; Ornstein et al., 2004). It is not known, however, whether MCB graft strengthens or weakens during the incorporation process (Ornstein et al., 2003). Some clinical factors have been proposed to affect the subsidence. It is widely believed that a thorough packing of the graft is essential for initial stem stability (Halliday et al., 2003; Cabanela et al., 2003). Graft particle size distribution is also believed important (Halliday et al., 2003). In vitro studies have characterized the mechanical behaviour of MCB in shear (Tanabe et al., 1999; Brewster et al., 1999; Dunlop et al., 2003). It was concluded that shear strength was affected by the particle size distribution (Tanabe et al., 1999; Dunlop et al., 2003), and by washing of the graft (Dunlop et al., 2003). Nonetheless, aggregates for which the grading was closest to the theoretically ideal particle size distribution did not have superior 14  shear strength compared with a theoretically inferior mix (Dunlop et al., 2003). Others, on the other hand, have reported that larger (>4.5-5 mm) particles yielded better stability in in vitro tests compared with smaller particles (Wallace et al., 1997; Eldridge et al., 1997b); however, no quantitative information was provided about the grading of the particle batches being compared. Finally, whether grading or washing of the graft truly affect implant stability is not certain, nor whether shear failures truly occur postoperatively within the graft bed. In some studies, there appeared to be a relationship between the severity of the bone stock deficiency and subsidence for the double-tapered Exeter stem (van Doorn et al., 2002; Nelissen et al., 2002; Halliday et al., 2003; Hassaballa et al., 2009). Other studies, however, did not observe such a relationship for the same stem (Meding et al., 1997; Ornstein, 2002; Ornstein et al., 2004; Gokhale et al., 2005), nor for collared stems (Leopold et al., 1999; Leopold et al., 1999; Boldt et al., 2001; Boldt et al., 2001; Ornstein, 2002; van Doorn et al., 2002; Ornstein et al., 2004; Gokhale et al., 2005; Gokhale et al., 2005). Similarly, it is unclear whether patient age at the time of revision plays a role in the development of subsidence. Some studies have reported higher angular stem displacement (Gokhale et al., 2005) and a higher incidence of massive distal subsidence (Hassabala et al., 2009) in older patients. Nonetheless, age was not found to affect subsidence in other studies (Eldridge et al., 1997a; van Doorn et al., 2002; Frances et al., 2007). In short, the relative impact of each proposed mechanism on the development of subsidence after impaction allografting remains unknown (Halliday et al., 2003; Gokhale et al., 2005), and the wide range of subsidence reported in the clinical literature is not fully understood.  1.3.5 Cement penetration into the graft bed Several clinical studies, retrieval analyses, and in vitro studies that focus on the mechanical characterization of MCB are available. However, it was not until recently that the morphology of the cement/graft layer was described. In a recent in vitro study, extensive cement penetration into the graft bed was observed when impaction allografting was performed in a cadaveric femur model (Frei et al., 2004). It was found that cement penetration was deeper than anticipated, reaching the endosteal surface of the femur in the diaphysis region, especially  15  around mid-stem (Figure 1.10). A finite element study showed that cement penetration depth was related to cement viscosity (e.g., how soon after mixing the cement was injected), how much pressure was applied, and for how long the pressure was maintained (Frei et al., 2006). It was also shown that the apparent strength of the endosteal interface was proportional to the amount of cement contact at that interface (Frei et al., 2005a). Host cortical bone Implant Pure cement mantle Graft/cement composite Pure graft  a.  b.  Figure 1.10 Extensive cement penetration observed after impaction allografting in a cadaveric femur model. (a) Cross-section taken from the metaphysis region, i.e., proximal femur. (b) Cross section from the diaphysis region, around mid-stem. Adapted from Frei et al., 2004.  1.4 Summary and Project Motivation In summary, various mechanisms have been proposed to contribute to subsidence, but the relative importance of each mechanism on the development of subsidence is not known. The quality of the graft impaction, i.e., the graft density achieved, is believed critical to the initial stability of the implant. Nonetheless, it has been demonstrated that high levels of cement penetration into the graft bed can occur after impaction allografting, and the roles of cement profile and graft packing density on implant stability have not yet been determined.  16  1.5 Objectives and Scope The focus of this thesis was to investigate the effects of impaction force and cement pressurization on morphological and mechanical characteristics of the femur-implant construct after impaction allografting. The objectives of this study were to: 1. Examine the effect of cement penetration into the graft bed on the initial stability of the femoral stem after impaction allografting (CHAPTER 2). 2. Determine the effect of impaction force on the bone density, shear strength, compression stiffness, and creep behavior of morsellized cancellous bone graft (CHAPTER 3). 3. Explore two alternative methods to graft impaction as a means to improve the mechanical characteristics of the graft (CHAPTER 3). 4. Estimate the effect of graft density, cement penetration profile, and the status of the stem/cement and endosteal interfaces, i.e., bonded or sliding, on the shear stresses at the endosteal interface and the distal migration and micromotion of the stem (CHAPTER 4). 5. Estimate the relative contribution of each of the following mechanisms on the stem motion: sliding at the stem-cement interface; deformation of the graft and/or graft/cement composite region; and slippage at the endosteal interface (CHAPTER 4). In Chapter 2, the effect of cement pressurization on cement penetration into the graft bed after impaction allografting was investigated in an in vitro human cadaveric femur model. The graft density, the cement penetration profiles, and the initial implant migration and micromotion were measured, and the relationships between these morphological and structural characteristics were studied. In Chapter 3, the effects of impaction force on density, compression stiffness and shear strength of the graft bed were examined using confined compression tests, simple shear tests, and histomorphometric analysis. Alternative graft compaction methods were proposed and these were compared with traditional impaction. Finally, in Chapter 4, a finite element model was used to estimate the relative importance of graft bed density, cement penetration profile, and debonding of the cement at the stem-cement and endosteal interfaces on the structural characteristics of the proximal femur after impaction allografting.  17  1.6 References American Academy of Orthopaedic Surgeons. www.orthoinfo.aaos.org accessed on March 23, 2009.  (2007)  Total  Hip  Replacement.  American Academy of Orthopaedic Surgeons. (2008) The Burden of Musculoskeletal Diseases in the United States. www.aaos.org accessed on March 3, 2009. Aulakh, T.S., Jayasekera, N., Kuiper, J.H., and Richardson, J.B. Long-term clinical outcomes following the use of synthetic hydroxyapatite and bone graft in impaction in revision tip arthroplasty. Biomaterials 30, 1732-8. Boldt, J.G., Dilawari, P., Agarwal, S., and Drabu, K.J. (2001) Revision total hip arthroplasty using impaction bone grafting with cemented nonpolished stems and charnley cups. J Arthroplasty 16, 943-52. Bozic, K.J., Kurtz, S.M., Lau, E., Ong, K., Vail, T.P., and Berry, D.J. (2009) The epidemiology of revision total hip arthroplasty in the United States. J Bone Joint Surg Am 91, 128-33. Brewster, N.T., Gillespie, W.J., Howie, C.R., Madabhushi, S.P., Usmani, A.S., and Fairbairn, D.R. (1999) Mechanical considerations in impaction bone grafting. J Bone Joint Surg Br 81, 118-24. Cabanela, M.E., Trousdale, R.T., and Berry, D.J. (2003) Impacted cancellous graft plus cement in hip revision. Clin Orthop Relat Res 417, 175-82. Canadian Institute for Health Information. (2008) Canadian Joint Replacement Registry (CJRR) 2007 Annual Report — Hip and Knee Replacements in Canada. Ottawa: CIHI. Canadian Institutes for Health Information. (2005) Canadian Joint Replacement Registry (CJRR). www.cihi.ca accessed on March 3, 2009. Note: original source: Hospital Morbidity Database, CIHI, 1994-1995 and 2004-2005. de Thomasson, E., Williams, J.B., Marmorat, J.L., Guidand, O., and Mazel, C. (2005) Modified Exeter technique in revision hip surgery: does distal fixation of the stem affect allograft transformation? J Arthroplasty 20, 473-80. Duncan, C.P., Masterson, E.L., and Masri, B.A. (1998) Impaction allografting with cement for the management of femoral bone loss. Orthop Clin North Am 29, 297-305. Dunlop, D.G., Brewster, N.T., Madabhushi, S.P., Usmani, A.S., Pankaj, P., and Howie, C.R. (2003) Techniques to improve the shear strength of impacted bone graft: the effect of particle size and washing of the graft. J Bone Joint Surg Am 85-A, 639-46. Eldridge, J.D., Smith, E.J., Hubble, M.J., Whitehouse, S.L., and Learmonth, I.D. (1997a) Massive early subsidence following femoral impaction grafting. J Arthroplasty 12, 535-40. Eldridge, J.D.J., Hubble, K., Nelson, K., Smith, E.J., and Learmonth, I.D. (1997b) The effect of bone chip size on initial stability following femoral impaction grafting. J Bone Joint Surg Br 18  79-B , 364. Elting, J.J., Mikhail, W.E., Zicat, B.A., Hubbell, J.C., Lane, L.E., and House, B. (1995) Preliminary report of impaction grafting for exchange femoral arthroplasty. Clin Orthop Relat Res 319, 159-67. Engelbrecht E. and Heinert K. (1987) Klassifikation und Behandlungsrichtlienen von Knochensubstanzverlusten bei Revisionsoperationen am Huftgelenk - mittelfristige Ergebnisse. In: Primär- und Revisions-Alloarthroplastik, Hüft- und Kniegelenk: 10 Jarhe Endo-Klinik Hamburg. Springer Verlag Berlin, 198-201. Fetzer, G.B., Callaghan, J.J., Templeton, J.E., Goetz, D.D., Sullivan, P.M., and Johnston, R.C. (2001) Impaction allografting with cement for extensive femoral bone loss in revision hip surgery: a 4- to 8-year follow-up study. J Arthroplasty 16, 195-202. Frances, A., Moro, E., Cebrian, J.L., Marco, F., Garcia-Lopez, A., Serfaty, D., and LopezDuran, L. (2007) Reconstruction of bone defects with impacted allograft in femoral stem revision surgery. Int Orthop 31, 457-64. Franzen, H., Toksvig-Larsen, S., Lidgren, L., and Onnerfalt, R. (1995) Early migration of femoral components revised with impacted cancellous allografts and cement. A preliminary report of five patients. J Bone Joint Surg Br 77, 862-4. Frei, H. (2003) Morphology and biomechanical characteristics of the proximal femur after impaction allografting. PhD Thesis. The University of British Columbia. Frei, H., Gadala, M.S., Masri, B.A., Duncan, C.P., and Oxland, T.R. (2006) Cement flow during impaction allografting: a finite element analysis. J Biomech 39, 493-502. Frei, H., Mitchell, P., Masri, B.A., Duncan, C.P., and Oxland, T.R. (2004) Allograft impaction and cement penetration after revision hip replacement. A histomorphometric analysis in the cadaver femur. J Bone Joint Surg Br 86, 771-6. Frei, H., Mitchell, P., Masri, B.A., Duncan, C.P., and Oxland, T.R. (2005a) Mechanical characteristics of the bone-graft-cement interface after impaction allografting. J Orthop Res 23, 9-17. Frei, H., O'Connell, J., Masri, B.A., Duncan, C.P., and Oxland, T.R. (2005b) Biological and mechanical changes of the bone graft-cement interface after impaction allografting. J Orthop Res 23, 1271-9. Gie, G.A., Linder, L., Ling, R.S., Simon, J.P., Slooff, T.J., and Timperley, A.J. (1993) Impacted cancellous allografts and cement for revision total hip arthroplasty. J Bone Joint Surg Br 75, 14-21. Giesen, E.B., Lamerigts, N.M., Verdonschot, N., Buma, P., Schreurs, B.W., and Huiskes, R. (1999) Mechanical characteristics of impacted morsellised bone grafts used in revision of total hip arthroplasty. J Bone Joint Surg Br 81, 1052-7.  19  Gokhale, S., Soliman, A., Dantas, J.P., Richardson, J.B., Cook, F., Kuiper, J.H., and Jones, P. (2005) Variables affecting initial stability of impaction grafting for hip revision. Clin Orthop Relat Res 432, 174-80. Gore, D.R. (2002) Impaction bone grafting for total hip revision. Int Orthop 26, 162-5. Halliday, B.R., English, H.W., Timperley, A.J., Gie, G.A., and Ling, R.S. (2003) Femoral impaction grafting with cement in revision total hip replacement. Evolution of the technique and results. J Bone Joint Surg Br 85, 809-17. Hassaballa, M., Mehendale, S., Poniatowski, S., Kalantzis, G., Smith, E., and Learmonth, I.D. (2009) Subsidence of the stem after impaction bone grafting for revision hip replacement using irradiated bone. J Bone Joint Surg Br 91, 37-43. Health Canada. (1998) Economic burden of illness in Canada. Health Canada (2003) Arthritis in Canada – an ongoing challenge. www.phac-aspc.gc.ca. Accessed on May 15, 2009. Karrholm, J., Garellick, G., and Herberts, P. (2007) Swedish hip arthroplasty register – annual report 2006. Department of Orthopaedics, Sahlgrenska University Hospital. Karrholm, J., Hultmark, P., Carlsson, L., and Malchau, H. (1999) Subsidence of a non-polished stem in revisions of the hip using impaction allograft. Evaluation with radiostereometry and dual-energy X-ray absorptiometry. J Bone Joint Surg Br 81, 135-42. Knight, J.L. and Helming, C. (2000) Collarless polished tapered impaction grafting of the femur during revision total hip arthroplasty: pitfalls of the surgical technique and follow-up in 31 cases. J Arthroplasty 15, 159-65. Kurtz, S., Mowat, F., Ong, K., Chan, N., Lau, E., and Halpern, M. (2005) Prevalence of primary and revision total hip and knee arthroplasty in the United States from 1990 through 2002. J Bone Joint Surg Am 87, 1487-97. Leopold, S.S., Berger, R.A., Rosenberg, A.G., Jacobs, J.J., Quigley, L.R., and Galante, J.O. (1999) Impaction allografting with cement for revision of the femoral component. A minimum four-year follow-up study with use of a precoated femoral stem. J Bone Joint Surg Am 81, 1080-92. Lind, M., Krarup, N., Mikkelsen, S., and Horlyck, E. (2002) Exchange impaction allografting for femoral revision hip arthroplasty: results in 87 cases after 3.6 years' follow-up. J Arthroplasty 17, 158-64. Linder, L. (2000) Cancellous impaction grafting in the human femur: histological and radiographic observations in 6 autopsy femurs and 8 biopsies. Acta Orthop Scand 71, 543-52. Masterson, E.L. and Duncan, C.P. (1997) Subsidence and the cement mantle in femoral impaction allografting. Orthopedics 20, 821-2.  20  Masterson, E.L., Masri, B.A., and Duncan, C.P. (1997a) The cement mantle in the Exeter impaction allografting technique. A cause for concern. J Arthroplasty 12, 759-64. Masterson, E.L., Masri, B.A., Duncan, C.P., Rosenberg, A., Cabanela, M., and Gross, M. (1997b) The cement mantle in femoral impaction allografting. A comparison of three systems from four centres. J Bone Joint Surg Br 79, 908-13. Meding, J.B., Ritter, M.A., Keating, E.M., and Faris, P.M. (1997) Impaction bone-grafting before insertion of a femoral stem with cement in revision total hip arthroplasty. A minimum two-year follow-up study. J Bone Joint Surg Am 79, 1834-41. Morgan, H.D., McCallister, W., Cho, M.S., Casnellie, M.T., and Leopold, S.S. (2004) Impaction allografting for femoral component revision: clinical update. Clin Orthop Relat Res 420, 160-8. Nelissen, R.G., Bauer, T.W., Weidenhielm, L.R., LeGolvan, D.P., and Mikhail, W.E. (1995) Revision hip arthroplasty with the use of cement and impaction grafting. Histological analysis of four cases. J Bone Joint Surg Am 77, 412-22. Nelissen, R.G., Valstar, E.R., Poll, R.G., Garling, E.H., and Brand, R. (2002) Factors associated with excessive migration in bone impaction hip revision surgery: a radiostereometric analysis study. J Arthroplasty 17, 826-33. Ornstein, E. (2002) Hip revisions with impacted morselized allograft bone and cement. Acta Orthop Scand Suppl 73, 1-69. Ornstein, E., Atroshi, I., Franzen, H., Johnsson, R., Sandquist, P., and Sundberg, M. (2001) Results of hip revision using the Exeter stem, impacted allograft bone, and cement. Clin Orthop Relat Res 398, 126-33. Ornstein, E., Atroshi, I., Franzen, H., Johnsson, R., Sandquist, P., and Sundberg, M. (2002) Early complications after one hundred and forty-four consecutive hip revisions with impacted morselized allograft bone and cement. J Bone Joint Surg Am 84-A, 1323-8. Ornstein, E., Franzen, H., Johnsson, R., Karlsson, M.K., Linder, L., and Sundberg, M. (2004) Hip revision using the Exeter stem, impacted morselized allograft bone and cement: a consecutive 5-year radiostereometric and radiographic study in 15 hips. Acta Orthop Scand 75, 533-43. Ornstein, E., Franzen, H., Johnsson, R., Stefansdottir, A., Sundberg, M., and Tagil, M. (2003) Hip revision with impacted morselized allografts: unrestricted weight-bearing and restricted weight-bearing have similar effect on migration. A radiostereometry analysis. Arch Orthop Trauma Surg 123, 261-7. Ornstein, E., Franzen, H., Johnsson, R., and Sundberg, M. (2000) Radiostereometric analysis in hip revision surgery – optimal time for index examination: 6 patients revised with impacted allografts and cement followed weekly for 6 weeks. Acta Orthop Scand 71, 360-4. Pekkarinen, J., Alho, A., Lepisto, J., Ylikoski, M., Ylinen, P., and Paavilainen, T. (2000) 21  Impaction bone grafting in revision hip surgery. A high incidence of complications. J Bone Joint Surg Br 82, 103-7. Piccaluga, F., Gonzalez Della Valle, A., Encinas Fernandez, J.C., and Pusso, R. (2002) Revision of the femoral prosthesis with impaction allografting and a Charnley stem. A 2- to 12-year follow-up. J Bone Joint Surg Br 84, 544-9. Robinson, D.E., Lee, M.B., Smith, E.J., and Learmonth, I.D. (2002) Femoral impaction grafting in revision hip arthroplasty with irradiated bone. J Arthroplasty 17, 834-40. Schreurs, B.W., Arts, J.J., Verdonschot, N., Buma, P., Slooff, T.J., and Gardeniers, J.W. (2005) Femoral component revision with use of impaction bone-grafting and a cemented polished stem. J Bone Joint Surg Am 87, 2499-507. Sierra, R.J., Charity, J., Tsiridis, E., Timperley, J.A., and Gie, G.A. (2008) The use of long cemented stems for femoral impaction grafting in revision total hip arthroplasty. J Bone Joint Surg Am 90, 1330-6. Sommerville, S.M., Johnson, N., Bryce, S.L., Journeaux, S.F., and Morgan, D.A. (2000) Contamination of banked femoral head allograft: incidence, bacteriology and donor follow up. Aust N Z J Surg 70, 480-4. Statistics Canada. (2008) 2005 Canadian Community Health Survey. www.statcan.gc.ca accessed on March 23, 2009. Tanabe, Y., Wakui, T., Kobayashi, A., Ohashi, H., Kadoya, Y., and Yamano, Y. (1999) Determination of mechanical properties of impacted human morsellized cancellous allografts for revision joint arthroplasty. J Mater Sci Mater Med 10, 755-60. Ullmark, G., Hallin, G., and Nilsson, O. (2002) Impacted corticocancellous allografts and cement for revision of the femur component in total hip arthroplasty. J Arthroplasty 17, 140-9. van Biezen, F.C., ten Have, B.L., and Verhaar, J.A. (2000) Impaction bone-grafting of severely defective femora in revision total hip surgery: 21 hips followed for 41-85 months. Acta Orthop Scand 71, 135-42. van Doorn, W.J., ten Have, B.L., van Biezen, F.C., Hop, W.C., Ginai, A.Z., and Verhaar, J.A. (2002) Migration of the femoral stem after impaction bone grafting. First results of an ongoing, randomised study of the exeter and elite plus femoral stems using radiostereometric analysis. J Bone Joint Surg Br 84, 825-31. Voor, M.J., Nawab, A., Malkani, A.L., and Ullrich, C.R. (2000) Mechanical properties of compacted morselized cancellous bone graft using one-dimensional consolidation testing. J Biomech 33, 1683-8. Voor, M.J., White, J.E., Grieshaber, J.E., Malkani, A.L., and Ullrich, C.R. (2004) Impacted morselized cancellous bone: mechanical effects of defatting and augmentation with fine hydroxyapatite particles. J Biomech 37, 1233-9.  22  Wallace, I.W., Ammon, P.R., Day, R., Lee, D.A., and Beaver, R.J. (1997) Does size matter?: an investigation into the effects of particle size on impaction grafting in vitro. J Bone Joint Surg Br 79-B, 366. Woolf, A.D. and Pfleger, B. (2003) Bulletin of the World Health Organization. Bull World Health Organ 81. Wraighte, P.J. and Howard, P.W. (2008) Femoral impaction bone allografting with an Exeter cemented collarless, polished, tapered stem in revision hip replacement: a mean follow-up of 10.5 years. J Bone Joint Surg Br 90, 1000-4. Yim, S.J., Kim, M.Y., and Suh, Y.S. (2007) Impaction allograft with cement for the revision of the femoral component. A minimum 39-month follow-up study with the use of the Exeter stem in Asian hips. Int Orthop 31, 297-302.  23  CHAPTER 2  CEMENT PENETRATION AND PRIMARY STABILITY OF THE FEMORAL PROSTHESIS A BIOMECHANICAL STUDY IN THE CADAVERIC FEMUR  A version of this chapter has been published. Albert C, Patil S, Frei H, Masri B, Duncan C, Oxland T, Fernlund G (2007). Cement penetration and primary stability of the femoral prosthesis after impaction allografting: A biomechanical study in the cadaveric femur. Journal of Bone and Joint Surgery British Edition. 89 (7): 962-70. 24  2.1 Introduction Failure of the femoral component in total hip replacement is often associated with a loss of bone stock in the proximal femur which can make revision procedures challenging. Impaction allografting has gained popularity over the last decade because of its potential for reconstitution of bone stock, which is particularly important when dealing with young patients with moderate to severe bone loss (Duncan et al., 1998). Impaction allografting, however, is associated with several clinical problems, including a high prevalence of intra- and postoperative fractures (Knight and Helming, 2000; Fetzer et al., 2001; Ornstein et al., 2002; Schreurs et al., 2005) and high levels of implant migration (> 10 mm) (Masterson and Duncan, 1997; Eldridge et al., 1997; van Biezen et al., 2000; Piccaluga et al., 2002). Although the mechanisms by which migration develops are not fully understood, inadequate compaction of the graft bed (Karrholm et al., 1999; Nelissen et al., 2002; Malkani et al., 1996; Gokhale et al., 2005), defects in the cement mantle (Nelissen et al., 2002; Masterson et al., 1997) and absorption of the endosteal surface (Frei et al., 2005b) are believed to be contributing factors. Many studies have shown radiological evidence of remodelling of the graft following impaction allografting (Gie et al., 1993; Meding et al., 1997; Cabanela et al., 2003). However, histological reports from autopsies and biopsies have revealed that the graft bed had not fully remodeled into viable bone, even up to eight years after the impaction allografting procedure (Linder, 2000; Nelissen et al., 1995). It has been shown in a cadaveric experiment that during femoral impaction allografting the penetration of cement into the graft bed is greater than expected, filling virtually the entire intramedullary canal at the level of mid-stem (Frei et al., 2004). Penetration of cement to the endosteal cortex is a limiting factor that prevents bone remodeling, as cement is not biodegradable (Frei et al., 2005b). Therefore, a construct having reduced cement penetration, particularly in the proximity of the endosteal surface, may enhance local vascularization and allow the formation of new bone in these cement-free areas. Results from a finite element analysis have suggested that lower levels of cement penetration could be achieved by reducing  25  cement pressure or increasing its viscosity (Frei et al., 2006). A potential drawback of reducing cement penetration in impaction allografting is that the shear strength at the endosteal interface is lower in the absence of cement contact (Frei et al., 2005a), which may lead to excessive migration of the stem. However, the importance of contact of the cement with the endosteal surface on the primary stability of the stem is not known. This study aimed to examine the effect of the extent of cement penetration into the graft bed on the primary stability of the femoral stem after impaction allografting. We hypothesized that by not pressurizing the cement a reduction in penetration would be achieved versus the conventional pressurized cementing technique. We also hypothesized that there is a relationship between cement contact with the endosteum and the primary stability of the stem. We tested our hypotheses in the human cadaveric femur.  2.2 Methods Impaction allografting was performed on eight pairs of human cadaveric femurs. The femoral heads were removed from the specimens and loss of bone was simulated with a high-speed burr. All the trabecular bone was removed from the proximal femur, and lytic defects were created in the cortical shell. The loss of bone achieved represented a class II defect according to the EndoKlinik classification system (Engelbrecht E. and Heinert K., 1987). Trabecular bone graft from 15 femoral heads and 15 femoral condyles was morsellized with a Lere bone mill (DePuy, Warsaw, Indiana), giving a particle size distribution of between 0.6 mm and 13 mm, with 50% by weight of the particles being smaller than 4 mm (Frei et al., 2004). The grafts were rinsed in saline and pooled together for better reproducibility. All the procedures were performed by a single surgeon (SP). Prophylactic application of Dall-Miles cables (Stryker Orthopaedics, Mahwah, New Jersey) was carried out to prevent iatrogenic fractures. For the impaction procedures the CPT instrumentation (Zimmer Inc., Warsaw, Indiana), with Simplex P bone cement (Stryker Howmedica Osteonics, Allendale, New Jersey) was used. In one femur from each pair, chosen randomly, the cement was pressurized, whereas in the other femur it was not. In both groups, the cement was injected in the neo-medullary canal in a retrograde fashion using a cement gun. In the pressure group, two packs of bone cement were used and the timing of pressurization was standardized. The cement was mixed for one minute,  26  injection began at two minutes from the onset of mixing, and pressurization at three minutes, ending at four minutes. Finally, ten seconds after the end of pressurization, an appropriate femoral component (CPT 12/14 Hip System, Zimmer Inc.) was inserted. In the no-pressure group, one pack of cement was used per specimen, and the same cement gun was used. The neo-medullary canal was filled with cement in a retrograde fashion, but the cement was not pressurized. Insertion of the stem began three minutes from the onset of cement mixing. Three specimens from two pairs fractured during the procedure, and therefore six pairs of specimens were included in the study (Table 2.1). Table 2.1 Description of specimens. Specimen pair Gender 1 Male 2 Female 3 Male 4 Male 5 Female 6 Male *EXT, extended offset  Age 85 84 Unknown 56 69 52  Implant size used 1 EXT* 1 EXT 2 EXT 1 EXT 1 EXT 2 EXT  The femurs were potted in dental stone (Tru-Stone, Heraeus Kulzer, Armonk, New York) at 13° of adduction and subjected to cyclical loading on a biaxial servohydraulic testing machine (Instron Model 8874, Instron, Canton, Massachusetts). The loads applied simulated 50% of expected walking loads (2500 cycles) followed by 100% of walking loads (5000 cycles). Two force components were applied sinusoidally at 1 Hz: a craniocaudal component (Fcc) with peak values of 0.4 to 2.3 times body-weight, and an anteroposterior component (Fap) with peak values of -0.1 to 0.3 times body-weight for the walking load cycles (Bergmann 2001 gait patterns). The loads were scaled for a 70 kg individual, and the two components were phased such that their maximum peak values coincided. The Fcc was applied at 13° from the longitudinal axis of the femur, resulting in a mediolateral component (Fml) and a proximodistal component (Fpd) as shown in Figure 2.1. The anteroposterior load (Fap) was applied with the rotary actuator by controlling the moment, M, with an offset of 32 mm between the femoral head and the line of action of the actuator. More details about the simulation of hip joint loading are provided in Appendix 1.1.  27  Fcc Biaxial actuator Load cell  M Fml Offset (32 mm)  Fap  Cap Implant motion reference point  Fcc 13°  Fpd  Linear guide Instron table  Figure 2.1 Diagram showing the loading set-up. The potted femur was mounted on a linear guide to avoid undesired horizontal reaction forces in the frontal plane. A craniocaudal load (Fcc) was applied with the linear actuator, generating the mediolateral (Fml) and proximodistal (Fpd) components. The anteroposterior load (Fap) was applied by controlling the moment, M, applied with the rotary actuator (M = Fap × offset).  The three-dimensional motion of the implant relative to the bone was measured at the reference point shown in Figure 2.1, using a custom-built system similar to designs used previously (Berzins et al., 1993; Chareancholvanich et al., 2002). The system, shown in Figure 2.2, consisted of six linear variable differential transformers (GCD-121-250, Shaevitz Sensors, Hampton, Virginia) mounted on an aluminum frame which was fixed rigidly to the femur with seven pyramid-tip set screws. In order to achieve adequate fixation of the screws to the femur, all local soft tissue was removed and the periosteal surface was sanded, cleaned with acetone, and sealed with cyanoacrylate. The set-screw-femoral interface was strengthened using polymethyl-methacrylate (Lecoset, LECO Corp, St Joseph, Michigan). The sensors measured the motion of a triangle that was rigidly fixed to the lateral side of the implant, 5 cm below its shoulder, with a 5 mm square steel pin through a hole in the femur. The relative motion between the implant and the bone was calculated from the measured motion of the triangle relative to the frame, using a custom program implemented in Matlab (MathWorks, Natick, Massachusetts). The measured motion was separated into migration (permanent) and micromotion (reversible), each of which had three translational components (posterior, lateral 28  and distal), and three rotational components, (valgus/varus, flexion/extension, and retroversion/anteversion), see Appendix 1, Figure A1.9. Migration was defined as the difference between the initial position of the stem and the average unloaded position during the last ten cycles of loading, and micromotion was defined as the amplitude of reversible motion during the last ten cycles. The accuracy of the system in measuring translation was evaluated against a dial gauge micrometer by attaching the sensors to an over-reamed composite femur (Model 3303, Pacific Research Laboratories, Vashon, Washington), and applying translations to the implant relative to the bone along each of the three axes. The mean absolute error observed was 2.1 µm (0.4 to 5.8) in 12 translation measurements over a range of 200 µm. The accuracy of each sensor was also measured. A maximum error of 2.2 µm was observed over a range of 450 µm (mean 0.8 µm (0.0 to 2.2) for 60 measurements). The accuracy of rotation was evaluated analytically from the maximum individual sensor errors, yielding a maximum rotation error of 0.004°. More details about the implant motion measurement methods can be found in Appendix 1.2.  Figure 2.2 Photograph of the motion measurement system. Six linear variable differential transformers were mounted on an aluminum frame that was rigidly attached to the femur. They measured the motion of a triangle that was fixed rigidly to the lateral side of the implant through a hole in the femur. The relative motion between implant and bone was determined from that of the triangle, relative to the frame.  29  Non-parametric statistical analysis was used to compare the results between the two groups because the variances differed substantially between the groups for some of the motion components. Motion was compared between the pressure group and the no-pressure group using Wilcoxon’s matched-pairs tests, and a significance level of 0.05 was used. For the components that showed a significant difference between the pressure and no-pressure groups, linear regression models were used to examine the relationship between migration and micromotion. After structural testing, the implant was removed and the cavity filled with coloured polymethyl-methacrylate (LECO Corp). The femurs were cut into transverse cross-sections, 7 mm thick, with a diamond saw (model 310 CP, Exakt Apparatebau, Norderstedt, Germany). The levels of cross-sections were matched between the specimens using the tip of the stem as the reference. Alternating cross-sections, corresponding to 11 levels, were processed for histomorphometric analysis. They were fixed in 10% buffered formalin for a minimum of two days, and dehydrated for four days: one day in 70% ethanol, one in 95% ethanol, and two in 100% ethanol (changing the solution after the first day). The cross-sections were embedded in resin (Buehler EpoThin, Lake Bluff, Illinois) to form blocks, which were mounted on a slide. Each block was cut to approximately 500 µm with a diamond saw and ground to 200 µm, and the slide was stained with a calcium stain, alizarin red S. Each slide was placed on a light box to enhance the contrast between graft and cement/voids, and photographed with a digital camera (resolution 3.2 megapixels). The graft porosity was characterized as the percentage of the graft area occupied by cement or void, using the Image-Pro 4.5 (Mediacybernetics, Silver Springs, Maryland) image analysis software. The slides were then placed on a green background, to distinguish more clearly the area of penetration of the cement within the graft bed. Each slide was photographed and the penetration of cement was measured using the Image-Pro 4.5 software. Penetration was characterized as the percentage of contact of the cement with the endosteal surface (cement contact, Equation 2.1) and the area of the canal occupied by cement (cement area, Equation 2.2):  Cement contact =  length of contact between cement and endosteum × 100% endosteum perimeter  (2.1)  30  Cement area =  area inside the cement penetration front × 100% total canal area  (2.2)  Further details about the slide preparation, staining and photography methods are provided in Appendix 2. The measurements of cement contact and cement area were performed by a single person (CA) and were each analyzed with a two-factor analysis of variance (ANOVA), in which the factors were group (pressure versus no-pressure), and level (defined as a repeated measure). The graft porosity was analyzed with a one-factor ANOVA (level) which was defined as a repeated measure. Student-Newman-Keuls analysis was used for post hoc comparisons, and the significance level was 0.05. Linear regression models were used to examine the relationship between the mean cement contact and cement area from all 11 matched levels, and the components of the motion that were significantly different between the two groups.  2.3 Results The cement contact and cement area results are shown in Figure 2.3 and Table 2.2 for each group at all cross-sectional levels. The main effect of pressure was significant for both cement contact and cement area (p = 0.0013 and p = 0.0014, respectively). There was an average of 30% more cement contact and 25% more cement area in the pressure group than in the nopressure group. The effect of level was also significant (p < 0.001 for both variables). Levels 4 to 10 had significantly higher cement contact and cement area than levels 1, 2, 3 and 11 (p < 0.05), but they did not differ significantly between levels 4 and 10, or between levels 1, 2, 3 and 11. There was no interaction effect between the two factors (i.e., pressure group and level). The results for graft porosity are shown in Table 2.2. The effect of level on graft porosity was significant (p < 0.001). Porosity was relatively constant along the length of the implant, averaging 66%, and did not differ significantly between levels 1 and 10 (post hoc comparisons with Student-Newman-Keuls p > 0.4). Below the tip of the implant, at level 11, the porosity  31  was slightly lower than at all other levels (post hoc comparisons with Student-Newman-Keuls p < 0.01).  Cement contact (%)  Pressure group  No-pressure group  100 90 80 70 60 50 40 30 20 10 0 0  1  2  3  4  5  6  7  8  9  10  11  12  Level  Figure 2.3 Graph showing cement contact with endosteum for the pressure and nopressure groups, at eleven matched-level cross-sections. Means (standard deviations).  32  Table 2.2 Cement contact with endosteum, cement area and graft porosity for the pressure and no-pressure group at eleven matched-level cross-sections.  Level 1 2 3 4 5 6 7 8 9 10 11  Cement contact (%) Pressure group No-pressure group Mean (range) Mean (range) 26 (0 to 56) 9 (0 to 32) 32 (25 to 44) 7 (0 to 20) 42 (28 to 73) 20 (4 to 29) 59 (32 to 86) 41 (10 to 82) 71 (49 to 98) 46 (28 to 91) 70 (49 to 97) 36 (0 to 52) 72 (48 to 100) 40 (0 to 60) 82 (67 to 100) 40 (0 to 100) 80 (45 to 100) 36 (0 to 100) 81 (27 to 97) 25 (0 to 65) 26 (0 to 91) 11 (0 to 38)  Cement area (%) Pressure group Mean (range) 52 (32 to 83) 53 (35 to 74) 57 (37 to 82) 73 (43 to 96) 83 (50 to 100) 80 (52 to 99) 80 (59 to 99) 87 (73 to 100) 89 (75 to 99) 91 (70 to 98) 50 (4 to 94)  No-pressure group Mean (range) 33 (23 to 46) 26 (20 to 29) 40 (24 to 69) 57 (35 to 97) 59 (38 to 97) 55 ( 38 to 76) 60 (43 to 76) 57 (32 to 100) 57 (30 to 100) 45 (16 to 73) 34 (0 to 57)  Graft porosity All specimens Mean (range) 68 (64 to 75) 66 (60 to 72) 68 (60 to 72) 66 (60 to 75) 65 (52 to 74) 64 (52 to 78) 66 (54 to 82) 69 (61 to 79) 66 (52 to 81) 65 (54 to 78) 57 (40 to 71)  The components of migration and micromotion for each group at the end of the cyclical loading are shown in Tables 2.3 and 2.4, respectively. Two of the migration components, distal translation and valgus rotation, differed significantly (p = 0.028) between the two groups (Figure 2.4). The median distal migration and valgus rotation migration were twelve and three times greater in the no-pressure group (Table 2.3). The resultant of the translational migration components ranged between 19 µm and 27 µm in the pressure group, and between 45 µm and 792 µm in the no-pressure group. Of all the micromotion components, only distal translation was significantly different between the groups (p = 0.028, Figure 2.4c). The median distal micromotion was almost five times greater in the no-pressure group (Table 2.4). The resultant of the translational micromotion components ranged between 10 µm and 31 µm in the pressure group, and between 22 µm and 70 µm in the no-pressure group. Distal migration also correlated significantly with the micromotion in the same direction (slope of the linear fit to the data (m) = 12.8, y-offset (b) = - 101 µm, R2 = 0.78, linear regression, p < 0.001). Table 2.3 Migration components for the pressure and no-pressure groups. Pressure group No-pressure group Median (Range ) Median (Range ) 8 (-15 to 22) 41 (-89 to 203) Posterior (µm) 2 (-9 to 20) 7 (-34 to 38) Lateral (µm) 18 (-6 to 26) 218 (42 to 765) Distal (µm) 0.04 (0.02 to 0.11) 0.13 (0.11 to 0.20) Valgus/varus (°) 0.02 (-0.08 to 0.05) 0.01 (-0.02 to 0.08) Flexion/extension (°) 0.01 (-0.01 to 0.05) 0.02 (-0.32 to 0.20) Retro/anteversion (°) Valgus, flexion and retroversion rotations are represented as positive rotations.  p-value 0.60 0.92 0.028 0.028 0.46 0.92  33  Table 2.4 Micromotion components for the pressure and no-pressure groups.  900 800 700 600 500 400 300 200 100 0 -100  Valgus/varus migration ( )  0.25 0.20 0.15 0.10 0.05 0.00  Pressure  Nopressure  p-value 0.75 0.75 0.028 0.17 0.60 0.46  70  °  Distal micromotion ( µm)  Distal migration ( µm)  Pressure group No-pressure group Median (Range ) Median (Range ) 9 (4 to 21) 8 (5 to 26) Posterior (µm) 7 (2 to 17) 8 (3 to 36) Lateral (µm) 7 (4 to 15) 32 (15 to 65) Distal (µm) 0.05 (0.02 to 0.08) 0.07 (0.04 to 0.13) Valgus/varus (°) 0.02 (0.01 to 0.04) 0.03 (0.01 to 0.04) Flexion/extension (°) 0.03 (0.01 to 0.08) 0.06 (0.02 to 0.08) Retro/anteversion (°) Valgus, flexion and retroversion rotations are represented as positive rotations.  60 50 40 30 20 10 0  Pressure  Nopressure  Pressure  Nopressure  Figure 2.4 Graphs showing distal migration (left), valgus/varus rotation migration (center) and distal micromotion (right) for the two groups (medians and ranges). Valgus rotation is represented as a positive rotation. All of these motions (distal migration, valgus/varus migration and distal micromotion) displayed a significant difference between the two groups (p<0.05).  Distal migration and micromotion throughout the 5000 cycles of walking loads are shown in Figure 2.5. In the pressure group, distal micromotion was mainly constant (Figure 2.5 a). The trend of micromotion in the no-pressure group was more variable, being roughly constant in three specimens, increasing slightly in one, and decreasing in two. The implant settled more quickly in the pressure group, where most of the migration occurred within the first 2000 cycles (Figure 2.5 b). In the no-pressure group, the implants were still migrating distally at 5000 cycles, with a median migration of 10.9 µm (0.3 to 27.1) during the last 1000 cycles, compared with 0.5 µm (-1.1 to 3.4) in the pressure group.  34  Pressure group  No-pressure group  Distal micromotion (µm)  90 80 70 60 50 40 30 20 10 0 0  1000  2000  3000  4000  5000  Cycle  Pressure group  No-pressure group  900  Distal migration (µm)  800 700 600 500 400 300 200 100 0 -100 0  1000  2000  3000  4000  5000  Cycle Figure 2.5 Graph showing distal micromotion (top) and migration (bottom) versus number of walking cycles for all specimens tested. Note that the migration data presented include the migration that had previously developed during the 2500 cycles at 50% of expected walking loads.  35  The three motion components that differed between the pressure and the no-pressure groups (distal migration, valgus rotation migration and distal micromotion) correlated significantly with both cement contact and cement area (Table 2.5 and Figure 2.6). Two specimens, both of which were in the no-pressure group, showed substantially higher distal migration than the others, and corresponded to the two largest specimens in that group (Pairs 3 and 6, Table 2.1). The average cement contact with the endosteum (averaged over the eleven levels) was correlated positively with the average cement area (slope = 1.07, R = 0.89).  Table 2.5 Pearson’s correlation coefficients (R) and slopes for the linear regressions between the motion components and the cement morphological parameters (average cement contact and average cement area). Parameter Distal migration (µm)  Factor Cement contact Cement area  Slope -9.557 -14.090  R 0.68 0.83  Valgus rotation migration (°)  Cement contact Cement area  -0.0021 -0.0028  0.65 0.75  Distal micromotion (µm)  Cement contact -0.574 Cement area -0.896 All slopes were significantly different from zero (p < 0.05).  0.59 0.77  36  900  Migration (µm)  800 700 600 500 400 300 200 100 0 0  20  40  60  80  Mean cement contact (%)  900  Migration (µm)  800 700 600 500 400 300 200 100 0 20  40  60  80  100  Mean cement area (%) Figure 2.6 Scattergraph showing distal migration as a function of: (top) the average cement contact with endosteum (averaged for each specimen over all 11 levels); (bottom) the average cement area. The regression parameters are given in Table 2.5.  37  2.4 Discussion When using the impaction allografting technique penetration of cement through the graft bed has been observed to reach the endosteal surface of the femur, creating an extensive cementhost bone interface at mid-stem level (Frei et al., 2004). The presence of cement at the endosteal surface compromises the potential for incorporation of the graft at that site (Frei et al., 2005b), but its importance for primary stability of the implant is not known. In this study, we carried out impaction allografting on cadaveric femurs with and without pressurizing the cement to examine the effect of cement penetration on the motion of the implant under simulated walking loads. In vitro mechanical tests are commonly performed to assess new hip implant designs or surgical techniques pre-clinically. The relevance of in vitro tests is supported by studies demonstrating that excessive micromotion at the bone-implant interface inhibits successful bone ingrowth in cementless implants, which may lead to early loosening of the implant (Pilliar et al., 1986; Engh et al., 1992; Jasty et al., 1997), and that cemented implants with inferior clinical results also display greater in vitro micromotion (Cristofolini et al., 2003). As an in vitro model our study did not model biological processes. There was no bleeding in the canal during the procedure, which may have affected cement penetration. However, it has been shown experimentally that the intramedullary blood pressure did not significantly affect the depth of penetration in primary cemented THA (Majkowski et al., 1994). It has been suggested that early migration of the implant could be a result of a combination of postoperative consolidation of the graft (Giesen et al., 1999; Voor et al., 2000; Voor et al., 2004), shear failure (Brewster et al., 1999; Dunlop et al., 2003) and/or slippage at the hostbone interface (Frei et al., 2005a). Other mechanisms that have been proposed to affect longerterm implant migration in impaction allografting, including graft incorporation/fibrous invasion (Linder, 2000; Nelissen et al., 1995), cement fatigue (Masterson et al., 1997), and cancellization of the femoral canal (Frei et al., 2005b), could not be modeled in this experiment. The effect of these mechanisms on long-term subsidence of the implant in impaction allografting has not yet been determined. The importance of primary stability is emphasized by reports of significant subsidence in the early postoperative days. In clinical  38  studies using roentgen stereophotogrammetric analysis (RSA) some implants were seen to have subsided more than 1 mm within the first three months after operation (Nelissen et al., 2002; Ornstein et al., 2001). It was also noted that more than half of the subsidence observed at six weeks was usually seen within the first two weeks (Ornstein et al., 2000). Our specimens represented class II bone defects. Most of the trabecular bone had been removed from the proximal femur, and the intramedullary canal was expanded. This type of bone loss is within the range commonly targeted by impaction allografting, but does not represent the most severe cases. As the migration of the implant may be proportional to the severity of the bone loss (Nelissen et al., 2002; Gokhale et al., 2005), we might see greater migration with a model of more severe bone loss. Migration of implants previously seen in cadaveric studies of impaction allografting has varied greatly, with average distal migrations ranging between 12 µm (Chassin et al., 1997) and 1 mm (Bolder et al., 2004). There were many variables in these studies including the type of specimen, the magnitude and orientation of the loading, the number of loading cycles, and the techniques used to measure motion (Chassin et al., 1997; Bolder et al., 2004; Berzins et al., 1996; Hostner et al., 2001; Kligman et al., 2003; van Haaren et al., 2005). Comparison of these results with those of our pressure group shows that two other studies used a six degree of freedom motion measurement system similar to ours (Chassin et al., 1997; Berzins et al., 1996) and found similar results in all motion components. Berzins et al. used loads approximately twice as high as ours but only 50 load cycles, whereas Chassin et al. used only ten cycles, with approximately half our loads. Two other studies reported distal migration of an order of magnitude higher than ours, but one used RSA measurement with loads approximately twice as high as ours and 13 times more cycles (Hostner et al., 2001), and the other an extensometer measurement with similar loads but almost 200 times as many cycles (Cornu et al., 2003). Other studies have used the position of the actuator as their measure of migration, and have consistently reported greater migration by at least an order of magnitude (Malkani et al., 1996; Kligman et al., 2003; van Haaren et al., 2005). However, the position of the actuator does not provide an accurate measure of actual implant motion relative to the bone.  39  Our results indicate that following impaction allografting there is a substantial amount of cement penetration in the graft, and cement contact with the endosteum, in particular around the distal half of the implant. Eliminating cement pressurization during impaction allografting reduced the cement area and cement contact with the endosteal surface, but migration and micromotion of the implant were significantly increased. Distal migration, valgus/varus migration, and distal micromotion were significantly higher without pressure than with, and correlated with the average amount of cement contact with the endosteum and with the average cement area. It can be argued that some initial settling of the implant may be acceptable, provided it then becomes stable. Because our study applied a limited number of cycles which were roughly equivalent to one or two days of loading, the initial settling process was not observed entirely. It is possible that the micromotion would ultimately be similar between the two groups because of further compaction of the graft during implant migration. During the cycles applied, however, the micromotion of the implant in the no-pressure group did not decrease consistently with the increasing number of cycles applied, and did not tend to converge towards the same magnitude as those in the pressure group. Furthermore, the results of a clinical impaction allografting study indicate that migration during the first two weeks may be a good predictor of long-term migration (Nelissen et al., 2002). Therefore, the differences in the patterns of early migration observed between our pressure and no-pressure groups indicate that longer term stability of the implant could be compromised if the cement is not pressurized. Previous research has demonstrated that the cement penetration in impaction allografting is affected by graft porosity (Frei et al., 2006). In our results, the lower porosity of the graft below the tip of the implant may explain the lower contact between the cement and the endosteum in that region due to a lower permeability of the graft in this region. However, the porosity of the graft was roughly uniform along the length of the implant, and the lower cement-endosteum contact in the proximal region is probably a result of the wider canal at that site. Our measurements of graft porosity at all levels were within the range previously reported in another in vitro study (Frei et al., 2004). It is not clear how much cement is necessary for the structural support of the implant in impaction allografting. Without cement, the morsellized bone does not provide sufficient 40  structural support for the stem (Robinson et al., 2005). Some clinical studies have reported a link between excessive subsidence and the presence of zones where the cement mantle was less than 2 mm thick (Nelissen et al., 2002; Masterson et al., 1997). This indicates the importance of sufficient thickness of the cement mantle between the implant and the graft bed, without which the cement could be more susceptible to early fatigue failure (Masterson and Duncan, 1997). A recent in vitro study found that penetration of cement into the graft bed reaches the endosteal surface, and that the strength of the endosteal interface with graft/cement is proportional to the amount of cement contact (Frei et al., 2004; Frei et al., 2005a). In the present study, constructs with more than 50% of cement contact with the endosteum generally resulted in substantially lower distal migration and micromotion than did those with less contact. This indicates that the regions of cement endosteum contact are potentially critical to the primary stability of the implant. There may be a conflict between the biological and structural goals of impaction allografting. A biologically favourable construct is one with limited cement penetration, but the morsellized graft bed may not provide sufficient support for the implant without contact of the cement with the endosteum. Attempts have been made to improve the mechanical properties of the morsellized graft bed by increasing the number of impactions (Bavadekar et al., 2001), rinsing the graft (Dunlop et al., 2003; Cornu et al., 2004), freeze-drying (Cornu et al., 2003) and/or irradiating the graft (Cornu et al., 2004; Butler et al., 2005), and by optimizing the size distribution of the graft particles (Dunlop et al., 2003). However, improvement of graft compaction will only be of clinical consequence if the implant is being supported largely by the graft bed. Future biomechanical studies should aim to determine the minimum graft packing necessary to support the implant without contact of the cement with the endosteum, how to achieve this level of packing consistently, and how to prevent the cement from reaching the endosteum. However, if the graft cannot provide sufficient support for the implant, efforts should be focused on how to control the cement penetration profile such that cement-endosteum contact occurs in regions not targeted for reconstitution of bone stock. The presence of cement at the endosteal surface in patients with impaction allografting has not yet been confirmed because published histological reports of biopsy and autopsy specimens used dehydrating solutions and embedding compounds that dissolve the bone cement (Linder, 2000; Ullmark and Linder, 1998; Ullmark and Obrant, 2002). The histological methods used in the present study did not dissolve the 41  cement. We found that the cement penetration at each level was most easily distinguished after each cross-section was mounted on a slide and ground to roughly 200 µm. Such biomechanical and retrieval studies could help determine whether cement contact with the endosteum is essential in preventing excessive migration in impaction allografting. Finally, surgeons should be aware that the potential for graft remodeling may be limited around the distal half of the implant, and that contact of the cement with the endosteum may be critically important for the initial stability of the implant in impaction allografting.  42  2.5 References Bavadekar, A., Cornu, O., Godts, B., Delloye, C., Van Tomme, J., and Banse, X. (2001) Stiffness and compactness of morselized grafts during impaction: an in vitro study with human femoral heads. Acta Orthop Scand 72, 470-6. Berzins, A., Sumner, D.R., Andriacchi, T.P., and Galante, J.O. (1993) Stem curvature and load angle influence the initial relative bone-implant motion of cementless femoral stems. J Orthop Res 11, 758-69. Berzins, A., Sumner, D.R., Wasielewski, R.C., and Galante, J.O. (1996) Impacted particulate allograft for femoral revision total hip arthroplasty. In vitro mechanical stability and effects of cement pressurization. J Arthroplasty 11, 500-6. Bolder, S.B., Schreurs, B.W., Verdonschot, N., Ling, R.S., and Slooff, T.J. (2004) The initial stability of an exeter femoral stem after impaction bone grafting in combination with segmental defect reconstruction. J Arthroplasty 19, 598-604. Brewster, N.T., Gillespie, W.J., Howie, C.R., Madabhushi, S.P., Usmani, A.S., and Fairbairn, D.R. (1999) Mechanical considerations in impaction bone grafting. J Bone Joint Surg Br 81, 118-24. Butler, A.M., Morgan, D.A., Verheul, R., and Walsh, W.R. (2005) Mechanical properties of gamma irradiated morselized bone during compaction. Biomaterials 26, 6009-13. Cabanela, M.E., Trousdale, R.T., and Berry, D.J. (2003) Impacted cancellous graft plus cement in hip revision. Clin Orthop Relat Res 417, 175-82. Chareancholvanich, K., Bourgeault, C.A., Schmidt, A.H., Gustilo, R.B., and Lew, W.D. (2002) In vitro stability of cemented and cementless femoral stems with compaction. Clin Orthop Relat Res 394, 290-302. Chassin, E.P., Silverton, C.D., Berzins, A., and Rosenberg, A.G. (1997) Implant stability in revision total hip arthroplasty: allograft bone packing following extended proximal femoral osteotomy. J Arthroplasty 12, 863-8. Cornu, O., Bavadekar, A., Godts, B., Van Tomme, J., Delloye, C., and Banse, X. (2003) Impaction bone grafting with freeze-dried irradiated bone. Part I. Femoral implant stability: cadaver experiments in a hip simulator. Acta Orthop Scand 74, 547-52. Cornu, O., Libouton, X., Naets, B., Godts, B., Van Tomme, J., Delloye, C., and Banse, X. (2004) Freeze-dried irradiated bone brittleness improves compactness in an impaction bone grafting model. Acta Orthop Scand 75, 309-14. Cristofolini, L., Teutonico, A.S., Monti, L., Cappello, A., and Toni, A. (2003) Comparative in vitro study on the long term performance of cemented hip stems: validation of a protocol to discriminate between "good" and "bad" designs. J Biomech 36, 1603-15. Duncan, C.P., Masterson, E.L., and Masri, B.A. (1998) Impaction allografting with cement for 43  the management of femoral bone loss. Orthop Clin North Am 29, 297-305. Dunlop, D.G., Brewster, N.T., Madabhushi, S.P., Usmani, A.S., Pankaj, P., and Howie, C.R. (2003) Techniques to improve the shear strength of impacted bone graft: the effect of particle size and washing of the graft. J Bone Joint Surg Am 85-A, 639-46. Eldridge, J.D., Smith, E.J., Hubble, M.J., Whitehouse, S.L., and Learmonth, I.D. (1997) Massive early subsidence following femoral impaction grafting. J Arthroplasty 12, 535-40. Engelbrecht E. and Heinert K. Klassifikation und Behandlungsrichtlienen von Knochensubstanzverlusten bei Revisionsoperationen am Huftgelenk - mittelfristige Ergebnisse. In: Primär- und Revisions-Alloarthroplastik, Hüft- und Kniegelenk: 10 Jarhe Endo-Klinik Hamburg. Springer Verlag Berlin, 198-201. 1987. Engh, C.A., O'Connor, D., Jasty, M., McGovern, T.F., Bobyn, J.D., and Harris, W.H. (1992) Quantification of implant micromotion, strain shielding, and bone resorption with porouscoated anatomic medullary locking femoral prostheses. Clin Orthop Relat Res 285, 13-29. Fetzer, G.B., Callaghan, J.J., Templeton, J.E., Goetz, D.D., Sullivan, P.M., and Johnston, R.C. (2001) Impaction allografting with cement for extensive femoral bone loss in revision hip surgery: a 4- to 8-year follow-up study. J Arthroplasty 16, 195-202. Frei, H., Gadala, M.S., Masri, B.A., Duncan, C.P., and Oxland, T.R. (2006) Cement flow during impaction allografting: a finite element analysis. J Biomech 39, 493-502. Frei, H., Mitchell, P., Masri, B.A., Duncan, C.P., and Oxland, T.R. (2004) Allograft impaction and cement penetration after revision hip replacement. A histomorphometric analysis in the cadaver femur. J Bone Joint Surg Br 86, 771-6. Frei, H., Mitchell, P., Masri, B.A., Duncan, C.P., and Oxland, T.R. (2005a) Mechanical characteristics of the bone-graft-cement interface after impaction allografting. J Orthop Res 23, 9-17. Frei, H., O'Connell, J., Masri, B.A., Duncan, C.P., and Oxland, T.R. (2005b) Biological and mechanical changes of the bone graft-cement interface after impaction allografting. J Orthop Res 23, 1271-9. Gie, G.A., Linder, L., Ling, R.S., Simon, J.P., Slooff, T.J., and Timperley, A.J. (1993) Impacted cancellous allografts and cement for revision total hip arthroplasty. J Bone Joint Surg Br 75, 14-21. Giesen, E.B., Lamerigts, N.M., Verdonschot, N., Buma, P., Schreurs, B.W., and Huiskes, R. (1999) Mechanical characteristics of impacted morsellised bone grafts used in revision of total hip arthroplasty. J Bone Joint Surg Br 81, 1052-7. Gokhale, S., Soliman, A., Dantas, J.P., Richardson, J.B., Cook, F., Kuiper, J.H., and Jones, P. (2005) Variables affecting initial stability of impaction grafting for hip revision. Clin Orthop Relat Res 432, 174-80.  44  Hostner, J., Hultmark, P., Karrholm, J., Malchau, H., and Tveit, M. (2001) Impaction technique and graft treatment in revisions of the femoral component: laboratory studies and clinical validation. J Arthroplasty 16, 76-82. Jasty, M., Bragdon, C., Burke, D., O'Connor, D., Lowenstein, J., and Harris, W.H. (1997) In vivo skeletal responses to porous-surfaced implants subjected to small induced motions. J Bone Joint Surg Am 79, 707-14. Karrholm, J., Hultmark, P., Carlsson, L., and Malchau, H. (1999) Subsidence of a non-polished stem in revisions of the hip using impaction allograft. Evaluation with radiostereometry and dual-energy X-ray absorptiometry. J Bone Joint Surg Br 81, 135-42. Kligman, M., Rotem, A., and Roffman, M. (2003) Cancellous and cortical morselized allograft in revision total hip replacement: A biomechanical study of implant stability. J Biomech 36, 797-802. Knight, J.L. and Helming, C. (2000) Collarless polished tapered impaction grafting of the femur during revision total hip arthroplasty: pitfalls of the surgical technique and follow-up in 31 cases. J Arthroplasty 15, 159-65. Linder, L. (2000) Cancellous impaction grafting in the human femur: histological and radiographic observations in 6 autopsy femurs and 8 biopsies. Acta Orthop Scand 71, 543-52. Majkowski, R.S., Bannister, G.C., and Miles, A.W. (1994) The effect of bleeding on the cement-bone interface. An experimental study. Clin Orthop Relat Res 299, 293-7. Malkani, A.L., Voor, M.J., Fee, K.A., and Bates, C.S. (1996) Femoral component revision using impacted morsellised cancellous graft. A biomechanical study of implant stability. J Bone Joint Surg Br 78, 973-8. Masterson, E.L. and Duncan, C.P. (1997) Subsidence and the cement mantle in femoral impaction allografting. Orthopedics 20, 821-2. Masterson, E.L., Masri, B.A., and Duncan, C.P. (1997) The cement mantle in the Exeter impaction allografting technique. A cause for concern. J Arthroplasty 12, 759-64. Meding, J.B., Ritter, M.A., Keating, E.M., and Faris, P.M. (1997) Impaction bone-grafting before insertion of a femoral stem with cement in revision total hip arthroplasty. A minimum two-year follow-up study. J Bone Joint Surg Am 79, 1834-41. Nelissen, R.G., Bauer, T.W., Weidenhielm, L.R., LeGolvan, D.P., and Mikhail, W.E. (1995) Revision hip arthroplasty with the use of cement and impaction grafting. Histological analysis of four cases. J Bone Joint Surg Am 77, 412-22. Nelissen, R.G., Valstar, E.R., Poll, R.G., Garling, E.H., and Brand, R. (2002) Factors associated with excessive migration in bone impaction hip revision surgery: a radiostereometric analysis study. J Arthroplasty 17, 826-33. Ornstein, E., Atroshi, I., Franzen, H., Johnsson, R., Sandquist, P., and Sundberg, M. (2001) 45  Results of hip revision using the Exeter stem, impacted allograft bone, and cement. Clin Orthop Relat Res 389, 126-33. Ornstein, E., Atroshi, I., Franzen, H., Johnsson, R., Sandquist, P., and Sundberg, M. (2002) Early complications after one hundred and forty-four consecutive hip revisions with impacted morselized allograft bone and cement. J Bone Joint Surg Am 84-A, 1323-8. Ornstein, E., Franzen, H., Johnsson, R., and Sundberg, M. (2000) Radiostereometric analysis in hip revision surgery--optimal time for index examination: 6 patients revised with impacted allografts and cement followed weekly for 6 weeks. Acta Orthop Scand 71, 360-4. Piccaluga, F., Gonzalez Della Valle, A., Encinas Fernandez, J.C., and Pusso, R. (2002) Revision of the femoral prosthesis with impaction allografting and a Charnley stem. A 2- to 12-year follow-up. J Bone Joint Surg Br 84, 544-9. Pilliar, R.M., Lee, J.M., and Maniatopoulos, C. (1986) Observations on the effect of movement on bone ingrowth into porous-surfaced implants. Clin Orthop Relat Res 208, 108-13. Robinson, M.C., Fernlund, G., Dominic Meek, R.M., Masri, B.A., Duncan, C.P., and Oxland, T.R. (2005) Structural characteristics of impaction allografting for revision total hip arthroplasty. Clin Biomech (Bristol, Avon) 20, 853-5. Schreurs, B.W., Arts, J.J., Verdonschot, N., Buma, P., Slooff, T.J., and Gardeniers, J.W. (2005) Femoral component revision with use of impaction bone-grafting and a cemented polished stem. J Bone Joint Surg Am 87, 2499-507. Ullmark, G. and Linder, L. (1998) Histology of the femur after cancellous impaction grafting using a Charnley prosthesis. Arch Orthop Trauma Surg 117, 170-2. Ullmark, G. and Obrant, K.J. (2002) Histology of impacted bone-graft incorporation. J Arthroplasty 17, 150-7. van Biezen, F.C., ten Have, B.L., and Verhaar, J.A. (2000) Impaction bone-grafting of severely defective femora in revision total hip surgery: 21 hips followed for 41-85 months. Acta Orthop Scand 71, 135-42. van Haaren, E.H., Smit, T.H., Phipps, K., Wuisman, P.I., Blunn, G., and Heyligers, I.C. (2005) Tricalcium-phosphate and hydroxyapatite bone-graft extender for use in impaction grafting revision surgery. An in vitro study on human femora. J Bone Joint Surg Br 87, 267-71. Voor, M.J., Nawab, A., Malkani, A.L., and Ullrich, C.R. (2000) Mechanical properties of compacted morselized cancellous bone graft using one-dimensional consolidation testing. J Biomech 33, 1683-8. Voor, M.J., White, J.E., Grieshaber, J.E., Malkani, A.L., and Ullrich, C.R. (2004) Impacted morselized cancellous bone: mechanical effects of defatting and augmentation with fine hydroxyapatite particles. J Biomech 37, 1233-9.  46  CHAPTER 3 THE EFFECT OF IMPACTION FORCE AND ALTERNATIVE COMPACTION METHODS ON THE MECHANICAL CHARACTERISTICS OF MORSELLIZED CANCELLOUS GRAFT  A version of this chapter has been published. Albert C, Masri B, Duncan C, Oxland T, Fernlund G (2008). Impaction allografting – The effect of impaction force and alternative compaction methods on the mechanical characteristics of the graft. Journal of Biomedical Materials Research Part B: Applied Biomaterials. 87 (2): 395-405. 47  3.1 Introduction Failure of total hip implants is often accompanied with problematic bone loss. The goal of femoral impaction allografting is to improve bone stock in patients with femoral deficiencies using impacted morsellized cancellous bone graft (MCB). Clinical problems often reported with impaction allografting, however, include intra- and postoperative fractures (Knight and Helming, 2000; Fetzer et al., 2001; Ornstein et al., 2002; Schreurs et al., 2005) and high levels of implant migration (Masterson et al., 1997; Eldridge et al., 1997; van Biezen et al., 2000; Piccaluga et al., 2002). The mechanisms through which this migration occurs are not yet fully understood, but the MCB region might be critical because its stiffness is lower than that of the other materials in the reconstruction. Compressive stiffness between 3 and 135 MPa has been reported for impacted MCB (Giesen et al., 1999; Voor et al., 2000; Verdonschot et al., 2001; Bavadekar et al., 2001; Cornu et al., 2003a; Cornu et al., 2004; Voor et al., 2004; Fosse et al., 2004; Phillips et al., 2006b; Fosse et al., 2006) compared with 2 GPa, 17 GPa and 220 GPa for the cement, cortical bone, and stem, respectively (Mow and Huiskes, 2005). The role of the graft in the migration process is supported by the results of a clinical study, in which an apparent relationship was reported between migration and the density of the graft bed (Gokhale et al., 2005). MCB is a biphasic material composed of bone particles and a fluid phase (marrow, fat and water). In addition to having a relatively low stiffness, impacted MCB exhibits viscoelasticviscoplastic, i.e., time dependent, material behavior (Giesen et al., 1999; Phillips et al., 2006a). Under compressive stress, the load is carried initially in part by the fluid, and the fluid pressure dissipates over a short period (a few seconds) through fluid flow (Voor et al., 2000). Beyond this point, MCB continues to deform (Voor et al., 2000), indicating that the bone particles themselves behave in a viscoelastic fashion. This process of creep consolidation is believed to contribute to implant migration in impaction allografting (Giesen et al., 1999; Voor et al., 2000). Particulate materials are susceptible to shear failures, which may also contribute to migration (Brewster et al., 1999; Ornstein et al., 2004). The shear strength (τf) of MCB has been  48  described using the Mohr-Coulomb criterion: τf = c + σn tanφ, in which σn is the normal (perpendicular) stress applied on the shear failure plane, and c and φ (the cohesion intercept and the friction angle, respectively) are constants for a given particulate material (Brewster et al., 1999; Dunlop et al., 2003; Tanabe et al., 1999). In soil mechanics, the shear strength of many particulate materials can be optimized by using a ‘well graded’ mix of particles, i.e., a wide range of particle sizes that allow maximum packing density (Craig, 1983). This, however, has not been shown to be consistently true for MCB (Dunlop et al., 2003). We suggest that, because MCB particles are highly deformable, the density of the graft bed may not depend only on the grading, but perhaps more importantly on the impaction force. Brewster et al. reported that the shear strength of dry MCB increased with increasing impaction energy, but the forces and energy levels used were not indicated (Brewster et al., 1999). We hypothesize that the shear strength of fresh frozen MCB increases with increasing impaction force, and that this strength increase is related to an increase in density of the graft bed. Furthermore, because of the damping effect of the fluid phase (Voor et al., 2000) and the short duration (~5 ms) of an impaction impulse (Fosse et al., 2006), it can be presumed that only a portion of the impaction force is exerted on the graft particles, while the remainder is carried by fluid pressure. We therefore hypothesize that a slower compaction technique that allows sufficient time for fluid exudation and graft deformation can produce a denser, stiffer and stronger graft bed than are obtained with impaction by allowing a larger portion of the compaction force to be transferred to the bone particles. Our first objective was to examine the relationship between the impaction force and the shear strength, compression stiffness, and creep behaviour of MCB over a range of impaction forces typical in impaction allografting. Our second objective was to examine the relationship between the above-mentioned mechanical characteristics and the density of impacted MCB. Finally, our third objective was to explore two alternative methods to graft impaction as a means to improve the mechanical properties of the graft bed compared with impaction at a given force level. We addressed our objectives with in vitro tests on human MCB.  49  3.2 Methods Thirty-nine human femoral heads were morsellized with a Lere bone mill (DePuy Orthopaedics, Inc.). Cortical bone from the femoral neck was removed prior to morsellization. The cartilage was also removed, as it has been demonstrated to decrease the stiffness of MCB (Bavadekar et al., 2001). The graft particle sizes obtained with this method ranged between 0.6 and 8 mm, with 50% of the particles (in weight) being finer than 2.4 mm. The particle size distribution, obtained using sieve analysis, is shown in Figure 3.1. The graft was pooled together to reduce variability and rinsed in saline, as this has been shown to improve its shear  Percent smaller (%)  strength (Dunlop et al., 2003).  100 90 80 70 60 50 40 30 20 10 0 0.1  1  10  Particle size (mm) Figure 3.1 Particle size distribution for the morsellized graft. The ordinate indicates the percentage by weight of particles smaller than the abscissa, e.g., 50% of the particles are smaller than 2.4 mm.  Graft compaction was done with an 18 mm diameter steel piston using a servohydraulic testing machine (Instron Model 8874, Instron, Canton, Massachussetts). The graft was compacted in a stainless steel cylinder, having a wall thickness of 6 mm, an inner diameter of 19 mm, and with 36 radial holes of 1 mm diameter to allow fluid exudation. The initial, unimpacted height of each specimen was approximately 35 mm (after light finger-packing). Compaction was done with one of the following compaction techniques: impaction, creep, and cyclic relaxation (Figure 3.2 and Table 3.1). The creep technique consisted of holding a constant compaction 50  force of 300 N for 90 s, and the cyclic relaxation consisted of ten cycles of compressing the graft to 300 N, then holding the displacement constant for 10 s each. The impaction technique simulated 20 impactions of a hammer with force levels of 300 N, 600 N, and 900 N, which were chosen to represent low, moderate, and high impaction forces (Fosse et al., 2004). Graft impaction was simulated on a servohydraulic materials testing machine, for which the load input was chosen to match as closely as possible the desired force peaks and short duration of an impaction impulse, i.e., approximately 0.005s (Fosse et al., 2004). A square load input with a frequency of 4 Hz – the highest frequency achievable by the testing machine – was used, which resulted in load impulses of 0.125s. The square load input generated a load profile as shown in Appendix 3 (Figure A3.3). A pilot experiment showed that this method of graft impaction produced equivalent graft stiffness to that achieved when using a hammer for a given peak force and number of impactions, while allowing controllable and repeatable forces between specimens. Stiffness development was monitored during impaction, Eimp, as the stress amplitude divided by the natural strain amplitude during the loading phase of each impaction.  Creep  (20 impactions)  (90s)  Time  Cyclic Relaxation (10 cycles)  Force  Force  Force  Impaction  Time  Time  Figure 3.2 Compaction techniques. Impaction (left): Peak compressive forces of 300, 600 and 900 N (1, 2, and 3 MPa) were used. Creep (center): a compressive force of 300 N was held for 90 s. Cyclic relaxation (right): a force of 300 N was applied ten times, holding displacement constant for 9 seconds after each loading. Table 3.1 Test matrix. Graft compaction techniques and number of specimens used for compression, shear, and density tests. Graft compaction technique Impaction 300 N (‘I300N’) 600 N (‘I600N’) 900 N (‘I900N’) Creep (‘C300N’) Cyclic Relaxation (‘R300N’)  Compression tests N=6 N=6 N=6 N=6 N=6  Normal stress, σn = 125 kPa N=6 N=6 N=6 N=6 -  Shear tests Normal stress, σn = 250 kPa N=6 N=6 N=6 N=6 -  Normal stress, σn = 375 kPa N=6 N=6 N=6 N=6 -  Density tests N=6 N=6 N=6 -  51  Following compaction, each specimen underwent either: compression testing, shear testing, or density measurement (Table 3.1 and Figure 3.3). Shear and compression tests were performed on the same testing machine as that used for compaction, which enabled us to regulate the time between compaction and testing. After compaction, the piston was lifted from the specimen surface and the graft was allowed to recoil for 60 s. This was done to simulate the recoil occurring during surgery in the time between the removal of the final impactor and the injection of the cement into the graft. One minute was chosen arbitrarily for the recoil, but most of the recoil is known to occur during the first 10 s (Ullmark and Nilsson, 1999). New graft was used for each specimen to ensure that the properties measured were not affected by previous testing. Impaction  Recoil  Creep test  σo (1.1 MPa) piston  (σo)  F (N)  Measure ε (t)  graft Time (not to scale)  Impaction  Recoil  σn (125, 250, or 375 kPa)  Shear test  (σn) F (N)  δ δ (mm)  τ  Measure τ (δ)  Time (not to scale)  Figure 3.3 Compression tests (top) consisted of one hour creep tests. Shear tests (bottom) were performed at three normal stress levels, σn.  Compression tests (Figure 3.3, top) consisted of one hour creep tests, for which the set-up is shown in Appendix 3. After recoil, a seating load of 10 N (0.035 MPa) was applied and held for ten minutes – the position of the piston at this time was recorded and used to obtain the initial specimen height. A load of 320 N (σo=1.1 MPa) was then applied in a 1 s ramp and held for one hour. The initial compression stiffness, Ei, was defined as the slope of the stress-natural strain curve during the initial loading ramp, and the compression stiffness at one hour, E1hr, was defined as the normal stress applied, σo, divided by the natural strain after one hour of creep, i.e., the inverse of creep compliance (Figure 3.4). Similarly, the initial strain, εi, was defined as the natural strain at the end of the initial loading ramp, and the creep constant, Cε, 52  was defined as the slope of the strain versus log time curve during the one hour creep period (Voor et al., 2000), Figure 3.5. A nonparametric approach was taken for the statistical analyses of these data because the stiffness and strain variances were not homogeneous between all the groups (Levene’s tests). Spearman rank R correlations were used to determine the relationships between Ei, E1hr, εi, and Cε and the impaction force, and these mechanical characteristics were compared between the three compaction techniques (impaction, creep, and cyclic relaxation) with a Kruskal Wallis tests.  1.4  Stress (MPa)  1.2 1.0 0.8 0.6 0.4  E1hr  1 Ei  1  0.2 0.0 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16  Strain (-) Figure 3.4 Definition of initial compression stiffness, Ei, and stiffness after one hour creep, E1hr, under an applied stress of 1.1 MPa.  53  0.16 0.14  Strain (-)  0.12  Loading Ramp  Cε  0.10 εi  0.08 0.06 0.04  Creep  0.02 0.00 0.01  0.1  1  10  100  1000  10000  Time (s) Figure 3.5 Strain during creep test. The specimen was compressed initially from 0.034 to 1.1 MPa during a 1 s loading ramp. Subsequently, the strain was measured during 1 hour of creep at 1.1 MPa. Initial strain, εi, and creep constant, Cε, are defined as shown.  Shear tests (Figure 3.3, bottom) were performed in a custom-built shear box (Appendix 3). The apparatus was designed to be mounted directly on the testing machine used to simulate graft compaction (Instron Model 8874, Instron, Canton, Massachusetts), such that the specimens did not need to be transferred, and potentially damaged, prior to shear testing. The shear box consisted of two 19 mm diameter stainless steel cylinders. The upper cylinder was fixed rigidly to the testing machine table with a custom rig, while the lower cylinder was mounted on a linear guide. During shear testing, the lower cylinder was displaced at a rate of 1.2 mm/min by an actuator (actuator: Instron Model A591-4, Instron, Canton, Massachusetts; load cell: Sensotec Model 75/C863-01, Honeywell, Columbus, Ohio) while the MCB was subjected simultaneously to a normal compressive stress by the testing machine. Three levels of normal compressive stress, σn, were used: 125, 250, and 375 kPa, similar in magnitude to previous studies (Brewster et al., 1999; Dunlop et al., 2003; Tanabe et al., 1999). During testing, shear stress, τ, was defined as the shearing force divided by the area of intersection between the inside of the upper and lower cylinders (Equation 3.1).  54  τ (δ ) =  Fshear (d / 2) 2 cos (δ / d ) − sin( 2 cos −1 (δ / d )) 2  [  −1  ]  (3.1)  where Fshear is the shearing force, d is the inner diameter of the cylinders, and δ is the displacement of the lower cylinder relative to its initial position. This equation accounts for the reduction in specimen cross-sectional area in the shearing plane due to the shear displacement,  δ. Shear strength was defined as the shear stress at a displacement of 1 mm (approximately 5% of the cylinder diameter). For each compaction technique, a linear curve fit was obtained from the shear strength versus normal stress curves, such that the parameters of the Mohr-Coulomb failure envelope, c and φ, could be determined. Shear strength was compared between the impaction forces with a 2-way ANOVA (impaction force, normal stress). Similarly, the shear strength was compared between the creep compaction technique and impaction of the same force (300 N) with a 2-way ANOVA (compaction technique, normal stress). Student-Newman Keuls tests were used for post hoc analyses, and a significance level of 0.05 was used. In addition to the specimens used for mechanical testing, six more specimens were compacted at each impaction force level (300, 600, and 900 N) in order to measure their density (Table 3.1). The graft specimens were impacted in a 3 mm thick polyurethane cylinder with an inner diameter of 19 mm, perforated with 18 radial 1 mm holes for fluid exudation during impaction. A single 2.3 mm slice located 10 mm from the bottom surface of the specimen was scanned using peripheral quantitative computed tomography, pQCT, (Norland/Stratec XCT 2000). The scan speed was 10 mm/s with an in-plane pixel size of 0.10 x 0.10 mm. All measurements were made by a single trained technician. The apparent bone density, ρ, was measured with the Norland/Stratec XCT 5.50 software, using Contour mode 1 and adjusting the threshold such that the region of interest contained only the graft region (i.e., area inside the polyurethane cylinder). After scanning, the specimens were processed into slides for histological analysis. The top and bottom 5 mm of each of these graft specimens were removed with a diamond saw (Model 310 CP, Exakt Apparatebau). Each remaining cross-section, approximately 10 mm thick, was set in 10% buffered formalin for two days, and dehydrated in 70%, 95%, and twice in 100% ethanol for a day each. The ethanol was then evaporated at 60°C for two hours, after 55  which the cross-sections were embedded in resin (Buehler EpoThin). The middle cross-section of each block was mounted onto a slide, ground down to 200 µm, and stained with Alizarin Red S, such that the areas of bone appeared in red. In contrast, the areas of resin, which originally contained the fluid phase, appeared as white (Figure 3.6). The slides were placed on a light box to enhance the distinction between graft and void areas, and were photographed with a digital camera (resolution 3.2 Megapixels). The percentage of each cross section occupied by bone, ‘percentage bone’, was determined using the Image-Pro 4.5 (Mediacybernetics) image analysis software. The area of interest was defined as the area inside the polyurethane cylinder, and the ‘percentage bone’ was quantified as the percentage of this area occupied by bone (Figure 3.6, dark regions). Linear regression models were used to determine the percentage bone, %b (histology) and graft density, ρ (pQCT) as a function of the impaction force, as well as to examine the relationship between graft density and percentage bone.  Figure 3.6 Histological measurement of graft density, quantified as the percentage of the specimen area occupied by bone (dark).  3.3 Results The results of the creep compression tests, i.e., Ei, E1hr, εi, and Cε are shown for each group in Table 3.2. The graft compression stiffness, Ei and E1hr, correlated positively with the impaction force (both with Rs=0.89, p<0.05), whereas εi, and Cε correlated negatively with the force (Rs=-0.86 and Rs=-0.92, respectively, with p<0.05). The median E1hr in the 600 N and 900 N 56  impaction groups were 64% and 93% higher, respectively, than in the 300 N impaction group. Similarly, the median creep constant, Cε, was 56% and 63% lower in the 600N and 900N groups compared with the 300 N impaction group.  Table 3.2 Results of the creep compression test. Graft compaction technique Impaction 300 N (‘I300N’) 600 N (‘I600N’) 900 N (‘I900N’) Creep (‘C300N’) Cyclic Relaxation (‘R300N’) a  b  Ei (MPa) Median (range) 11.0 (9.5-11.5) 15.6 (12.9-17.8) 18.6 (16.5-22.6) 12.2 (11.6-12.6)a 11.9 (9.8-11.9)  εi (%) Median (range) 9.7 (9.0-10.9) 6.8 (6.0-8.0) 5.8 (4.7-6.6) 8.5 (8.2-8.7)a 9.4 (8.5-10.6)  E1hr (MPa) Median (range) 7.4 (6.6-7.7) 12.2 (10.4-13.7) 14.3 (12.5-16.6) 8.6 (8.3-8.9)a 7.4 (6.5-8.0)  Cε (% / log time) Median (range) 1.10 (1.02-1.25) 0.48 (0.44-0.53) 0.41 (0.37-0.46) 0.94 (0.84-1.01)b 1.22 (1.04-1.41)  p<0.01 compared to Impaction 300 N p<0.05 compared to Impaction 300 N  During shear testing, we did not observe a shear stress plateau when plotting shear stress versus shear displacement (Figure 3.7). By defining shear strength as the shear stress at a displacement of 1 mm (~5% of specimen diameter), we obtained the shear strength versus normal stress curves shown in Figure 3.8. From these results, the main effect of the impaction force on shear strength was significant (p<0.001). The average shear strength in the 600 N and 900 N impaction groups was 79% and 164% greater, respectively, than in the 300 N impaction group (Figure 3.8). Note that the shear strength increased with increasing normal stress (p<0.0001), and that there was no interaction effect between normal stress and impaction force. The lack of a shear stress plateau led us to use a more arbitrary shear strength definition; nevertheless, similar observations were made when comparing the shear strength results between the compaction groups regardless of the threshold displacement value used in the definition of shear strength (Table 3.3). The cohesion intercept, c, increased consistently with increasing impaction force, but friction angle, φ, did not (Table 3.3).  57  Shear stress, τ (kPa)  1200 1000 800 1 mm displacement  600 400 200 0 0  2  4  6  8  Shear displacement, δ (mm)  10  Shear strength, τ f (kPa)  Figure 3.7 Shear stress (black) versus shear displacement curve for a typical specimen. In grey is the same curve when the shear stress is derived with the initial specimen area, i.e., without accounting for the changing cross-sectional area.  900 800 I900N  700 600  I600N  500 400  C300N  300  I300N  200 100 0 100  150  200  250  300  350  400  Normal stress, σn (kPa) Figure 3.8 Results of the shear tests - shear strength, τf, versus normal stress, σn, for all impaction forces: 300 N (I300N), 600 N (I600N) and 900 N (I900N), and for the creep compaction method (C300N). Results in the C300N group were offset on the abscissa for clarity.  58  Table 3.3 Mohr-Coulomb parameters versus strength definition for three impaction forces: 300 N (I300N), 600 N (I600N) and 900 N (I900N), and for the creep compaction method (C300N). Also shown is the shear strength, τf, (median and range) under a normal stress, σn, of 375 kPa. Shear strength definition shear stress at 1 mm shear stress at 2 mm shear stress at 4 mm  c (kPa) φ (°) τf (kPa) c (kPa) φ (°) τf (kPa) c (kPa) φ (°) τf (kPa)  I300N  I600N  I900N  C300N  214 15 288 (247-315) 247 41 389 (358-462) 410 60 641 (588-797)  311 28 485 (461-626) 394 42 720 (695-832) 579 52 1044 (1022-1140)  510 29 724 (660-791) 633 41 961 (866-1064) 648 60 1254 (1188-1478)  166 18 310 (287-328) 303 19 459 (387-469) 424 41 755 (620-817)  The graft density, measured as apparent bone density, ρ (pQCT) and as the percentage bone, %b (histology) was found to correlate positively with the impaction force, F: ρ = 0.197ּF + 217 mg/cm3 (R2=0.84, p<0.0001), %b = 0.018ּF + 21.5% (R2=0.60, p=0.0002). The ρ and %b measurements were also correlated with each other: ρ = 7.04ּ%b + 108 mg/cm3, (R2=0.58, p=0.0003). Finally, the relationships between the graft density and the compressive stiffness as well as the creep constant are shown in Figures 3.9 and 3.10, respectively.  59  25  I900N  Ei (MPa)  20 15  I600N  10  I300N E i = 0.068 ρ - 7.39  5 0 200  250  300  350  400  450  500  Graft density, ρ (mg/cm3) 18 16  I900N  E1hr (MPa)  14  I600N  12 10 8  I300N E 1hr = 0.0618ρ - 9.24  6 4 2 0 200  250  300  350  400  450  500  Graft density, ρ (mg/cm3) Figure 3.9 Post-impaction stiffness, Ei (top) and E1hr (bottom) versus graft density for each force group. Shown are medians and ranges, and the median values were used for curve fitting. See Tables 3.2 and 3.4.  60  Cε (% / log time (s))  1.4 1.2  C ε = 12.111e-0.009ρ 2 R = 0.9447  I300N  1.0 0.8 0.6  I600N  0.4  I900N  0.2 0.0 200  250  300  350  400  450  500  Graft density, ρ (mg/cm ) 3  Figure 3.10 Creep constant during post-impaction creep compression versus graft density for each force group (medians and ranges, see Tables 3.2 and 3.4). Table 3.4 Graft density and percentage bone as a function of impaction force. Medians (ranges).  Impaction force 300 N 600 N 900 N  Graft density, ρ Percentage bone, (mg/cm3) %b (%) 272 (251-301) 27 (21-32) 342 (306-365) 31 (28-36) 385 (371-453) 38 (33-45)  The main effect of the compaction technique on the graft compression behaviour, Ei, E1hr, εi, and Cε was statistically significant (p<0.01) but for practical purposes the differences in median values between the groups were small, Table 3.2. Compared to impaction, the creep technique resulted in a small increase in stiffness and a small decrease in εi, and Cε, e.g.,16% higher median E1hr, whereas the cyclic relaxation technique did not affect the compression behaviour. Similarly, the main effect of the compaction technique on shear strength was significant (p<0.0001, Figure 3.8) but the difference in average shear strength between the two groups is for practical purposes small. The creep compaction technique resulted in 14% greater shear strength, on average, compared with impaction at the same force - however, its effect on  61  the Mohr Coulomb coefficients was not clear as it varied as a function of the definition of shear strength used (Table 3.3). Stiffness as measured during impaction, Eimp, increased with each impact (Figure 3.11). The Eimp during the twentieth impact was proportional to the impaction force (Rs=0.90 p>0.05). The median Eimp was 70.8 MPa (range 66.0-75.6 MPa), 78.6 MPa (71.6-84.2 MPa), and 87.0 MPa (84.6-96.8 MPa), in the 300 N, 600N and 900 N impaction groups, respectively. Finally, the stiffness of the impacted MCB, Ei and E1hr, were correlated with Eimp at the twentieth impaction (Rs=0.92 for both, with p>0.05). 60  100 90 80 70 60 50 40 30 20 10 0  Stiffness (MPa)  50  Recoil (60 s) 40  30  I300N I600N I900N  Creep (3600 s) Ei  E1hr  20  Impaction  10  0  0  10  Number of impactions  20  10 30  100  1000  10000  Time following impaction (s)  Figure 3.11 Median stiffness of the graft during impaction, Eimp, and during subsequent creep test for all impaction forces.  3.4 Discussion Morsellized cancellous bone grafts exhibit viscoelastic-viscoplastic behaviour. In other words, MCB deforms under load in a way that is partly recoverable (elastic) and partly nonrecoverable (plastic), in a process that is time-dependant (viscous). Simple confined compression and shear tests are useful in estimating the effect of surgical parameters on the mechanical performance of impacted MCB. However, it must be emphasized that viscoelasticviscoplastic materials do not exhibit distinctive stiffness or shear strength properties, and that the apparent stiffness or shear strength measurement of such a material are sensitive to timerelated testing parameters such as the rate and duration of loading. Moreover, MCB has been 62  described as a nonlinear viscoelastic material (Phillips et al., 2006a), indicating that its mechanical properties also depend on the stress applied. One must therefore be careful when comparing the results of MCB characterization studies. A wide range of stiffness values have been reported for fresh frozen, impacted MCB: 3-135 MPa (Giesen et al., 1999; Voor et al., 2000; Verdonschot et al., 2001; Bavadekar et al., 2001; Cornu et al., 2003b; Cornu et al., 2004; Voor et al., 2004; Fosse et al., 2004; Phillips et al., 2006a; Phillips et al., 2006b; Fosse et al., 2006). This large range of results is likely due to variation in the test methods among the studies. Our post-impaction MCB stiffness under creep loading, E1hr, ranged between 7 and 17 MPa – which is comparable to the results of other studies that have similarly measured stiffness in creep under a similar stress: 3-8 MPa (Voor et al., 2000; Voor et al., 2004; Fosse et al., 2004). Studies that measured stiffness immediately after impaction, without allowing the graft to recoil, consistently reported higher stiffness: 1070 MPa (Phillips et al., 2006b; Fosse et al., 2006). Similarly, stiffness measured during a loading ramp ranged 40-70 MPa (Cornu et al., 2003a; Cornu et al., 2004) and during impaction or cyclic loading, 15-135 MPa (Verdonschot et al., 2001; Phillips et al., 2006b). Our stiffness measurements during the twentieth impaction, Eimp, 66-97 MPa, were within the range reported by others (Phillips et al., 2006b). Dynamic stiffness measurements, however, are influenced by the fluid phase, rendering the measurements sensitive to the loading rate and that of fluid dissipation, which is affected by the geometry and permeability of the graft and its environment. By characterizing MCB stiffness in creep, we were able to account for fluid effects and, therefore, to measure the response of the graft particles to the applied load. Finally, we believe that measuring stiffness post-impaction, and after having allowed the graft to recoil, provides a more clinically relevant measure of graft stiffness since recoil is unavoidable during surgery. It has been demonstrated that MCB stiffness increases with the number of impactions (Bavadekar et al., 2001; Cornu et al., 2003a). In these studies, however, the impaction force was not reported. In the current study, Eimp appeared to have reached a plateau within ten impactions at forces of 600 N and 900 N (2 and 3 MPa), whereas with a force of 300 N, Eimp was still increasing beyond the first ten impactions (Figure 3.11). It did not appear, however, that further increase in stiffness would be achieved beyond 20 impactions. In another study, the 63  stiffness of bovine MCB was shown to increase with the number of impactions, or ‘cycles’, and increasing compacting force (Phillips et al., 2006b). Their specimens, however, were impacted in force increments, thus the effect of force was not isolated from that of the number of impactions – furthermore, no recoil was allowed prior to stiffness measurement, therefore it is not clear whether the effect of the force they observed would remain following the removal of the impaction tool. Another study demonstrated that the graft compression stiffness is proportional to the drop height of a slap hammer, and reported a correlation between MCB stiffness and density (Fosse et al., 2006). In their study, however, the impaction force was neither controlled nor reported specifically. Furthermore, their bone mineral density measurements appeared to have been taken following creep testing at 2.3 MPa and they reported using a correction factor, but it was not described. Our results agree that increasing the impaction force increases the post-impaction compressive stiffness of MCB, and demonstrate that this improvement is related to the graft density. Since our density measurements were taken after impaction without subjecting the specimens to further testing that may affect the density, we believe this provides a more clinically relevant measure of the MCB density at the time that the stem is cemented in place. Nonetheless, it is worth noting that the shear strength and stiffness of the graft bed may be affected by the endosteum roughness, and the fluid paths, which can be influenced by the absence of drainage holes in the femur, the varying canal geometry, and the size of the gap around the impactor. For this reason, the stiffness and shear strength measured in this study should be regarded as approximate values. The stiffness of structural cancellous bone is known to be roughly proportional to the square (Hayes and Bouxein, 1997) or cube (Carter and Hayes, 1977) of its apparent density. The nature of the relationship between the stiffness and density of morsellized cancellous graft, however, is not known. A linear regression offered a reasonable fit to our stiffness versus apparent density data (Figure 3.9), but it is not clear whether this linear relationship would apply over a wider range of apparent graft density. The correlation observed between the two density measurements, %b and ρ, was somewhat weak, i.e., R2=0.58. This may be explained by the following observations. The ρ and %b measurements were taken at approximately but not exactly the same location. Furthermore, while the pQCT measurements of ρ were obtained from three-dimensional voxels (0.1 mm x 64  0.1 mm x 2.3+/-0.2 mm), the histological measurements of %b were obtained from a single two-dimensional cross section. Finally, a small swelling of the graft occurred during the preparation of the histological slides, which may have led to a lower %b measurement, and this swelling was not controllable between the specimens. Based on the current results, the relationship between the compression properties of MCB and its density can be described by the following equations, with: A  specimen cross sectional area  Ho  height of the graft specimen prior to impaction  Vo  specimen volume prior to impaction  ρo  graft apparent bone density prior to impaction  H  instantaneous specimen height during impaction  V  instantaneous specimen volume  ρ  instantaneous graft density  εN imp  MCB strain during impaction (natural)  The natural compressive strain of the graft during impaction, εN imp can be calculated as: ⎛ Ho ⎞ ⎟ ⎝ H ⎠  ε N imp = ln⎜  (3.2)  During impaction, fluid flows out of the MCB, and conservation of mass of the graft particles gives:  ρ o V o= ρ V  (3.3)  Because the cross sectional area is constant during the test: H V = H o Vo  (3.4)  From (2-4) we can estimate the instantaneous bone density of the MCB:  ρ = ρoe  ε N imp  (3.5)  65  We can estimate ρo from our correlation results (graft density versus impaction force) as ρ resulting from an impaction force of zero, i.e., approximately 217 mg/cm3. Under post-impaction compression loading, we observed that the initial stiffness, Ei, of the impacted MCB (loading ramp to 1.1 MPa in 1 s) and the stiffness after one hour creep, E1hr, were proportional to the graft density achieved during impaction (Figure 3.9):  and  Ei ( ρ ) = C1 ⋅ ρ + C 2  (3.6)  E1hr ( ρ ) = C 3 ⋅ ρ + C 4  (3.7)  where C1 ≈ 0.068 MPa cm3/mg, C2 ≈-7.39 MPa, C3 ≈0.062 MPa cm3/mg, and C4 ≈-9.24 MPa. Given the relationship in Equation 3.6 and the data shown in Figure 3.5, knowing the MCB density post-impaction, we can estimate the total creep strain, ε, that will develop over time under a sustained confined compression stress, σo :  ε (t ) ≈ ε i ( ρ ) + Cε ( ρ ) ⋅ log t =  σo + Cε ( ρ ) ⋅ log t Ei ( ρ )  (3.8)  Furthermore, from the results shown in Figure 3.10, with a compressive stress σo of 1.1 MPa, the creep constant of the MCB, Cε, was related to the graft density as follows: Cε ( ρ ) = D1 e − D2 ρ  (3.9)  where D1 ≈ 12.1, and D2 ≈ 0.009 cm3/mg. Finally, by combining Equations 3.6, 3.8 and 3.9, we can express the creep compression strain  ε(t) as a function of the post-impaction graft density (Equation 3.10a) or its density before impaction, ρo, and total strain during impaction, εN imp (Equation 3.10b):  ε (t ) = ε (t ) =  σo + D1 e −( D ⋅ρ ) ⋅ log t C1 ρ + C 2  (3.10a)  2  σo C1 ρ o e  ε N imp  ε N imp  + C2  + D1 e −( D2 ⋅ρoe  )  ⋅ log t  (3.10b)  66  Equations 3.2-3.10 illustrate how the density achieved during compaction affects its stiffness and creep behaviour. Constitutive equations have been previously proposed to describe the time-dependent behaviour of MCB under compressive loading (Giesen et al., 1999), and the relationship between stiffness of MCB and the compressive stress to which it is subjected (Phillips et al., 2006a). In these studies, the graft was subjected to 23 and 750 cycles of preconditioning at high loads (2.3-3 MPa) prior to the stiffness measurement. Consequently, their specimens would have been of roughly uniform, high density. Our work expands on those findings by factoring in the effect of the graft density achieved during impaction. Further work, however, is required to determine if and how Cε, Ei, and E1hr vary as a function of the compression stress, σo. The creep technique brought some improvement in compression properties compared with impaction: e.g., 16% increase in E1hr (median values, Table 3.2). We can use the derived constitutive equations explain this finding. From Equation 3.7, we can estimate that for a 25% gain in E1hr, we would need to achieve a graft density of approximately 298 mg/cm3 (i.e., a compaction strain εN  imp  of 0.317, assuming an unimpacted density of 217 mg/cm3). When  extrapolating the strain-time data obtained during compaction with the creep technique, we estimate that to achieve such graft density, the 300 N compression load would have to be held for approximately 4 hours. This would not be practical in a clinical setting. Finally, the cyclic relaxation technique did not offer any benefit to the compressive properties of MCB. A recent study reported that applying a small pressure to the graft bed following impaction can offer a considerable increase in stiffness compared with impaction alone (Lunde et al., 2008). The current study used considerably larger pressure and yet offered only a modest increase in graft stiffness compared with impaction. Nonetheless, further consideration may be warranted towards a technique that incorporates both impaction and a slow compression of the graft. The shear strengths of MCB reported in the literature also vary greatly between studies. For example, under a normal stress of 350-370 kPa, shear strengths between 225 kPa (Dunlop et al., 2003) and 1700 kPa (Tanabe et al., 1999) have been reported. The current results are within the range of values seen in the literature. In contrast to what is often reported for hard particulate materials, no shear stress plateau was observed (Figure 3.7, black line). More 67  specifically, if we defined shear stress as the shearing force divided by the initial crosssectional area, the shear stress would have reached a plateau after approximately 3 mm of shear displacement (Figure 3.7, grey line). This shear strength definition, however, would be inaccurate for a large displacement. We therefore accounted for the changing specimen crosssectional area in our definition of shear stress, after which no shear stress plateau was observed even at a displacement of 9 mm (50% of specimen diameter). For this reason, we used a threshold displacement of 1 mm (5% of specimen diameter) in our definition of shear strength. This makes the definition of shear strength ambiguous and direct comparison with literature values difficult. It is not known what causes the lack of a shear stress plateau, but qualitative observations during the tests revealed a substantial amount of deformation and cohesiveness of the MCB particles – this behaviour is in sharp contrast to what is seen when a stiff particulate material such as sand or gravel is tested in shear. The lack of plateau may also be attributed partly to an increase in ‘true’ normal stress due to the decreasing cross-sectional area in the shear plane. This kind of error would have probably been much smaller with harder particles for which shear failure would have occurred at a smaller shear displacement. Our results demonstrate that the shear strength of MCB is proportional to the impaction force. Because MCB particles are highly deformable, the ‘packing density’ may therefore depend more on the impaction force than on the particle size distribution. Finally, the creep technique offered a 14% average increase in shear strength compared with impaction. Shear failure and creep consolidation of the graft bed are not the only mechanisms that are believed to influence implant migration in impaction allografting. Other mechanisms such as cement fatigue (Masterson et al., 1997), graft incorporation/fibrous invasion (Linder, 2000; Nelissen et al., 1995), and cancellization of the femoral canal (Frei et al., 2005b) are also suspected to play a role in implant stability, while previous studies also point to the importance of cement contact with the endosteum (Frei et al., 2005a; Albert et al., 2007). Nonetheless, clinical studies have shown that a substantial amount of subsidence develops in the early postoperative days: more than 1 mm subsidence was often observed within the first three months after surgery (Ornstein et al., 2001; Nelissen et al., 2002) and more than half of the subsidence observed at six weeks was seen typically within the first two weeks (Ornstein et al., 2000). Furthermore, a recent clinical study indicated that there appears to be relationship 68  between implant migration and the density of the graft bed (Gokhale et al., 2005). These observations point to the importance of adequate MCB compaction, and the resulting shear and compression properties. A number of surgical parameters have been reported to affect graft properties. The following parameters were observed to affect the shear strength of MCB: rinsing the graft increased shear strength by 15-25% (Dunlop et al., 2003); controlling the graft particle size distribution affected shear strength by 8-14% (Dunlop et al., 2003); and using a Bioglass extender to optimize the particle size distribution increased shear strength by 7% (Brewster et al., 1999). Washing the graft with a detergent increased the compression stiffness by 5% (Voor et al., 2004), while combining washing, freeze-drying and irradiation increased graft stiffness by 17% when compared with fresh-frozen graft (Cornu et al., 2004). With increases of 14% in stiffness and 16% in shear strength, our proposed creep technique appears to offer a benefit comparable to other surgical parameters such as optimizing the particle size distribution. However, it is not yet clear whether this small increase in graft stiffness is of clinical relevance, and for this reason, we do not recommend changing the current technique of impaction. A recent study investigated the potential benefit of vibration during graft compaction in a composite femur model of impaction allografting – on average, the use of vibration compacted 9% more graft (by weight) in the medullary canal compared with standard impaction, and this resulted in 28% less migration (Bolland et al., 2007). These results indicate that a small increase in graft density may have a large effect on implant stability. Other recent studies, however, have observed extensive cement contact with the endosteum after impaction allografting in a cadaver model (Frei et al., 2005a; Albert et al., 2007; Frei et al., 2004). This leads to the following questions. Does implant stability rely on a cement-endosteum interface rather than on the stiffness and strength of the graft bed? Alternatively, how dense must the graft bed be to adequately support the implant without cement-endosteum contact? We believe that it is essential to answer these two questions prior to proposing changes to the current graft impaction technique. Within the range of forces used in impaction allografting, the impaction force has a substantial effect on the graft compression stiffness and shear strength. The current results point to the 69  importance of maximizing impaction forces – therefore it would be beneficial to reinforce the femur during surgery to minimize the risk of intraoperative fracture, especially when dealing with severe bone loss. A graft compaction method yielding greater strength and stiffness without increasing the compaction force could reduce the risk of fracture. Our ‘creep method’ resulted in a small increase in stiffness and shear strength compared with standard graft impaction; however it is not clear whether this would be clinically beneficial in impaction allografting.  70  3.5 References Albert, C., Patil, S., Frei, H., Masri, B., Duncan, C., Oxland, T., and Fernlund, G. (2007) Cement penetration and primary stability of the femoral component after impaction allografting. A biomechanical study in the cadaveric femur. J Bone Joint Surg Br 89, 962-70. Bavadekar, A., Cornu, O., Godts, B., Delloye, C., Van Tomme, J., and Banse, X. (2001) Stiffness and compactness of morselized grafts during impaction: an in vitro study with human femoral heads. Acta Orthop Scand 72, 470-6. Bolland, B.J., New, A.M., Madabhushi, S.P., Oreffo, R.O., and Dunlop, D.G. (2007) Vibration-assisted bone-graft compaction in impaction bone grafting of the femur. J Bone Joint Surg Br 89, 686-92. Brewster, N.T., Gillespie, W.J., Howie, C.R., Madabhushi, S.P., Usmani, A.S., and Fairbairn, D.R. (1999) Mechanical considerations in impaction bone grafting. J Bone Joint Surg Br 81, 118-24. Carter, D.R., Hayes, W.C. (1977) The behavior of bone as a two-phase porous structure. J Bone and Joint Surg Am 59, 954-62. Cornu, O., Bavadekar, A., Godts, B., Van Tomme, J., Delloye, C., and Banse, X. (2003a) Impaction bone grafting with freeze-dried irradiated bone. Part II. Changes in stiffness and compactness of morselized grafts: experiments in cadavers. Acta Orthop Scand 74, 553-8. Cornu, O., Bavadekar, A., Godts, B., Van Tomme, J., Delloye, C., and Banse, X. (2003b) Impaction bone grafting with freeze-dried irradiated bone. Part I. Femoral implant stability: cadaver experiments in a hip simulator. Acta Orthop Scand 74, 547-52. Cornu, O., Libouton, X., Naets, B., Godts, B., Van Tomme, J., Delloye, C., and Banse, X. (2004) Freeze-dried irradiated bone brittleness improves compactness in an impaction bone grafting model. Acta Orthop Scand 75, 309-14. Craig RF. (1983) Soil Mechanics. 3rd Edition. London, UK: Nostrand Reinholds, pp.6-7. Dunlop, D.G., Brewster, N.T., Madabhushi, S.P., Usmani, A.S., Pankaj, P., and Howie, C.R. (2003) Techniques to improve the shear strength of impacted bone graft: the effect of particle size and washing of the graft. J Bone Joint Surg Am 85-A, 639-46. Eldridge, J.D., Smith, E.J., Hubble, M.J., Whitehouse, S.L., and Learmonth, I.D. (1997) Massive early subsidence following femoral impaction grafting. J Arthroplasty 12, 535-40. Fetzer, G.B., Callaghan, J.J., Templeton, J.E., Goetz, D.D., Sullivan, P.M., and Johnston, R.C. (2001) Impaction allografting with cement for extensive femoral bone loss in revision hip surgery: a 4- to 8-year follow-up study. J Arthroplasty 16, 195-202. Fosse, L., Muller, S., Ronningen, H., Irgens, F., and Benum, P. (2006) Viscoelastic modelling of impacted morsellised bone accurately describes unloading behaviour: an experimental study of stiffness moduli and recoil properties. J Biomech 39, 2295-302. 71  Fosse, L., Ronningen, H., Lund-Larsen, J., Benum, P., and Grande, L. (2004) Impacted bone stiffness measured during construction of morsellised bone samples. J Biomech 37, 1757-66. Frei, H., Mitchell, P., Masri, B.A., Duncan, C.P., and Oxland, T.R. (2004) Allograft impaction and cement penetration after revision hip replacement. A histomorphometric analysis in the cadaver femur. J Bone Joint Surg Br 86, 771-6. Frei, H., Mitchell, P., Masri, B.A., Duncan, C.P., and Oxland, T.R. (2005a) Mechanical characteristics of the bone-graft-cement interface after impaction allografting. J Orthop Res 23, 9-17. Frei, H., O'Connell, J., Masri, B.A., Duncan, C.P., and Oxland, T.R. (2005b) Biological and mechanical changes of the bone graft-cement interface after impaction allografting. J Orthop Res 23, 1271-9. Giesen, E.B., Lamerigts, N.M., Verdonschot, N., Buma, P., Schreurs, B.W., and Huiskes, R. (1999) Mechanical characteristics of impacted morsellised bone grafts used in revision of total hip arthroplasty. J Bone Joint Surg Br 81, 1052-7. Gokhale, S., Soliman, A., Dantas, J.P., Richardson, J.B., Cook, F., Kuiper, J.H., and Jones, P. (2005) Variables affecting initial stability of impaction grafting for hip revision. Clin Orthop Relat Res 432, 174-80. Hayes, W.C., Bouxsein, M.L. (1997) Biomechanics of cortical and trabecular bone: implications for assessment of fracture risk. In: Mow, V.C., Hayes, W.C., editors. Basic Orthopaedic Biomechanics. Philadelphia: Lippincott-Raven. 69–111. Knight, J.L. and Helming, C. (2000) Collarless polished tapered impaction grafting of the femur during revision total hip arthroplasty: pitfalls of the surgical technique and follow-up in 31 cases. J Arthroplasty 15, 159-65. Linder, L. (2000) Cancellous impaction grafting in the human femur: histological and radiographic observations in 6 autopsy femurs and 8 biopsies. Acta Orthop Scand 71, 543-52. Lunde, K.B., Kaehler, N., Ronningen, H., and Fosse, L. (2008) Pressure during compaction of morsellised bone gives an increase in stiffness: an in vitro study. J Biomech 41, 231-4. Masterson, E.L., Masri, B.A., and Duncan, C.P. (1997) The cement mantle in the Exeter impaction allografting technique. A cause for concern. J Arthroplasty 12, 759-64. Mow, V.C., Huiskes, R. (2005) Basic Orthopaedic Biomechanics and Mechano-Biology. 3rd Edition. Philadelphia: Lippincott Williams & Wilkins. Nelissen, R.G., Bauer, T.W., Weidenhielm, L.R., LeGolvan, D.P., and Mikhail, W.E. (1995) Revision hip arthroplasty with the use of cement and impaction grafting. Histological analysis of four cases. J Bone Joint Surg Am 77, 412-22. Nelissen, R.G., Valstar, E.R., Poll, R.G., Garling, E.H., and Brand, R. (2002) Factors associated with excessive migration in bone impaction hip revision surgery: a 72  radiostereometric analysis study. J Arthroplasty 17, 826-33. Ornstein, E., Atroshi, I., Franzen, H., Johnsson, R., Sandquist, P., and Sundberg, M. (2001) Results of hip revision using the Exeter stem, impacted allograft bone, and cement. Clin Orthop Relat Res 389, 126-33. Ornstein, E., Atroshi, I., Franzen, H., Johnsson, R., Sandquist, P., and Sundberg, M. (2002) Early complications after one hundred and forty-four consecutive hip revisions with impacted morselized allograft bone and cement. J Bone Joint Surg Am 84-A, 1323-8. Ornstein, E., Franzen, H., Johnsson, R., Karlsson, M.K., Linder, L., and Sundberg, M. (2004) Hip revision using the Exeter stem, impacted morselized allograft bone and cement: a consecutive 5-year radiostereometric and radiographic study in 15 hips. Acta Orthop Scand 75, 533-43. Ornstein, E., Franzen, H., Johnsson, R., and Sundberg, M. (2000) Radiostereometric analysis in hip revision surgery--optimal time for index examination: 6 patients revised with impacted allografts and cement followed weekly for 6 weeks. Acta Orthop Scand 71, 360-4. Phillips, A., Pankaj, P., May, F., Taylor, K., Howie, C., and Usmani, A. (2006a) Constitutive models for impacted morsellised cortico-cancellous bone. Biomaterials 27, 2162-70. Phillips, A.T., Pankaj, Brown, D.T., Oram, T.Z., Howie, C.R., and Usmani, A.S. (2006b) The elastic properties of morsellised cortico-cancellous bone graft are dependent on its prior loading. J Biomech 39, 1517-26. Piccaluga, F., Gonzalez Della Valle, A., Encinas Fernandez, J.C., and Pusso, R. (2002) Revision of the femoral prosthesis with impaction allografting and a Charnley stem. A 2- to 12-year follow-up. J Bone Joint Surg Br 84, 544-9. Schreurs, B.W., Arts, J.J., Verdonschot, N., Buma, P., Slooff, T.J., and Gardeniers, J.W. (2005) Femoral component revision with use of impaction bone-grafting and a cemented polished stem. J Bone Joint Surg Am 87, 2499-507. Tanabe, Y., Wakui, T., Kobayashi, A., Ohashi, H., Kadoya, Y., and Yamano, Y. (1999) Determination of mechanical properties of impacted human morsellized cancellous allografts for revision joint arthroplasty. J Mater Sci Mater Med 10, 755-60. Ullmark, G. and Nilsson, O. (1999) Impacted corticocancellous allografts: recoil and strength. J Arthroplasty 14, 1019-23. van Biezen, F.C., ten Have, B.L., and Verhaar, J.A. (2000) Impaction bone-grafting of severely defective femora in revision total hip surgery: 21 hips followed for 41-85 months. Acta Orthop Scand 71, 135-42. Verdonschot, N., van Hal, C.T., Schreurs, B.W., Buma, P., Huiskes, R., and Slooff, T.J. (2001) Time-dependent mechanical properties of HA/TCP particles in relation to morsellized bone grafts for use in impaction grafting. J Biomed Mater Res 58, 599-604.  73  Voor, M.J., Nawab, A., Malkani, A.L., and Ullrich, C.R. (2000) Mechanical properties of compacted morselized cancellous bone graft using one-dimensional consolidation testing. J Biomech 33, 1683-8. Voor, M.J., White, J.E., Grieshaber, J.E., Malkani, A.L., and Ullrich, C.R. (2004) Impacted morselized cancellous bone: mechanical effects of defatting and augmentation with fine hydroxyapatite particles. J Biomech 37, 1233-9.  74  CHAPTER 4  INFLUENCE OF CEMENT PROFILE AND GRAFT PROPERTIES ON STEM MIGRATION AND MICROMOTION A FINITE ELEMENT STUDY  A version of this chapter has been accepted for publication. Albert C, Masri B, Duncan C, Oxland T, Fernlund G (2009). Influence of cement penetration and graft density on stem stability in impaction allografting: A finite element study. Clinical Biomechanics. 75  4.1 Introduction Implant loosening is the leading cause of femoral implant failure in total hip arthroplasty (THA) (Karrholm et al., 2007; Canadian Institute for Health Information, 2008; Bozic et al., 2009). Loosening is often accompanied by bone loss, which can pose a challenge during revision surgery (Engelbrecht et al., 1990; Chandler et al., 1994; Haddad and Duncan, 1999). Impaction allografting was introduced to address the problem of bone loss in revision THA, however, some concerns have been raised about the high incidences of femoral fractures (Knight and Helming, 2000; Fetzer et al., 2001; Schreurs et al., 2005) and excessive implant migration relative to bone (Meding et al., 1997; Eldridge et al., 1997; van Biezen et al., 2000; Robinson et al., 2002; Piccaluga et al., 2002). While implant migration may not necessarily lead to failure, excessive migration (>10 mm) increases the risk of thigh pain, dislocation and the need for re-revision – for this reason, excessive migration is often deemed a failure of the reconstruction (Meding et al., 1997; Eldridge et al., 1997; Masterson et al., 1997a; Robinson et al., 2002). It has been suggested that early implant migration may be caused by: settling of the stem into the cement mantle (Malkani et al., 1996); postoperative compaction of the graft (Nilsson and Karrholm, 1996; Linder, 2000, Halliday et al. 2003); shear failures in the graft (Brewster et al., 1999; Dunlop et al., 2003; Halliday et al., 2003; Ornstein et al., 2004; Bolland et al., 2006; Bolland et al., 2007); and sliding of the graft at the endosteum (Ornstein et al., 2004; Frei et al., 2005a). In the long-term, other variables such as cement fatigue (Ornstein et al., 2004), femoral expansion due to ageing or interruption of vascular supply (Ornstein et al., 2004; Frei et al., 2005b), and remodeling of the bone graft (Franzen et al., 1995) may also contribute to migration. However, it is not known what effect these events and mechanisms have on longterm implant stability, nor is the wide range of migration seen in impaction allografting understood. Morsellized cancellous bone graft has substantially lower stiffness compared to the other materials in an impaction allografting construct (Voor et al., 2004; Albert et al., 2008), and it may be susceptible to shear failure due to its particulate nature (Brewster et al., 1999). It is therefore believed that graft bed density plays a role in implant stability (Lind et al., 2002;  76  Cabanela et al., 2003; Pekkarinen et al., 2000; Morgan et al., 2004; Gokhale et al., 2005; Bolland et al., 2007). For this reason, numerous studies have focused on the mechanical characteristics of morsellized bone graft (Brodt et al., 1998; Brewster et al., 1999; Tanabe et al. 1999; Giesen et al., 1999; Voor et al., 2000; Schreurs et al., 2001; Bavadekar et al., 2001; Dunlop et al., 2003; Cornu et al., 2003; Voor et al., 2004; Cornu et al., 2004a; Cornu et al., 2004b; Fosse et al., 2004; Butler et al., 2005; Fosse et al., 2006; Phillips et al., 2006a; Phillips et al., 2006b; Albert et al., 2008). In impaction allografting, however, the graft bed is infiltrated in part with bone cement, and for this reason the role of the graft in the migration process may be more complex. In recent in vitro studies, high levels of cement penetration into the graft bed were observed after impaction allografting, resulting in substantial cement contact with the endosteum around the distal half of the stem (Frei et al., 2004; Albert et al., 2007). Stem migration was also found to increase with a decrease in amount of cement and cementendosteum contact (Albert et al., 2007). The relative importance of graft bed density and cement penetration profile, however, is not known. The first objective of this study was to determine the effects of graft density and cement penetration profile on the shear stresses at the endosteum interface. The second objective was to estimate the effects of four variables on stem migration and micromotion: graft density, cement penetration profile, and the status of the stem-cement and endosteum interfaces; i.e., bonded or sliding. The third objective was to estimate the relative contribution of each of the following mechanisms on implant motion: sliding at the stem-cement interface, shear deformation of the cement mantle layer, shear deformation of the graft and/or graft/cement composite region, and sliding at the endosteum interface – and how these four contributions vary as a function of the four study variables. These objectives were addressed using a finite element model, and the resulting migration and micromotion were validated against experimental in vitro results.  4.2 Methods A finite element mesh of 2004 eight-noded axisymmetric elements was built using ANSYS 11.0 (ANSYS, Inc., Canonsburg, PA, USA) to simulate the proximal femur after impaction allografting. Four variables were incorporated into the model. The first variable was the cement  77  penetration profile; nine profiles were simulated, ranging from little cement penetration into the graft to extensive cement contact with the endosteum (Figure 4.1). The second variable was graft density, which was simulated by controlling the stiffness and shear strength of the graft and graft-cement composite regions to represent low, moderate, and high graft impaction forces. The last two variables were the status of the stem-cement and cement-endosteum interfaces; i.e., bonded or sliding. The effects of these parameters on implant motion were examined during cyclic loading corresponding to a walking cycle. A total of 90 simulations were conducted, and the simulations and post-processing were performed using Abaqus/Standard and Abaqus/CAE 6.6-1 (Abaqus Inc./Simulia Inc. Rhode Island, USA).  Profile Profile Profile Profile Profile A E C D B  Profile Profile Profile F H G  Profile I Stem-cement interface Cement-endosteum interface Stem Cement mantle Graft/cement composite Femur Centralizer Graft Graft-endosteum interface Distal plug  Figure 4.1 Geometries of the nine cement penetration profiles modeled. Profiles A-C: idealized constructs with uniform cement penetration that reached depths of 25%, 50%, and 75% of the graft bed width, respectively. Profiles D-I: constructs with 11%, 24%, 36%, 49%, 61% and 80% cement contact with the endosteal surface, respectively (the width-to-length aspect ratio was increased to facilitate viewing the cement profiles).  The distal 13 cm of a stem was modeled as an axisymmetric solid body with a total taper angle of 4.4°, i.e., the average between the taper angles in the frontal and lateral planes for a CPT stem (CPT 12/14 Hip System, Size 2, Zimmer Inc.). The femur was modeled as a tapered axisymmetric cylinder mimicking the dimensions of proximal human femurs with moderate bone loss and cortical thinning. The following morphometric features were chosen based on measurements from radiographs of twelve human femur specimens (six donors) on which impaction allografting had been performed: an inner diameter of 14 mm at stem tip level; an 78  endosteal taper angle of 2.4° from the axis of symmetry (total taper angle of 4.8°); and a cortical thickness of 5 mm at stem tip level and 2 mm at 13 cm above the stem tip. The top of the distal plug was located 38 mm distal from the stem tip. Nine distinct profiles of cement penetration into the graft were simulated (Figure 4.1). In all profiles, a 2 mm thick layer of pure cement (cement mantle) was modeled adjacent to the stem, and a polyethylene centralizer was included at the stem tip. The remaining canal content consisted of areas of pure graft and of graft/cement composite, forming nine configurations, each representing a distinct cement penetration profile. Profiles A, B and C represented idealized constructs with uniform cement penetration that did not reach the endosteum, but reached a depth of 25%, 50%, and 75% of the graft bed width, respectively. In profiles D through I, the graft/cement composite region was in contact with 11%, 24%, 36%, 49%, 61% and 80% of the endosteum interface, respectively. The centers of the cement-endosteum contact regions were located 50 mm above the stem tip, to mimic cement profiles that were observed in previous in vitro studies, in which cement-endosteum contact was seen predominantly around the mid and distal stem (Albert et al., 2007; Frei et al., 2004). The stem, cement and femur were modeled with linear elastic behaviour. The stem consisted of a CoCrMo alloy and was given a Young’s modulus, E, of 210 GPa (Hallab et al., 2004) and a Poisson’s ratio, ν, of 0.3. The femoral cortex was assumed isotropic and assigned E = 20 GPa (Reilly and Burnstein, 1975) and ν = 0.28 (Pidaparti and Vogt, 2002); and the bone cement E = 2 GPa (Kindt-Larsen et al., 1995) and ν = 0.37 (Guild et al., 1994). The graft was modeled with elasto-plastic behaviour. In tension and compression, the graft behaviour was assumed linear elastic, while in shear it was assumed linear elastic up the onset of failure where it exhibited plastic behaviour with a pressure-dependent Mohr-Coulomb shear failure envelope that was defined based on experimental data. The graft properties, shown in Table 4.1, represented low, moderate and high graft densities based on experimental data (Chapter 3). The Mohr-Coulomb cohesion and friction parameters were assumed constant for each graft density. The graft was assigned a Poisson’s ratio, νgraft, of 0.2 (Brodt et al., 1998).  79  Table 4.1 Graft properties used in the finite element model, representing low, moderate or high graft density. Graft impaction Graft apparent Volume fraction pressure density of bone, f (MPa) (mg/cm3) (%) 1 272 (low) 27 2 342 (moderate) 31 3 385 (high) 38 Values taken from (Albert et al., 2008).  Egraft (MPa) 7.4 12.2 14.3  Mohr-Coulomb parameters c (kPa) φ (°) 214 15 311 28 510 29  The graft/cement composite region was assumed isotropic. Its modulus and Poisson’s ratio were estimated for each graft density level using Equations 4.1-4.4 (Halpin and Tsai, 1969; Halpin 1984). The volume fraction of graft in the composite regions, f, was estimated based on the percentage of bone observed experimentally in morsellized graft specimens for each graft density level (Chapter 3); these values are shown in Table 4.1. It was assumed that cement filled all of the spaces that were not occupied by bone in the graft/cement composite regions, such that the volume fraction of cement was equal to 1-f. From Equations 4.1-4.4, it was estimated that Ecomposite and νcomposite were 1289 MPa and 0.32, 1201 MPa and 0.32, and 1053 MPa and 0.31 for the low, moderate and high density level, respectively. ⎛ 1 + ζηf E composite = E cement ⎜⎜ ⎝ 1 − ηf  ⎞ ⎟⎟ ⎠  ⎛ E graft ⎞ ⎜⎜ ⎟ −1 E cement ⎟⎠ ⎝ with η = ⎛ E graft ⎞ ⎜⎜ ⎟⎟ + ζ E ⎝ cement ⎠ and  (4.1)  (4.2)  ζ ≈ 2 + 40 f 10  (4.3)  ν composite = fν graft + (1 − f )vcement  (4.4)  The stem-cement interface was modeled as either bonded or as sliding with a Coulomb friction model and µ = 0.3 (Nuno et al., 2002; Nuno et al., 2006). The graft-endosteum interface was modeled as sliding with µ = 0.61 (Zhang et al., 1999), and the cement-endosteum interface was modeled as bonded or sliding with µ = 0.61.  80  Loading was applied in two steps. In step 1, an axial force of 1646 N (14.5 MPa) was applied to the proximal cross section of the stem in the distal direction. In step 2, the force was reduced to 213 N (1.9 MPa). These two force levels were chosen to represent the minimum and maximum peak distal loads on the hip joint during walking for a 75 kg patient; i.e., 2.24 and 0.29 times body weight (Bergmann et al., 2001). The shear stress was calculated at the endosteal interface, and the implant motion relative to the femur was calculated at the most proximal cross section. Migration was defined as the vertical position of the proximal section of the stem relative to the bone while it was loaded (step 1), minus its initial position. Micromotion was defined as the vertical stem position relative to the bone while loaded (step 1), minus its position after it was unloaded (step 2). Micromotion and migration (δ) were each decomposed into: motion at the stem-cement and endosteum interfaces (δstem-cement and δendosteum, respectively); motion as a result of shear deformation of the cement mantle (δmantle); and motion due to shear deformation of the graft and/or graft/cement composite regions (δgraft/composite), see Appendix 4, Figure A4.2. In other words, the migration and micromotion were decomposed as:  δ = δ stem−cement + δ mantle + δ graft / composite + δ endosteum  (4.5)  Migration and micromotion were validated against in vitro experimental data from Chapter 2.  4.3 Results The cement penetration profile had a substantial effect on shear stresses at the endosteum (Figure 4.2). For profiles A-C, representing increasing uniform cement penetration into the graft, an average endosteum shear stress of 0.14 MPa was observed above the stem tip (Figure 4.2), where a gradual load transfer occurred from the stem to the femur. For these constructs, the endosteum shear stress peaked at 10 mm below the stem tip, where the load remaining on the distal stem was transferred sharply to the femur through the graft. For profiles D-I, representing increasing amounts of cement-endosteum contact, the load transfer occurred almost entirely through the cement-endosteum contact regions, where the average shear stress was 1.12, 0.52, 0.35, 0.26, 0.20, and 0.16 MPa, respectively. For these profiles, the shear stress was less than 0.05 MPa in the graft-endosteum interface regions. 81  Shear stress at endosteum (MPa)  0.30 0.25 Profile C Profile B Profile A  0.20 0.15 0.10 0.05 Stem tip  Distal  0.00 -50  0  Proximal  50  100  150  -0.05  y (mm) Shear stress at endosteum (MPa)  1.6 1.4  Profile D  1.2 1.0 Profile E  0.8  Profile F  0.6 Stem tip  Profile G Profile H  0.4  Profile I  0.2 0.0 -0.2  Proximal  Distal  -50  0  50  100  150  y (mm) Figure 4.2 Shear stress at endosteum interface for profiles A-C (top) and D-I (bottom). For profiles D-I, simulations in which the cement-endosteum contact region was bonded are shown in grey, and those in which the entire endosteum interface was sliding are shown in black.  For profiles A-C, where there was no cement-endosteum contact, the migration and micromotion were 578-3323 µm and 48-289 µm, respectively (Figure 4.3). For these profiles, the cement penetration depth (i.e., profile A, B, or C) had a substantial effect on the implant motion. For example, when comparing equivalent cases (i.e., having the same graft density and 82  status of the stem-cement interface), profile A resulted in much higher migration and micromotion (by 968-2071 µm 102-199, respectively) than did profile C. For profiles A-C, a low graft density resulted in much greater migration and micromotion (by 552-1649 µm and 36-132 µm, respectively) than did a high graft density.  Debonding of the stem-cement  interface, however, had little effect on implant motion, increasing migration and micromotion by only 74-89 µm (2-11%) and 5-6 µm, respectively. The motion results for profiles D-I, where there were varying amounts of cement-endosteum contact, are shown in Figure 4.4, together with the in vitro experimental results. For these profiles, migration and micromotion (5-356 µm and 4-37 µm, respectively) were much lower than for profiles A-C, and the motion decreased with increasing cement-endosteum contact. With profiles D-I, however, graft density had little effect on the implant motion – migration and micromotion differences of only 0-13 µm and 0-3 µm, respectively, were observed between the three graft density levels, with all other variables being equal. For these profiles, debonding of the stem-cement interface increased migration and micromotion by 80-170 µm and 4-7 µm, respectively. A breakdown of the contributions of each material and interface to total migration and micromotion (Equation 4.5) is presented in Table 4.2 and Appendix 4. Most of the migration occurred at the interfaces. For all cement profiles and graft densities, slippage at the stemcement interface accounted for 104-124 µm of the migration when this interface was defined as sliding. Slippage at the endosteum, however, was much more varied: it accounted for 0-2982 µm of the migration, and its contribution was largest when there was no cement-endosteum contact, in which case it decreased with increasing graft density. Shear deformation of the graft and graft/cement composite regions, however, accounted for only 42-262 µm (6-8%) of the migration for profiles A-C, and 1-24 µm of the migration for profiles D-I. For all profiles, micromotion occurred mainly within the graft and graft/cement composite regions and at the endosteum interface.  83  stem-cement interface sliding stem-cement interface bonded 3500  Migration (microns)  3000  Profile A  2500 2000  B  A  1500 1000 500  A  B  C  B  C low graft density  0 225  275  C  moderate  high  325  375  425  3  Graft density (mg/cm ) stem-cement interface sliding stem-cement interface bonded  Micromotion (microns)  350 300  Profile A  250 200  B  A  A  150  B 100 50 0 225  C  B  C  low graft density 275  C  moderate  high  325  375  425  3  Graft density (mg/cm )  Figure 4.3 Implant migration (top) and micromotion (bottom) for cement penetration profiles A-C. The grey data points represent constructs in which the stem-cement interface was bonded, while the black points represent those in which this interface was sliding, i.e., debonded.  84  FE (cement-endosteum sliding) experimental data (cycle 5000) 800  FE (cement-endosteum bonded) experimental data (cycle 10)  Migration (microns)  700 600  stem-cement sliding stem-cement bonded  Profile D  500 400  E  300  F  G  200  H  I  100 0 -100 0  20 40 60 80 Cement contact with endosteum (%)  FE (cement-endosteum sliding) experimental data  100  FE (cement-endosteum bonded)  70 Micromotion (microns)  60 50  Profile D  40  E  30  F  20  G  H  I  10 0 0  20 40 60 80 Cement contact with endosteum (%)  100  Figure 4.4 Implant migration (top) and micromotion (bottom) for profiles D-I. Simulation results are shown in black and grey, and experimental data are shown as white triangles (after 10 cycles of walking loads) and white squares (after 5000 cycles). Note that there are six migration data points in grey (cement-endosteum bonded) and six in black (cement-endosteum sliding) for each profile. These six points form two clusters: the top cluster represents constructs with a sliding stem-cement interface while the bottom one represents those with a bonded stem-cement interface. Each cluster contains three overlapping points, representing the three graft densities (low, moderate and high).  85  Table 4.2 Contribution of each material and interface to the distal implant motion relative to the bone. Micromotion (migration). Profile A B C D E F G H I  Interfaces δstem-cement (µm) 0-6 (0-130) 0-6 (0-123) 0-6 (0-120) 0-3 (0-124) 0-3 (0-119) 0-3 (0-116) 0-3 (0-115) 0-4 (0-117) 0-5 (0-114)  δendosteum (µm) 7-8 (1421-2982) 4-5 (996-1994) 8-9 (537-1096) 5-26 (32-221) 6-16 (17-121) 6-11 (12-83) 5-8 (8-56) 4-6 (5-48) 0-6 (0-35)  δmantle (µm) 2 (2) 2 (2) 2 (2) 0-1 (0-1) 0-1 (0-1) 0-1 (0- 1) 0-1 (0-1) 0-1 (0-1) 1-2 (0-1)  δgraft/composite (µm) 141-273 (127-262) 87-168 (82-159) 38-74 (42-82) 3-29 (3-24) 2-19 (2-16) 2-14 (2-12) 1-10 (1-10) 1-7 (1-8) 2-3 (2-3)  4.4 Discussion A wide range of stem migration has been reported in impaction allografting patients, and the reason for this variability has not yet been determined. In this study, a finite element model was used to estimate the effects of cement penetration profile and graft density on implant motion after impaction allografting, and to identify key mechanisms responsible for early implant migration. While graft density and the status of the stem-cement interface (bonded vs. sliding) had little effect on shear stresses at the endosteum, the cement penetration profile had a large effect. When there was cement-endosteum contact, very low stresses were observed in the graft regions and at the graft-endosteum interface, and virtually all load transfer to the femur occurred in the cement contact region (Figure 4.2, bottom). This is not surprising due to the higher stiffness of the cement compared to the graft. The average shear stress at the cementendosteum interface was inversely proportional to the area of cement contact. Shear stresses were also found to peak at the distal and/or proximal edges of the cement-endosteum interface. In this finite element model the transitions between cement and graft contact regions with the endosteum were very sharp, which can have a large effect on local stresses while having a small effect on the overall migration and micromotion. For this reason these local stresses should be interpreted in a qualitative rather than quantitative manner. In all constructs in which there is cement-endosteum contact, the radial compressive stresses in the graft-cement composite were highest in the proximal region. When the cement-endosteum interface was  86  sliding the shear stress at this interface was proportional to the normal stress (due to friction), thus peaking proximally. When the cement-endosteum interface was bonded, however, shear stress concentrations occurred at both ends of the of the cement contact region due to a shear lag effect, a phenomenon that is well-known in adhesively bonded joints.  Although bone cement does not bond chemically with bone, the cement-bone interface exhibits some strength through mechanical interlocking (Race et al., 2007). In impaction allografting, the apparent shear strength of the endosteal interface was found to be proportional to the amount of cement contact at that surface (Frei et al., 2005a). In the studied range of cement content, an apparent cement-endosteum shear strength of 0.8 MPa (0.25 to 1.25 MPa), average (range), has been reported (Frei et al., 2005a). Assuming that shear stresses in excess of 0.25 and 0.8 MPa indicate ‘possible’ and ‘likely’ cement-endosteum debonding, respectively, we can conclude that debonding of the cement-endosteum interface would be likely for Profile D (12% cement contact), and possible for profiles E-H (24-61% contact), particularly in the proximal and distal regions. For profile I (80% contact), however, this interface is likely to remain ‘bonded’. Migration and micromotion were highest in constructs in which the cement penetration front had not reached the endosteum. In these constructs, implant motion increased with decreasing graft density and increasing cement content. When there was cement-endosteum contact, however, implant motion was much lower and decreased with increasing contact. For these profiles, implant motion was not affected by graft density, which is not surprising because the load transfer occurred throught the cement while the graft carried little load. Finally, migration was attributed primarily to slippage at the endosteum, and this mechanism was responsible for most of the variability between constructs. Conversely, while shear deformation of the graft contributed to micromotion, shear deformation and failures within the graft had little effect on migration. A simplified geometry was used to facilitate the generation of multiple cement profiles. The model was axisymmetric and bending and/or torsional loading, which can increase implant motion, were not simulated. The loading was limited to one cycle and the model did not account for long-term processes such as creep and fatigue of the cement, and biological 87  processes such as graft incorporation or osteolysis, which could affect long term implant fixation. However, because a substantial part of the migration often occurs early after surgery; i.e., weeks to months (Nelissen et al., 2002; Ornstein et al., 2002), we incorporated the mechanisms of shear failure and compression of the graft, and slippage at the interfaces, which could induce early migration. Although the creep behaviour of the graft was not incorporated explicitly, creep was taken into account indirectly by using graft stiffness values equal to the inverse of the creep compliance (one hour under 1.1 MPa of confined compression) that were measured Chapter 3. A sensitivity analysis confirmed that the resulting stress and implant motion trends remained essentially the same when the graft was modeled as having half the graft stiffness (assuming further creep) or using shear failure parameters that were reported in another study (Dunlop et al., 2003). Nonetheless, despite the simplifying assumptions made in the model, there was reasonable agreement between model results and those of in vitro experiments. Migration and micromotion were in the same range as was measured experimentally, and in both cases the motion decreased with increasing cement-endosteum contact. In light of the simplifying assumptions made in this study, however, it should be noted that the trends observed are more important than the actual numerical results. A sensitivity analysis was also performed to assess how the current findings would be affected by changes in νgraft and the endosteum friction coefficient, µendosteum. The actual values of these parameters are not known, and these are likely to vary within each construct. In this model, estimates for these parameters were taken from the literature. More specifically, νgraft was assumed to be 0.2, a value obtained experimentally for unimpacted morsellized cancellous graft (Brodt et al., 1998). A µendosteum value of 0.61 was assumed, based on a value determined experimentally for a cancellous-cortical bone interface (Zhang et al., 1999). This value, however, was obtained for a smooth and dry surface, whereas the endosteum interface seen in impaction allografting is rougher and wet (with blood, fat and marrow). A sensitivity analysis was done, in which νgraft and µendosteum were varied between 0.2-0.4 and 0.61-1.0, respectively. It was found that the migration results were affected by changes in νgraft and µendosteum, however, the trends observed in this study – i.e., the relative effects of cement content and graft density on migration, the important role played by the endosteum interface, and the small  88  contribution of shear failures in the graft bed – remained the same regardless of νgraft and  µendosteum (Appendix 4, Figures A4.12 and A4.13). Previous research studies have been focused primarily on the graft rather than on the cement profile. Earlier findings have raised concerns about the risk of cement mantle fracture due to insufficient mantle thickness (Masterson et al., 1997a), which has prompted the re-design of impaction tools to allow for a wider gap between the stem and the graft bed. Over the last decade, however, very few studies have looked at the cement penetration profile (Frei et al., 2004; Frei et al., 2005a; Albert et al., 2007; Bolland et al., 2008). Clinically, it is difficult to distinguish the cement profile due to the radiographic similarity between the cement and the graft regions (Masterson 1997b; Robinson et al., 2002; Cabanela et al., 2003; Krupp et al., 2006). The use of cortical struts, wire meshes, and other hardware further obscures the radiographic observation of the graft and cement regions (Morgan et al., 2004). Moreover, in autopsy and biopsy studies, the cement was often dissolved during the histological process (Ling et al., 1993; Linder, 2000; Ullmark and Obrant, 2002). Most clinical reports, therefore, do not include cement mantle analysis. With limited research data concerning the cement penetration profile in impaction allografting, surgeons may overlook its relevance. While the presence of cement at the endosteum can hinder new bone formation in the graft (Frei et al., 2005b), some cement-endosteum contact may be unavoidable during surgery, particularly around the distal half of the stem (Frei et al., 2004; Albert et al., 2007). The depth of cement penetration in the graft bed has been shown to be proportional to both the cement pressure applied and the permeability of the graft bed, the latter of which is inversely proportional to its density (Frei et al., 2006). In a recent in-vitro study, the use of vibration and drainage holes in the impaction tool was found to improve stem stability in impaction allografting (Bolland et al., 2007), and this improvement was associated with an increase in graft density (Bolland et al., 2008). Contrary to other studies (Frei et al., 2004; Albert et al., 2007), the cement penetration profiles observed by Bolland et al. appeared to have little cement-endosteum contact. These differing morphological observations may be explained by differences in the impaction forces used. In one study (Bolland et al., 2007), the specimens were composite femurs with no cortical 89  thinning, enabling the use of very high impaction forces, 2.8-4.7 kN, and resulting in a very dense graft bed (65-95% bone). In the other studies, the specimens were human cadaveric femurs with simulated cortical thinning, enabling impaction forces of only 400-1200 N (Frei et al., 2004) and 200-300 N (Albert et al., 2007), and resulting in much lower graft densities (1860% bone, Albert et al., 2007). The results of the current study concur with Bolland’s observation that implant migration decreased with increasing graft density, but only when there was no cement-endosteum contact. However, since the risk of fracture is greater in the presence of extensive bone loss, it may not be possible to achieve a sufficiently dense graft bed to prevent cement from reaching the endosteum when the bone loss is moderate or severe. A certain amount of cement contact with the endosteum may thus be unavoidable in severely deficient femurs. Nonetheless, the current results indicate that some cement contact (e.g., 12% or more) may be beneficial to stem stability, particularly if post-operative adhesion of the peripheral graft to the endosteum is not achieved early. In conclusion, the cement penetration profile has a considerable effect on implant motion. Migration and micromotion decrease with increasing cement penetration into the graft bed, and some cement contact with the endosteum is beneficial for stability. Without cement-endosteum contact, implant motion increases with decreasing graft density, but the effect of graft density is negligible in constructs in which the cement profile reaches the endosteum. The results of this study emphasize that the cement penetration profile may have thus far been widely overlooked in studies of femoral impaction allografting. The results of this study also provide valuable insight into the mechanisms responsible for excessive implant migration in some impaction allografting patients. Slippage at the endosteum interface was a major contributor to implant migration and was responsible for most of the variability between constructs – this observation indicates that early adhesion of the peripheral graft to the endosteum may be critical to implant stability. Shear failures in the graft, on the other hand, played little role in implant migration.  90  4.5 References Albert, C., Masri, B., Duncan, C., Oxland, T., and Fernlund, G. (2008) Impaction allografting – the effect of impaction force and alternative compaction methods on the mechanical characteristics of the graft. J Biomed Mater Res B Appl Biomater 87, 395-405. Albert, C., Patil, S., Frei, H., Masri, B., Duncan, C., Oxland, T., and Fernlund, G. (2007) Cement penetration and primary stability of the femoral component after impaction allografting. A biomechanical study in the cadaveric femur. J Bone Joint Surg Br 89, 962-70. Bavadekar, A., Cornu, O., Godts, B., Delloye, C., Van Tomme, J., and Banse, X. (2001) Stiffness and compactness of morselized grafts during impaction: an in vitro study with human femoral heads. Acta Orthop Scand 72, 470-6. Bergmann, G., Deuretzbacher, G., Heller, M., Graichen, F., Rohlmann, A., Strauss, J., and Duda, G.N. (2001) Hip contact forces and gait patterns from routine activities. J Biomech 34, 859-71. Bolland, B.J., New, A.M., Madabhushi, G., Oreffo, R.O., and Dunlop, D.G. (2008) The role of vibration and drainage in femoral impaction bone grafting. J Arthroplasty 23, 1157-64. Bolland, B.J., New, A.M., Madabhushi, S.P., Oreffo, R.O., and Dunlop, D.G. (2007) Vibration-assisted bone-graft compaction in impaction bone grafting of the femur. J Bone Joint Surg Br 89, 686-92. Bolland, B.J., Partridge, K., Tilley, S., New, A.M., Dunlop, D.G., and Oreffo, R.O. (2006) Biological and mechanical enhancement of impacted allograft seeded with human bone marrow stromal cells: potential clinical role in impaction bone grafting. Regen Med 1, 457-67. Bozic, K.J., Kurtz, S.M., Lau, E., Ong, K., Vail, T.P., and Berry, D.J. (2009) The epidemiology of revision total hip arthroplasty in the United States. J Bone Joint Surg Am 91, 128-33. Brewster, N.T., Gillespie, W.J., Howie, C.R., Madabhushi, S.P., Usmani, A.S., and Fairbairn, D.R. (1999) Mechanical considerations in impaction bone grafting. J Bone Joint Surg Br 81, 118-24. Brodt, M.D., Swan, C.C., and Brown, T.D. (1998) Mechanical behavior of human morselized cancellous bone in triaxial compression testing. J Orthop Res 16, 43-9. Butler, A.M., Morgan, D.A., Verheul, R., and Walsh, W.R. (2005) Mechanical properties of gamma irradiated morselized bone during compaction. Biomaterials 26, 6009-13. Cabanela, M.E., Trousdale, R.T., and Berry, D.J. (2003) Impacted cancellous graft plus cement in hip revision. Clin Orthop Relat Res 417, 175-82. Canadian Institute for Health Information. (2008) Canadian Joint Replacement Registry (CJRR) 2007 Annual Report — Hip and Knee Replacements in Canada. Ottawa: CIHI. Chandler, H., Clark, J., Murphy, S., McCarthy, J., Penenberg, B., Danylchuk, K., and Roehr, 91  B. (1994) Reconstruction of major segmental loss of the proximal femur in revision total hip arthroplasty. Clin Orthop Relat Res 298, 67-74. Cornu, O., Bavadekar, A., Godts, B., Van Tomme, J., Delloye, C., and Banse, X. (2003) Impaction bone grafting with freeze-dried irradiated bone. Part II. Changes in stiffness and compactness of morselized grafts: experiments in cadavers. Acta Orthop Scand 74, 553-8. Cornu, O., Libouton, X., Naets, B., Godts, B., Van Tomme, J., Delloye, C., and Banse, X. (2004a) Freeze-dried irradiated bone brittleness improves compactness in an impaction bone grafting model. Acta Orthop Scand 75, 309-14. Cornu, O., Manil, O., Godts, B., Naets, B., Van Tomme, J., Delloye, C., and Banse, X. (2004b) Neck fracture femoral heads for impaction bone grafting: evolution of stiffness and compactness during impaction of osteoarthrotic and neck-fracture femoral heads. Acta Orthop Scand 75, 303-8. Dunlop, D.G., Brewster, N.T., Madabhushi, S.P., Usmani, A.S., Pankaj, P., and Howie, C.R. (2003) Techniques to improve the shear strength of impacted bone graft: the effect of particle size and washing of the graft. J Bone Joint Surg Am 85-A, 639-46. Eldridge, J.D., Smith, E.J., Hubble, M.J., Whitehouse, S.L., and Learmonth, I.D. (1997) Massive early subsidence following femoral impaction grafting. J Arthroplasty 12, 535-40. Engelbrecht, D.J., Weber, F.A., Sweet, M.B., and Jakim, I. (1990) Long-term results of revision total hip arthroplasty. J Bone Joint Surg Br 72, 41-5. Fetzer, G.B., Callaghan, J.J., Templeton, J.E., Goetz, D.D., Sullivan, P.M., and Johnston, R.C. (2001) Impaction allografting with cement for extensive femoral bone loss in revision hip surgery: a 4- to 8-year follow-up study. J Arthroplasty 16, 195-202. Fosse, L., Muller, S., Ronningen, H., Irgens, F., and Benum, P. (2006) Viscoelastic modelling of impacted morsellised bone accurately describes unloading behaviour: an experimental study of stiffness moduli and recoil properties. J Biomech 39, 2295-302. Fosse, L., Ronningen, H., Lund-Larsen, J., Benum, P., and Grande, L. (2004) Impacted bone stiffness measured during construction of morsellised bone samples. J Biomech 37, 1757-66. Franzen, H., Toksvig-Larsen, S., Lidgren, L., and Onnerfalt, R. (1995) Early migration of femoral components revised with impacted cancellous allografts and cement. A preliminary report of five patients. J Bone Joint Surg Br 77, 862-4. Frei, H., Gadala, M.S., Masri, B.A., Duncan, C.P., and Oxland, T.R. (2006) Cement flow during impaction allografting: a finite element analysis. J Biomech 39, 493-502. Frei, H., Mitchell, P., Masri, B.A., Duncan, C.P., and Oxland, T.R. (2004) Allograft impaction and cement penetration after revision hip replacement. A histomorphometric analysis in the cadaver femur. J Bone Joint Surg Br 86, 771-6. Frei, H., Mitchell, P., Masri, B.A., Duncan, C.P., and Oxland, T.R. (2005a) Mechanical 92  characteristics of the bone-graft-cement interface after impaction allografting. J Orthop Res 23, 9-17. Frei, H., O'Connell, J., Masri, B.A., Duncan, C.P., and Oxland, T.R. (2005b) Biological and mechanical changes of the bone graft-cement interface after impaction allografting. J Orthop Res 23, 1271-9. Giesen, E.B., Lamerigts, N.M., Verdonschot, N., Buma, P., Schreurs, B.W., and Huiskes, R. (1999) Mechanical characteristics of impacted morsellised bone grafts used in revision of total hip arthroplasty. J Bone Joint Surg Br 81, 1052-7. Gokhale, S., Soliman, A., Dantas, J.P., Richardson, J.B., Cook, F., Kuiper, J.H., and Jones, P. (2005) Variables affecting initial stability of impaction grafting for hip revision. Clin Orthop Relat Res 174-80. Guild, F.J., Young, R.J., and Lovell, P.A. (1994) The influence of material properties on the predicted behaviour of rubber-toughened polymers. J Mater Sci Lett 13, 10-14. Haddad, F.S. and Duncan, C.P. (1999) Impaction allografting of the proximal femur: fact or fad? Orthopedics 22, 855-8. Halliday, B.R., English, H.W., Timperley, A.J., Gie, G.A., and Ling, R.S. (2003) Femoral impaction grafting with cement in revision total hip replacement. Evolution of the technique and results. J Bone Joint Surg Br 85, 809-17. Karrholm, J., Garellick, G., and Herberts, P. (2007) Swedish Hip Arthroplasty Register Annual report 2006. Department of Orthopaedics, Sahlgrenska University Hospital. Kindt-Larsen, T., Smith, D.B., and Jensen, J.S. (1995) Innovations in acrylic bone cement and application equipment. J Appl Biomater 6, 75-83. Knight, J.L. and Helming, C. (2000) Collarless polished tapered impaction grafting of the femur during revision total hip arthroplasty: pitfalls of the surgical technique and follow-up in 31 cases. J Arthroplasty 15, 159-65. Krupp, R.J., Malkani, A.L., Crawford, C.H., Capello, W.N., Feinberg, J.R., and Baker, D. (2006) Impaction grafting for femoral component revision using a non-polished bead-blasted chrome cobalt stem-average 8 1/2-year follow-up. J Arthroplasty 21, 1180-6. Lind, M., Krarup, N., Mikkelsen, S., and Horlyck, E. (2002) Exchange impaction allografting for femoral revision hip arthroplasty: results in 87 cases after 3.6 years' follow-up. J Arthroplasty 17, 158-64. Linder, L. (2000) Cancellous impaction grafting in the human femur: histological and radiographic observations in 6 autopsy femurs and 8 biopsies. Acta Orthop Scand 71, 543-52. Ling, R.S., Timperley, A.J., and Linder, L. (1993) Histology of cancellous impaction grafting in the femur. A case report. J Bone Joint Surg Br 75, 693-6.  93  Malkani, A.L., Voor, M.J., Fee, K.A., and Bates, C.S. (1996) Femoral component revision using impacted morsellised cancellous graft. A biomechanical study of implant stability. J Bone Joint Surg Br 78, 973-8. Masterson, E.L., Masri, B.A., and Duncan, C.P. (1997a) The cement mantle in the Exeter impaction allografting technique. A cause for concern. J Arthroplasty 12, 759-64. Masterson, E.L., Masri, B.A., Duncan, C.P., Rosenberg, A., Cabanela, M., and Gross, M. (1997b) The cement mantle in femoral impaction allografting. A comparison of three systems from four centres. J Bone Joint Surg Br 79, 908-13. Meding, J.B., Ritter, M.A., Keating, E.M., and Faris, P.M. (1997) Impaction bone-grafting before insertion of a femoral stem with cement in revision total hip arthroplasty. A minimum two-year follow-up study. J Bone Joint Surg Am 79, 1834-41. Morgan, H.D., McCallister, W., Cho, M.S., Casnellie, M.T., and Leopold, S.S. (2004) Impaction allografting for femoral component revision: clinical update. Clin Orthop Relat Res 420, 160-8. Nelissen, R.G., Valstar, E.R., Poll, R.G., Garling, E.H., and Brand, R. (2002) Factors associated with excessive migration in bone impaction hip revision surgery: a radiostereometric analysis study. J Arthroplasty 17, 826-33. Nilsson, K.G. and Karrholm, J. (1996) RSA in the assessment of aseptic loosening. J Bone Joint Surg Br 78, 1-3. Nuno, N., Amabili, M., Groppetti, R., and Rossi, A. (2002) Static coefficient of friction between Ti-6Al-4V and PMMA for cemented hip and knee implants. J Biomed Mater Res 59, 191-200. Nuno, N., Groppetti, R., and Senin, N. (2006) Static coefficient of friction between stainless steel and PMMA used in cemented hip and knee implants. Clin Biomech (Bristol, Avon) 21, 956-62. Ornstein, E., Atroshi, I., Franzen, H., Johnsson, R., Sandquist, P., and Sundberg, M. (2002) Early complications after one hundred and forty-four consecutive hip revisions with impacted morselized allograft bone and cement. J Bone Joint Surg Am 84-A, 1323-8. Ornstein, E., Franzen, H., Johnsson, R., Karlsson, M.K., Linder, L., and Sundberg, M. (2004) Hip revision using the Exeter stem, impacted morselized allograft bone and cement: a consecutive 5-year radiostereometric and radiographic study in 15 hips. Acta Orthop Scand 75, 533-43. Pekkarinen, J., Alho, A., Lepisto, J., Ylikoski, M., Ylinen, P., and Paavilainen, T. (2000) Impaction bone grafting in revision hip surgery. A high incidence of complications. J Bone Joint Surg Br 82, 103-7. Phillips, A., Pankaj, P., May, F., Taylor, K., Howie, C., and Usmani, A. (2006a) Constitutive models for impacted morsellised cortico-cancellous bone. Biomaterials 27, 2162-70. 94  Phillips, A.T., Pankaj, Brown, D.T., Oram, T.Z., Howie, C.R., and Usmani, A.S. (2006b) The elastic properties of morsellised cortico-cancellous bone graft are dependent on its prior loading. J Biomech 39, 1517-26. Piccaluga, F., Gonzalez Della Valle, A., Encinas Fernandez, J.C., and Pusso, R. (2002) Revision of the femoral prosthesis with impaction allografting and a Charnley stem. A 2- to 12-year follow-up. J Bone Joint Surg Br 84, 544-9. Pidaparti, R.M. and Vogt, A. (2002) Experimental investigation of Poisson's ratio as a damage parameter for bone fatigue. J Biomed Mater Res 59, 282-7. Race, A., Mann, K.A., and Edidin, A.A. (2007) Mechanics of bone/PMMA composite structures: an in vitro study of human vertebrae. J Biomech 40, 1002-10. Ratner, B.D., Hoffman, A.S., Schoen, F.J., and Lemons, J.E. (2004) In Biomaterials science An introduction to materials in medicine Pp. 535. Elsevier Academic Press. Reilly, D.T. and Burnstein, A.H. (1975) The elastic and ultimate properties of compact bone tissue. J Biomech 8, 393-405. Robinson, D.E., Lee, M.B., Smith, E.J., and Learmonth, I.D. (2002) Femoral impaction grafting in revision hip arthroplasty with irradiated bone. J Arthroplasty 17, 834-40. Schreurs, B.W., Arts, J.J., Verdonschot, N., Buma, P., Slooff, T.J., and Gardeniers, J.W. (2005) Femoral component revision with use of impaction bone-grafting and a cemented polished stem. J Bone Joint Surg Am 87, 2499-507. Schreurs, B.W., Slooff, T.J., Buma, P., and Verdonschot, N. (2001) Basic science of bone impaction grafting. Instr Course Lect 50, 211-20. Tanabe, Y., Wakui, T., Kobayashi, A., Ohashi, H., Kadoya, Y., and Yamano, Y. (1999) Determination of mechanical properties of impacted human morsellized cancellous allografts for revision joint arthroplasty. J Mater Sci Mater Med 10, 755-60. Ullmark, G. and Obrant, K.J. (2002) Histology of impacted bone-graft incorporation. J Arthroplasty 17, 150-7. van Biezen, F.C., ten Have, B.L., and Verhaar, J.A. (2000) Impaction bone-grafting of severely defective femora in revision total hip surgery: 21 hips followed for 41-85 months. Acta Orthop Scand 71, 135-42. Voor, M.J., Nawab, A., Malkani, A.L., and Ullrich, C.R. (2000) Mechanical properties of compacted morselized cancellous bone graft using one-dimensional consolidation testing. J Biomech 33, 1683-8. Voor, M.J., White, J.E., Grieshaber, J.E., Malkani, A.L., and Ullrich, C.R. (2004) Impacted morselized cancellous bone: mechanical effects of defatting and augmentation with fine hydroxyapatite particles. J Biomech 37, 1233-9.  95  Zhang, Y., Ahn, P.B., Fitzpatrick, D.C., Heiner, A.D., Poggie, R.A., and Brown, T.D. (1999) Interfacial frictional behaviour: cancellous bone, cortical bone, and a novel porous tantalum biomaterial. Journal of Musculoskeletal Research 3, 425-251.  96  CHAPTER 5 GENERAL DISCUSSION AND CONTRIBUTIONS  97  5.1 Discussion Failure of femoral implants in THA is often accompanied by problematic bone loss in the proximal femur. Impaction allografting has the potential to restore bone stock in deficient femurs. However, femoral fractures are common with this procedure, as are high levels of implant subsidence which can cause thigh pain and hip dislocation. A number of mechanisms have been proposed to contribute to subsidence, but their relative impact on implant stability is not known. Much research has focused on improving the stiffness and shear strength of the graft, however, it is not known how much effect postoperative graft compression and shear failure have on implant subsidence. Moreover, in a recent in vitro study of impaction allografting, a substantial portion of the graft bed was found to be saturated with bone cement, and some cement was found to have reached the host bone. The consequence of such extensive cement penetration on implant stability has not been determined. This thesis examined the effects of cement profile and graft properties on initial stem stability in revision THA with impaction allografting. Key mechanisms that contribute to early implant subsidence were identified. The results of this study provide valuable information that will help surgeons, scientists, and engineers to develop strategies to minimize the risk of problematic subsidence in femoral impaction allografting.  5.1.1 Cement penetration profile Limited information is available about the cement morphology within the graft bed in impaction allografting patients. As mentioned earlier, bone cement is not clearly distinguishable from the graft radiographically (Robinson et al., 2002; Krupp et al., 2006), and histological studies often use methods that dissolve the cement (Ling et al., 1993; Linder, 2000; Ullmark and Obrant, 2002). In this study, extensive cement-endosteum contact was seen around the distal half of the stem, and the same was true in a previous cadaveric femur study (Frei et al., 2004). In these two studies, the surgical procedures were performed by three surgeons using two different sets of graft impaction tools: X-change (Howmedica Inc, Rutherford, NJ, USA) and CPT (Zimmer Inc., Warsaw, IN, USA). Cement-endosteum contact was reported in two other in vitro studies (Berzins et al., 1996; Ohashi et al., 2009). While no clinical studies of femoral impaction allografting were found that mention cement-endosteum contact explicitly, published radiographs appear to confirm regions of cement contact with the 98  host bone in some patients. In personal communications with surgeons from Vancouver General Hospital, Dr. Bassam Masri and Dr. Clive Duncan, the surgeons were asked to describe the cement profiles in a selection of radiographs taken from the clinical literature – both surgeons agreed that cement-endosteum contact appeared to be present in some of the radiographs, usually around the distal half of the stem, but that no cement contact was apparent in others. The surgeons, however, cautioned that the true cement profiles could not be inferred definitively from the radiographs. Nonetheless, although the actual cement profile is essentially impossible to confirm in individual impaction allografting patients, a wide range of cement penetration appears to occur.  5.1.2 Effects of cement profile and graft density on initial implant subsidence The results of the current study demonstrate that cement penetration profile has a sizeable effect on implant motion. In vitro stem subsidence correlated negatively with the amounts of cement in the construct and cement-endosteum contact (Chapter 2). These results were corroborated by those of the finite element model (Chapter 4), in which simulations without cement-endosteum contact resulted in substantially greater migration than did those with cement contact, and the implant motion decreased with increasing cement contact. Cement contact with host bone has not received much attention in the literature and could explain, at least in part, the variability in early implant subsidence seen between impaction allografting patients. Achieving a dense graft bed during impaction is believed essential for initial stem stability (Franzen et al., 1995; Knight and Helming, 2000; Pekkarinen et al., 2000; Gore, 2002; Halliday et al., 2003; Morgan et al., 2004; Krupp et al., 2006). In a clinical study, subsidence was reported to correlate negatively with graft density around the stem tip (Gokhale et al., 2005). A relationship between subsidence and graft density was also reported in an in vitro study (Bolland et al., 2007; Bolland et al., 2008b). However, no such relationship was found in other clinical studies (Ornstein, 2002, Nelissen et al., 2002). These seemingly contradictory findings can be explained by the results of the current study, in which the relationship between graft density and subsidence was demonstrated to be dependent upon the cement penetration profile. Without cement-endosteum contact, implant motion decreases with increasing graft 99  density, however, graft density has little effect on implant motion in the presence of cementendosteum contact (Chapter 4 and Figure 5.1). The apparent relationship between subsidence and graft density around the stem tip reported in Gokhale’s study indicates that their cohort of patients may have had minimal cement-endosteum contact, or that the apparence of dense graft may have been the result of extensive cement penetration in that region. The relationship between graft density and in vitro migration observed by Bolland et al. can be explained by an apparent lack of cement-endosteum contact in their specimens. In Bolland’s study, very high impaction forces were used, generating high graft densities that may not be achievable in many patients. Impaction allografting is commonly performed in patients with considerable femoral bone stock deficiencies, for whom the use of such high impaction forces is likely to result in iatrogenic fracture. Furthermore, because the depth of cement penetration in the graft increases with decreasing graft density (Frei et al., 2006), the likelihood of producing regions of cementendosteum contact is probably very high in patients with moderate and extensive bone stock deficiencies. For this reason, the effect of cement profile on stem stability probably explains why graft density and the extent of bone stock deficiency have not been found to correlate consistently with subsidence (Meding et al., 1997; van Doorn et al., 2002; Nelissen et al., 2002; Ornstein, 2002; Halliday et al., 2003; Ornstein et al., 2004; Gokhale et al., 2005; Hassaballa et al., 2009).  70  Distal micromotion (µm)  Distal migration (µm)  1000 800 600 400 R2 = 0.0012  200 0  60 50 40 30  R2 = 0.0006  20 10 0  -200 28  30  32  34  36  38  Average graft density (%bone)  40  28  30  32  34  36  38  40  42  Average graft density (%bone)  Figure 5.1 Relationship between distal migration (left) and micromotion (right), and graft density after impaction allografting in cadaveric femurs (Chapter 2). Lines represent linear regressions.  In conclusion, the current study demonstrates that the risk of excessive subsidence is highest in constructs with low cement content and low graft density, and that achieving a dense graft bed is helpful for stability, but only when there is no cement-endosteum contact. 100  5.1.3 Subsidence mechanisms The clinical literature reveals a wide range of implant subsidence after impaction allografting, and several attempts have been made to identify the clinical factors and the mechanisms that play a role in its development. There is radiographic evidence that longitudinal cement fracture can cause excessive subsidence (Masterson and Duncan, 1997; Masterson et al., 1997a; Masterson et al., 1997b). Since these studies were published, graft impaction tools have been redesigned to address this problem by ensuring a thicker cement mantle. Nonetheless, high levels of subsidence have continued to occur in some patients (Nelissen et al., 2002; Gokhale et al., 2005; Deakin and Bannister, 2007; Sierra et al., 2008). Therefore, other mechanisms are likely to contribute to subsidence, namely: cement creep/settling of the stem within the cement, graft compression, shear failures within the graft, sliding at the endosteum interface, allograft resorption, and expansion of the femoral canal. The relative impact of each mechanism on stem subsidence is not known. In most clinical studies, subsidence is measured from plain radiographs (e.g., Gie et al., 1993; Eldridge et al., 1997; Masterson et al., 1997a; Knight and Helming, 2000; Boldt et al., 2001; Gore, 2002; Halliday et al., 2003; Gokhale et al., 2005; Deakin and Bannister, 2007; Sierra et al., 2008; Hassaballa et al., 2009). With this method, subsidence less than 4 mm cannot be measured reliably (Brand et al., 1986), and the subsidence reported in impaction allografting patients is usually less than that. It is therefore no surprise that no clinical factors have yet been identified that correlate consistently with subsidence. A more accurate measurement of implant motion is obtained with roentgen stereophotogrammetric analysis (RSA). In RSA, two radiographs are taken simultaneously from different angles, allowing the calculation of the three-dimensional stem displacement relative to the femur by using implanted metal beads as reference points. The estimated error when measuring distal migration using RSA is 0.2-0.3 mm (Karrholm et al., 1999). However, as most centres do not have the necessary equipment, few RSA studies on impaction allografting patients have been published (Karrholm et al., 1999; Ornstein et al., 2000; van Doorn et al., 2002; Nelissen et al., 2002; Ornstein et al., 2003; Ornstein et al., 2004). Nonetheless, from these studies we have learned that a large part of the implant migration often develops early after surgery. For example, 0.4-2 mm can occur within 101  the first week, i.e., approximately half of that seen in the same patients at six months (Ornstein et al., 2000). We have also learned that unstable implants are often distinguishable from stable ones within the first three months (Nelissen et al., 2002; Ornstein et al., 2004). These reports therefore highlight the relevance of initial implant subsidence – and the mechanisms that affect it – on long-term stability. Some settling of the stem within the cement is expected when using smooth-tapered stems in impaction allografting (Gie et al., 1993). The observation of high levels of subsidence in some patients (e.g., Eldridge et al., 1997; Meding et al., 1997), however, has led some surgeons to substitute the popular smooth-tapered stem designs with rough-collared ones (Leopold et al., 1999; Karrholm et al., 1999; Fetzer et al., 2001; de Roeck and Drabu, 2001; Boldt et al., 2001; Ullmark and Obrant, 2002; Ullmark et al., 2002; Piccaluga et al., 2002; van Kleunen et al., 2003; Sorensen et al., 2003; Krupp et al., 2006; Van Kleunen et al., 2006; Nich and Sedel, 2006; Ullmark et al., 2007). Migration reported with rough-collared stems has ranged between 0 and 8 mm (e.g., Boldt et al., 2001; Nich and Sedel, 2006), compared with 0 to 80 mm for smooth-tapered stems (e.g., Sierra et al., 2008). It is not clear whether rough stems are truly superior to smooth ones, because although very high subsidence can occur with smooth stems, the same stems are stable with less than 2 mm subsidence in many patients (Nelissen et al., 2002). These studies, however, support the assumption that settling of the stem within the cement accounts for some but not all of the subsidence observed clinically after impaction allografting. In our finite element simulations (Chapter 4), regardless of cement profile or graft density, sliding at the stem/cement interface accounted for roughly 100 µm (104 to 124 µm) of the implant migration when this interface was debonded (Appendix 4). This contribution represented 30-96% of total migration for profiles D-H, but only 4-17% of the large migration seen in profiles A-C. This mechanism therefore did not explain the wide range of migration observed. Finally, while our model did not incorporate cement creep, it was demonstrated in a previous finite element study that subsidence of polished stems due to creep in primary hip arthroplasty did not exceed 50 µm (Verdonschot and Huiskes, 1997). Therefore, creep of the cement does not appear to be a major contributor to stem subsidence in impaction allografting. Slippage between graft particles due to shear stresses, i.e., the formation of shear failure planes in the graft bed, has also been speculated to contribute to subsidence (Ornstein et al., 2001; 102  Halliday et al., 2003; Ornstein et al., 2004; Bolland et al., 2007; Bolland et al., 2008b). The results of the current study, however, do not point to this mechanism as a major contributor. Instead, our results emphasize the importance of the endosteum interface in the development of subsidence. In vitro migration correlated negatively with the amount of cement at the endosteum (Table 2.5). In our finite element simulations, slippage at the endosteum accounted for a substantial portion of the implant migration, and this slippage decreased with increasing cement contact (Appendix 4). Without cement-endosteum contact, migration was almost entirely attributed to slippage at the endosteum interface, and this slippage decreased with increasing graft density (Figures A4.3 to A4.5). Hence, not only was the contribution of slippage at the endosteum substantial, it also accounted for a large portion of the migration variability between simulations. Slippage at the endosteum is accompanied by some radial and longitudinal compression of the graft bed due to the tapered and confined geometry of the intramedullary canal. Our model simulated creep deformation of the graft indirectly by using stiffness values equal to the inverse of the creep compliance obtained after one hour of loading. We can presume that any additional creep occurring postoperatively in the graft would further contribute to the amount of slippage at the endosteum interface. The portion of the implant migration that was attributed to shear deformation and slippage within the graft bed, on the other hand, was much smaller. In summary, for all graft densities and cement penetration profiles, migration was attributed primarily to slippage at the stem-cement and the endosteal interfaces (Figure 5.2), and the latter was responsible for most of the variability between constructs.  103  Figure 5.2 Major mechanisms of subsidence for femoral impaction allografting constructs without (left) and with (right) cement-endosteum contact.  The small effect of shear failure within the graft upon subsidence is further confirmed when comparing the motion obtained by including, versus without including, graft shear failure in our finite element model. An additional simulation was performed that replicated one of the simulations from Chapter 4, but this time without defining a graft shear failure criterion. The resulting micromotion was unaffected by shear failures while the migration was only 6% lower than that obtained when the shear failure criterion was modeled (Figure 5.3). Shear failure in the graft bed, therefore, does not appear to be an important contributor to early subsidence.  104  stem-cement interface graft/graft-cement composite  cement mantle endosteum interface  3500  Migration (microns)  3000 2500 2000 1500 1000 500 0 Shear failure stem-cement interface graft/graft-cement composite  No shear failure cement mantle endosteum interface  350  Micromotion (microns)  300 250 200 150 100 50 0 Shear failure  No shear failure  Figure 5.3 Comparison of migration (top) and micromotion (bottom) finite element results when the graft shear failure criterion was included in the model versus when it was not, for cement profile A (Figure 4.1), with a low graft density and when the stemcement interface was defined as sliding.  Finally, we can use the knowledge gained in Chapter 4 to speculate about mechanisms that took place during implant migration in each cadaveric specimens in Chapter 2 (Figure 2.5). In Chapter 4, micromotion was found to result from shear deformation of the graft and cement regions and slippage at the endosteum, while large migrations were attributed primarily to 105  slippage at the endosteum. Three of the four specimens with the highest migration in Chapter 2 experienced decreasing micromotion as the implant migrated. In these three specimens, the migration and micromotion trends are consistent with the assumption that the large migration was attributed to slippage at the endosteum, which caused a stiffening of the graft as it became increasingly compressed. In the eight specimens with the lowest migration, however, the micromotion was stable – it did not increase or decrease during the simulated walking load cycles. In these specimens, the migration due to endosteal slippage appears to have been too small to have a pronounced effect on graft density or stiffness. Finally, only one specimen did not fit the abovementioned trends. In the specimen with the third highest migration (264 µm), micromotion actually increased during migration (from 31 to 42 µm), indicating that the graft did not become denser. Migration in this specimen may therefore have been attributed to mechanisms such as a gradual debonding of the stem cement interface and/or slippage between the distal plug and the femur rather than endosteal slippage. In summary, the results of the current study indicate that early implant stability relies largely on the endosteum. Excessive implant migration may be avoided if early post-operative adhesion of the graft onto the endosteum surface is achieved. There is clinical evidence that early graft adhesion to the endosteum is possible after impaction allografting (Ullmark and Obrant, 2002; Leopold et al., 1999); however, this does not occur consistently (Ullmark et al., 2002; personal communications with Dr. Clive Duncan). Without graft adhesion to the endosteum, the presence of cement-endosteum contact may be necessary to prevent excessive migration, nonetheless compromising the potential for new bone formation. Our results further indicate that high levels of cement penetration into the graft may be unavoidable in some impaction allografting patients. A lesser extent of cement penetration may be obtained if a very dense graft bed is achieved, however, this may only be possible in less defective (i.e., stronger) femurs. In moderately or severely defective femurs, the higher risk of fracture may preclude the use of high impaction forces, and extensive cement penetration may be unavoidable in some regions. Thus, patients with less defective femurs may have a better potential for new bone formation than those with severe bone loss. Finally, the development of methods to ensure early graft adhesion to the endosteum would be useful in minimizing the risk of excessive migration.  106  5.1.4 Postoperative biological changes The process of graft incorporation is difficult to evaluate clinically. There is no widely accepted guideline on how to infer graft incorporation from radiographs in femoral impaction allografting. Interpretation of the radiographic appearance of the graft bed is imprecise due to the difficulty in distinguishing the graft from the cement (Duncan et al., 1998; Mikhail et al., 1999; Lind et al., 2002). Furthermore, the presence of external hardware such as reinforcement wires, meshes, plates, or dense cortical graft struts that are often required to prevent or repair fractures only further obscures the appearance of the graft (Elting et al., 1995; van Biezen et al., 2000; Ullmark et al., 2002; Schreurs et al., 2005). Many clinical studies, for that matter, simply do not comment on the appearance of the graft bed (Eldridge et al., 1997; Masterson et al., 1997b; Pekkarinen et al., 2000; Knight and Helming, 2000; Ornstein et al., 2001; Fetzer et al., 2001; Ornstein et al., 2002; Gore, 2002; Atroshi et al., 2004; Ornstein et al., 2004; Mahoney et al., 2005; Sierra et al., 2008). Nonetheless, radiographic evidence of cortical repair and/or cancellous remodeling has been reported in some patients (Franzen et al., 1995; Elting et al., 1995; Flugsrud et al., 2000; Boldt et al., 2001; Deakin and Bannister, 2007), and changes that are consistent with graft incorporation, i.e., changes in radiopacity, have been noted in others (Meding et al., 1997; Leopold et al., 1999; Lind et al., 2002). Histological studies have indeed revealed incorporation, at least in part, of the dead graft (Ling et al., 1993; Nelissen et al., 1995; Ullmark and Linder, 1998; Mikhail et al., 1999; Linder, 2000; Weidenhielm et al., 2001; Ullmark and Obrant, 2002). Early autopsy and biopsy case reports described three zones in the graft bed: an outer zone consisting of viable cortical bone, presumably regenerated from peripheral graft; a middle zone consisting of viable trabecular bone and cement; and an inner zone consisting of partly necrotic graft and cement, with occasional regions of fibrous tissue and viable bone (Ling et al., 1993; Nelissen et al., 1995; Mikhail et al., 1999). A larger study with 14 autopsy and biopsy specimens was later published that revealed that graft incorporation is highly variable (Linder, 2000). While one patient showed complete bony reconstitution, variable amounts of dead graft remained in the others even after eight years. Cortical healing, trabecular incorporation, and/or trabecular remodeling were observed only in some femurs. The author commented that the graft appeared to have been first invaded by fibrovascular tissue from the periphery, eventually embedding the graft particles, but that the vascular front did not always reach the cement surface, sometimes 107  leaving many millimeters of avascular dead graft adjacent to the cement. In another biopsy study, fibrous tissue invasion was observed as early as one month post-surgery, and many graft particles had layers of living bone at four months (Ullmark and Obrant, 2002). Finally, graft incorporation in the distal region was reported to be more extensive than in the proximal region in three autopsies performed after six months to three years (Ullmark and Linder, 1998; Weidenhielm et al., 2001; Ullmark and Obrant, 2002). The appearance of the graft was also described during a few re-revision surgeries. During one re-revision performed three weeks after impaction allografting due to a fracture, the inner surface of the femur was said to have changed from “smooth” to having a “pronounced rough appearance, resulting from firmly attached small graft pieces” (Ullmark and Obrant, 2002). In three other re-revisions at three to six years, the graft was described as “bleeding and healed to the cortex” (Leopold et al., 1999), and as “soundly incorporated” (Deakin and Bannister, 2007). Nonetheless, this is certainly not always true. In one re-revision at five years, the graft was described as partly necrotic without ingrowth of fibrous tissue or bone in most areas (Ullmark et al., 2002). Moreover, in the few femoral impaction allografting patients who have been re-revised at Vancouver General Hospital, the graft bed was not found to have incorporated; it essentially disintegrated during the re-revision procedure (personal communication with Dr. Clive Duncan). Some recent studies have used nuclear imaging methods to monitor blood flow and bone turnover after impaction allografting (Boldt et al., 2001; Sorensen et al., 2003; van Kleunen et al., 2003; Hisatome et al., 2004; Van Kleunen et al., 2006; Ullmark et al., 2007; Temmerman et al., 2008). These methods involve the injection of a radioactive blood tracer from which the emitted gamma rays are captured by cameras. Increases in blood flow and tracer uptake, indicating revascularization and graft incorporation, were observed in and around graft regions (Boldt et al., 2001; Sorensen et al., 2003; van Kleunen et al., 2003; Hisatome et al., 2004; Van Kleunen et al., 2006; Ullmark et al., 2007; Temmerman et al., 2008). In areas adjacent to the graft, the increase in blood flow and tracer uptake was seen within two weeks after surgery, indicating a swift onset of neovascularization and osteoblastic activity (Sorensen et al., 2003; Temmerman et al., 2008). The uptake was noted to decrease significantly within a year (Hisatome et al., 2004), however, at four to six years it remained elevated in the trochanters 108  (Hisatome et al., 2004) and around the stem tip (Hisatome et al., 2004; Ullmark et al., 2007), indicating that remodeling was probably still incomplete in those regions. In short, without invasive biopsies, clinical evaluation of the graft is difficult, and therefore the actual status of the graft incorporation remains unknown in most impaction allografting patients. However, these studies demonstrate collectively that graft incorporation is variable, usually incomplete, and often limited to the distal region and the periphery. Nonetheless, it is apparent that some peripheral graft incorporation can take place as early as within a few weeks. The effect of the graft incorporation process on implant migration is not clear. Structural cortical grafts have been shown to lose strength during the early stages of their incorporation due to resorption-induced porosity (Enneking et al., 1975). Their density and strength eventually return to normal as new bone gradually forms. Cancellous structural grafts see a reverse effect gaining density as new bone apposes onto trabeculae, without experiencing initial resorption (Abbott et al., 1947; Burchardt, 1983). For morsellized cancellous graft, however, it is not known how the mechanical properties evolve during incorporation. Nonetheless, the biological changes occurring in the graft after surgery could be the cause of some implant subsidence. In an in vivo experiment in goats, cancellous graft particles implanted in a defect in the distal femur were subjected to loading through a piston, which was reported to have moved by 2 mm at two weeks in two out of four goats (Lamerigts et al., 2000). The same experiment was repeated in three femurs ex vivo, but no piston displacement was reported after 198,000 cycles of loading. Thus, biological changes in the graft may be significant enough to cause some implant subsidence. Furthermore, in addition to the changes occurring within the graft bed, the host bone can also experience changes postoperatively. Cancellization of the cortex has been observed following the implantation of graft particles and cement against the inner cortical surface of rat tibiae (Frei et al., 2005). It was suggested that the formation of pores in the cortex may be the result of impaired vascular circulation, and that some subsidence of the stem and cement/graft conglomerate relative to the host bone after impaction allografting may be caused by a widening of the femoral canal due to resorption of the host bone.  109  Although our in vitro and numerical studies did not model biological processes that take place in the graft post-surgery, our results indicate that postoperative bonding of the peripheral graft to the endosteum may be a critical factor to implant stability. As discussed earlier, in our finite element results, substantial migration occurred at the graft-endosteum interface. From the results presented in Appendix 4, we can estimate that if peripheral graft bonding to the endosteum is successful and prompt after surgery, the migration may be reduced by over 75% if there is no cement-endosteum contact. It has been demonstrated that fibrous invasion of the graft increases the unconfined compression strength of morsellized cancellous graft (Tagil and Aspenberg, 2001), as well as the shear strength of the graft/host-bone interface (Frei et al., 2005). From a mechanical viewpoint, this fibrous invasion may provide sufficient anchoring of the peripheral graft to prevent excessive subsidence. Nonetheless, unless the fibrous tissue is eventually replaced with new bone, it would likely be removed during any future re-revision. Finally, the effects of cement profile and graft density on graft incorporation remain unknown. In a bone chamber model in the rat, impacted graft particles were found to yield less bone ingrowth than unimpacted ones (Tagil and Aspenberg, 1998); however, the effect of graft density on incorporation has not been confirmed in the impaction allografting graft bed. And although the presence of cement at the endosteal surface presents a mechanical advantage, it does not allow restoration of the host bone stock in those regions. Furthermore, it is not clear how the presence of cement-endosteum contact affects incorporation of the surrounding graft regions. The resulting reduction in micromotion may provide a more favourable environment for bone ingrowth in the graft regions; however, it is not clear how the resulting reduced loading in graft regions would affect the incorporation process. While loading of the graft appeared to promote its incorporation in a study in the rabbit tibia (Wang et al., 2000), no difference in density was seen between loaded and non-loaded grafts in the sheep (van der Donk et al., 2002).  5.1.5 Relevance of current work with respect to other in vitro studies The results of the current study, which emphasize the role of the endosteum and the effect of the cement penetration profile on initial stem stability in impaction allografting, are valuable to future research studies. Factors such as fat or moisture content and graft irradiation have been shown to affect the stiffness and shear strength of cancellous graft particles (Voor et al., 2000; 110  Dunlop et al., 2003; Voor et al., 2004; Cornu et al., 2004; Butler et al., 2005). Based on the results of our structural tests (Chapter 2) and our finite element model (Chapter 4), however, changes in graft stiffness would only affect initial subsidence in the absence of cementendosteum contact. Other in vitro studies have explored how initial stability would be affected if the graft particles were processed differently or replaced with other materials. It was reported that freeze-drying the graft (Cornu et al., 2003), replacing it with cortical particles (Kligman et al., 2003), and using calcium phosphate particles as graft extenders (Grimm et al., 2001; Blom et al., 2002; van Haaren et al., 2005; Fujishiro et al., 2005) or as graft substitutes (Munro et al., 2006) reduced in vitro subsidence. It was also reported that fusion of the proximal graft region produced a greater reduction in subsidence than did fusion of the distal graft, indicating that emphasis should be placed upon the proximal graft region during surgery (Heiner et al., 2008). None of these studies, however, have factored in the cement penetration profile, which may have differed between the groups studied. For example, in the study examining the effect of graft fusion (Heiner et al., 2008), the fusion was simulated using epoxy, which may have hindered cement penetration compared to the clinical scenario. Therefore, in order to eliminate the possibility that a difference in cement content may be responsible for the observed effects, cement profiles must be examined. Other than the current study, only one in vitro study has considered the cement penetration profile. In that study, the use of vibration and drainage during graft compaction improved in vitro stability compared with traditional graft impaction (Bolland et al., 2007), and further investigation with microcomputed tomography revealed that the reduction in migration was attributed to differences in graft density rather than cement content (Bolland et al., 2008b).  5.1.6 Trends in impaction allografting research In addition to problems of fracture and excessive subsidence, the risk of disease transfer is a significant concern with the use of allografts. Transplantation of bone tissue involves the risk of transmitting viruses such as human immunodeficiency virus (HIV), and hepatitis B and C (Yao et al., 2007), as well as various bacterial contaminants. In one study, as many as 22% of femoral heads donated to an Australian bone bank were found to be contaminated (Sommerville et al., 2000). The most common contaminants found in bone grafts are in the Staphylococcus family (Sommerville et al., 2000; James et al., 2004; van de Pol et al., 2007), which could lead to deep infection in the recipient (van de Pol et al., 2007). 111  In order to address the problems of disease transmission and graft availability, various synthetic materials have been studied in mechanical and/or biological environments representative of impaction allografting. These materials include bioglass (Fujishiro et al., 1997; Brewster et al., 1999) and calcium phosphate ceramics such as hydroxyapatite (Fujishiro et al., 1997; Voor et al., 2004; van Haaren et al., 2005; Fujishiro et al., 2005; Coathup et al., 2008) or hydroxyapatite combined with tricalcium phosphate (Grimm et al., 2001; Blom et al., 2002; Voor et al., 2004; Blom et al., 2005; Hannink et al., 2006; Arts et al., 2006a; Arts et al., 2006b; Hannink et al., 2007). Calcium phosphate ceramic particles have shown promising results in in vivo studies. Fibrous ingrowth and new bone formation were seen in and around ceramic particles, whether these were used alone or as an extender to cancellous graft particles in bone chamber models (Arts et al., 2006b; Hannink et al., 2007). In one study, calcium phosphate granules smaller than 150 µm appeared to have resorbed within eight weeks in a rabbit model (Arts et al., 2006b). In a femoral impaction allografting model in sheep, the use of ceramic graft extender in one group yielded comparable postoperative outcomes (ground reaction forces, bone density and subsidence) to those of another group of sheep in which cancellous graft alone was used (Blom et al., 2005). In a similar sheep study, some new bone formation was observed whether 90% or 50% hydroxyapatite was added to the graft (Coathup et al., 2008). From a mechanical viewpoint, adding small particles of bioglass to morsellized cancellous graft can increase its resistance to shear forces (Brewster et al., 1999). The use of calcium phosphate particles as a graft substitute or extender in impaction allografting reduced in vitro stem subsidence (Grimm et al., 2001; Blom et al., 2002; van Haaren et al., 2005; Fujishiro et al., 2005). Nonetheless, calcium phosphate particles are much stiffer than graft (Verdonschot et al., 2001; Voor et al., 2004), and the use of hard synthetic particles in impaction allografting can increase the risk of femoral fracture during impaction (van Haaren et al., 2005). Three recent clinical studies have reported the use of calcium phosphate particles in femoral impaction allografting (Nich and Sedel, 2006; Fujishiro et al., 2008; Aulakh et al., 2009). The results have been encouraging. In all three studies, the hip score – which denotes pain and 112  function – improved in most patients. In one study, in which calcium phosphate particles were used as either extenders or substitutes, radiological evidence of cortical repair was noted in most patients (Nich and Sedel, 2006). In another study, when hydroxyapatite was added to the graft, the 13 year survival rate and function were reported to be comparable to that of patients in which graft alone was used (Aulakh et al., 2009). Nonetheless, the appearance of the graft and ceramic particles as noted during subsequent re-revisions varied. During re-revision surgeries, the synthetic particles were found to be partly integrated in two patients at 20 and 33 months (Nich and Sedel, 2006; Fujishiro et al., 2008), but mostly unincorporated in another patient revised 16 months (Fujishiro et al., 2008). Over the last few years, a number of research studies have been aimed at enhancing the incorporation of cancellous graft particles through additives such as bone morphogenetic protein, bisphosphonate, and stem cells. The use of bone morphogenetic protein BMP-7 was found to increase the distance of bone ingrowth into the graft particles in a bone chamber model in the rat (Tagil et al., 2000). In later studies, however, BMP-7 did not enhance new bone formation in rabbit and goat models (Tagil et al., 2003; Hannink et al., 2006; Buma et al., 2008), and it was found to increase resorption during the early stages of graft incorporation in a sheep model (McGee et al., 2004). Finally, the use of BMP-7 did not appear to affect implant migration in a clinical study (Karrholm et al., 2007). Bisphosphonate, a compound known to inactivate osteoclasts (bone-resorbing cells) has also been investigated as a graft additive (Jeppsson et al., 2003). However, while it was found to increase graft density, the use of bisphosphonate resulted in a shorter bone ingrowth distance than the use of BMP-7 alone. It was concluded that although the problems associated with the use of BMP-7 in impaction allografting may be solved by adding bisphosphonate, some of the benefits of BMP-7 would also be lost. Bone marrow stromal cells (BMSCs) have also generated recent research interest in impaction allografting (Tilley et al., 2006; Bolland et al., 2006; Korda et al., 2008; Bolland et al., 2008a; Green et al., 2009). Similar to mesenchymal stem cells, BMSCs have the potential to differentiate into many types of specialized cells including osteoblasts (bone-laying cells), but they are harvested from bone marrow rather than embryos. BMSCs have been shown to 113  survive impaction (Bolland et al., 2006), and adding them to morsellized graft yielded greater new bone formation around a cementless stem in a sheep model than did the use of graft alone (Korda et al., 2008). In a study investigating the use of poly-lactic acid as a graft substitute, adding BMSCs increased vascularization and new bone formation in a subcutaneous model in mice (Bolland et al., 2008a). Their potential benefit in enhancing bone formation is further supported by a recent clinical study in which BMSCs were used successfully in two patients in conjunction with cancellous graft particles to repair femoral bone defects (Tilley et al., 2006). In those patients, postoperative radiographs revealed that the lesions were replaced by bone having a higher density than the surrounding bone. Finally, the use of bone marrow (without extracting the stromal cells) as a graft additive has also been investigated in a group of patients undergoing femoral and/or acetabular impaction allografting (Deakin and Bannister, 2007). The results were encouraging, with radiological signs of incorporation observed in most patients. Bone marrow and bone marrow stromal cells could therefore prove useful in enhancing graft incorporation in impaction allografting. In conclusion, impaction allografting offers a unique potential to reverse problematic bone loss in hip arthroplasty. The success of this procedure, however, relies upon achieving adequate initial implant stability and graft incorporation. While its clinical results have varied in terms of both stability and graft incorporation, recent research aimed at finding solutions to these problems has generated valuable information. Mechanical studies such as those presented in this thesis provide orthopaedic surgeons and researchers with a growing understanding of the mechanisms that cause problematic implant subsidence, while biological studies are exploring promising strategies aimed at enhancing new bone formation. Moreover, the growing popularity of stem cells in medical research, as well as the recent lifting of a ban on embryonic stem cell studies in the United States, could lead to new methods for improving graft incorporation. Continuing research aimed at minimizing the risks of subsidence and incomplete graft remodeling is critical to improving the effectiveness of impaction allografting as a revision hip arthroplasty procedure.  114  5.2 Contributions 1. The relationship between initial implant subsidence and the cement penetration profile after femoral impaction allografting had not been determined previously. Constructs in which cement reached the endosteum were found to subside less than those with no cementendosteum contact. 2. It has thus far been widely assumed that a dense graft bed is critical in ensuring initial stability. This study has demonstrated that this assumption is only true if the cement profile has not reached the host bone interface – thus, the cement profile was shown to play a dominant role over that of graft density on implant stability. This finding emphasizes the importance of taking the cement profile into consideration in any study aiming to optimize the properties of the graft bed in impaction allografting. 3. This study developed constitutive equations that describe how the morsellized cancellous graft material behaviour in compression and shear are affected by the graft bed density achieved during impaction. No other study had described this relationship. These constitutive equations can be useful in future finite element simulations or mathematical models aiming to study the structural behaviour of femoral impaction allografting constructs and other surgical applications where morsellized cancellous graft is used. 4. This thesis identified slippage at the endosteal interface as a major contributor to subsidence and to variability in subsidence between constructs. This information will be valuable in developing strategies to minimize the risk of excessive implant subsidence in impaction allografting.  115  5.3 Conclusions 1. In femurs with moderate or severe bone loss, impaction allografting with cement yielded extensive cement contact with the host bone, particularly around the distal half of the stem. 2. The cement penetration profile had a large effect on initial stem stability after femoral impaction allografting – micromotion and migration were correlated negatively with the amount of cement in the construct and with the amount of cement-endosteum contact. 3. Graft density, compression stiffness and shear strength were proportional to the impaction force used. 4. A slower graft compaction method that allowed more time for the fluid to exudate out of the graft bed resulted in increased graft density, stiffness, and shear strength, but these increases were small relative to the effect of increasing the impaction force. 5. The relationship between graft density and initial implant subsidence was dependent on the cement profile: in the absence of cement-endosteum contact, subsidence decreased with increasing graft density; however, density did not affect implant subsidence in the presence of cement-endosteum contact. 6. Initial subsidence was primarily attributed to slippage at the endosteum and stem-cement interfaces, and most of the variability in subsidence between constructs of varying cement penetration profiles and graft densities was the result of slippage at the endosteum. 7. Some cement contact with the endosteum (e.g., 10% or more) is advisable to reduce the risk of excessive implant subsidence.  116  5.4 Future Work •  The relationship between the degree of bone deficiency, the magnitude of impaction force that can be achieved without causing fracture, and the resulting graft density and cement penetration profile should be investigated in cadaveric femurs.  •  Future autopsy and biopsy studies describing the cement penetration profile in impaction allografting patients would be useful. The histological method described in this thesis would be helpful in preserving the cement in the processing of histological slides.  •  The relationship between graft incorporation and cement profile should be explored. The potential for new bone formation is compromised in regions where cement is present; however, the increased stability provided by the presence of cement-endosteum contact may be favourable in promoting new bone formation in the surrounding pure graft regions.  •  The development of a non-invasive method that would enable surgeons to distinguish the cement penetration profile from the morsellized graft in impaction allografting patients would be valuable.  •  Methods aiming at enhancing graft incorporation at the host-bone interface should be explored. For example, coating the endosteum with marrow stromal cells or stem cells prior to graft impaction may improve peripheral graft incorporation and reduce implant subsidence.  •  The development of a method that would enable surgeons to control the cement penetration profile such that cement-endosteum contact occurs in regions that are not targeted for bone stock reconstitution could be helpful.  •  Future research aimed at enhancing the mechanical characteristics of the graft region in order to improve implant stability should also consider the cement penetration profile, as the effects of these factors on implant stability are interdependent.  117  5.5 References Abbott, L.C., Schottstaedt, E.R., Saunders, J.B., and Bost, F.C. (1947) The evaluation of cortical and cancellous bone as grafting material: a clinical experimental study. J Bone Joint Surg Am 29, 381-414. Arts, J.J.C., Verdonschot, N., Schreurs, B.W., and Buma, P. (2006a) The use of a bioresorbable nano-crystalline hydroxyapatite paste in acetabular bone impaction grafting. Biomaterials 27, 1110-8. Arts, J.J.C., Walschot, L.H.B., Verdonschot, N., Schreurs, B.W., and Buma, P. (2006b) Biological activity of tri-calciumphosphate/hydroxyl-apatite granules mixed with impacted morsellized bone graft. A study in rabbits. J Biomed Mater Res Part B. Appl Biomater 81B, 476-85. Atroshi, I., Ornstein, E., Franzen, H., Johnsson, R., Stefansdottir, A., and Sundberg, M. (2004) Quality of life after hip revision with impaction bone grafting on a par with that 4 years after primary cemented arthroplasty. Acta Orthop Scand 75, 677-83. Aulakh, T.S., Jayasekera, N., Kuiper, J.H., and Richardson, J.B. (2009) Long-term clinical outcomes following the use of synthetic hydroxyapatite and bone graft in impaction in revision hip arthroplasty. Biomaterials 30, 1732-8. Berzins, A., Sumner, D.R., Wasielewski, R.C., and Galante, J.O. (1996) Impacted particulate allograft for femoral revision total hip arthroplasty. In vitro mechanical stability and effects of cement pressurization. J Arthroplasty 11, 500-6. Blom, A.W., Cunningham, J.L., Hughes, G., Lawes, T.J., Smith, N., Blunn, G., Learmonth, I.D., and Goodship, A.E. (2005) The compatibility of ceramic bone graft substitutes as allograft extenders for use in impaction grafting of the femur. J Bone Joint Surg Br 87, 421-5. Blom, A.W., Grimm, B., Miles, A.W., Cunningham, J.L., and Learmonth, I.D. (2002) Subsidence in impaction grafting: the effect of adding a ceramic bone graft extender to bone. Proc Inst Mech Eng [H] 216, 265-70. Boldt, J.G., Dilawari, P., Agarwal, S., and Drabu, K.J. (2001) Revision total hip arthroplasty using impaction bone grafting with cemented nonpolished stems and charnley cups. J Arthroplasty 16, 943-52. Bolland, B.J., Kanczler, J.M., Ginty, P.J., Howdle, S.M., Shakesheff, K.M., Dunlop, D.G., and Oreffo, R.O. (2008a) The application of human bone marrow stromal cells and poly(dl-lactic acid) as a biological bone graft extender in impaction bone grafting. Biomaterials 29, 3221-7. Bolland, B.J., New, A.M., Madabhushi, G., Oreffo, R.O., and Dunlop, D.G. (2008b) The role of vibration and drainage in femoral impaction bone grafting. J Arthroplasty 23, 1157-64. Bolland, B.J., New, A.M., Madabhushi, S.P., Oreffo, R.O., and Dunlop, D.G. (2007) Vibration-assisted bone-graft compaction in impaction bone grafting of the femur. J Bone Joint Surg Br 89, 686-92. 118  Bolland, B.J., Partridge, K., Tilley, S., New, A.M., Dunlop, D.G., and Oreffo, R.O. (2006) Biological and mechanical enhancement of impacted allograft seeded with human bone marrow stromal cells: potential clinical role in impaction bone grafting. Regen Med 1, 457-67. Brand, R.A., Pedersen, D.R., and Yoder, S.A. (1986) How definition of "loosening" affects the incidence of loose total hip reconstructions. Clin Orthop Relat Res 210, 185-91. Brewster, N.T., Gillespie, W.J., Howie, C.R., Madabhushi, S.P., Usmani, A.S., and Fairbairn, D.R. (1999) Mechanical considerations in impaction bone grafting. J Bone Joint Surg Br 81, 118-24. Buma, P., Arts, J.J., Gardeniers, J.W., Verdonschot, N., and Schreurs, B.W. (2008) No effect of bone morphogenetic protein-7 (OP-1) on the incorporation of impacted bone grafts in a realistic acetabular model. J Biomed Mater Res B Appl Biomater 84, 231-9. Burchardt, H. (1983) The biology of bone graft repair. Clin Orthop 174, 28-42. Butler, A.M., Morgan, D.A., Verheul, R., and Walsh, W.R. (2005) Mechanical properties of gamma irradiated morselized bone during compaction. Biomaterials 26, 6009-13. Coathup, M., Smith, N., Kingsley, C., Buckland, T., Dattani, R., Ascroft, G.P., and Blunn, G. (2008) Impaction grafting with a bone-graft substitute in a sheep model of revision hip replacement. J Bone Joint Surg Br 90, 246-53. Cornu, O., Bavadekar, A., Godts, B., Van Tomme, J., Delloye, C., and Banse, X. (2003) Impaction bone grafting with freeze-dried irradiated bone. Part I. Femoral implant stability: cadaver experiments in a hip simulator. Acta Orthop Scand 74, 547-52. Cornu, O., Libouton, X., Naets, B., Godts, B., Van Tomme, J., Delloye, C., and Banse, X. (2004) Freeze-dried irradiated bone brittleness improves compactness in an impaction bone grafting model. Acta Orthop Scand 75, 309-14. de Roeck, N.J. and Drabu, K.J. (2001) Impaction bone grafting using freeze-dried allograft in revision hip arthroplasty. J Arthroplasty 16, 201-6. Deakin, D.E. and Bannister, G.C. (2007) Graft incorporation after acetabular and femoral impaction grafting with washed irradiated allograft and autologous marrow. J Arthroplasty 22, 89-94. Duncan, C.P., Masterson, E.L., and Masri, B.A. (1998) Impaction allografting with cement for the management of femoral bone loss. Orthop Clin North Am 29, 297-305. Dunlop, D.G., Brewster, N.T., Madabhushi, S.P., Usmani, A.S., Pankaj, P., and Howie, C.R. (2003) Techniques to improve the shear strength of impacted bone graft: the effect of particle size and washing of the graft. J Bone Joint Surg Am 85-A, 639-46. Eldridge, J.D., Smith, E.J., Hubble, M.J., Whitehouse, S.L., and Learmonth, I.D. (1997) Massive early subsidence following femoral impaction grafting. J Arthroplasty 12, 535-40. Elting, J.J., Mikhail, W.E., Zicat, B.A., Hubbell, J.C., Lane, L.E., and House, B. (1995) 119  Preliminary report of impaction grafting for exchange femoral arthroplasty. Clin Orthop Relat Res 319, 159-67. Enneking, W.F., Burchardt. H. , and Piotrowski, G. (1975) Physical and biological aspects of repair in dog cortical-bone transplants. J Bone Joint Surg Am 57, 237-52. Fetzer, G.B., Callaghan, J.J., Templeton, J.E., Goetz, D.D., Sullivan, P.M., and Johnston, R.C. (2001) Impaction allografting with cement for extensive femoral bone loss in revision hip surgery: a 4- to 8-year follow-up study. J Arthroplasty 16, 195-202. Flugsrud, G.B., Ovre, S., Grogaard, B., and Nordsletten, L. (2000) Cemented femoral impaction bone grafting for severe osteolysis in revision hip arthroplasty. Good results at 4year follow-up of 10 patients. Arch Orthop Trauma Surg 120, 386-9. Franzen, H., Toksvig-Larsen, S., Lidgren, L., and Onnerfalt, R. (1995) Early migration of femoral components revised with impacted cancellous allografts and cement. A preliminary report of five patients. J Bone Joint Surg Br 77, 862-4. Frei, H., Gadala, M.S., Masri, B.A., Duncan, C.P., and Oxland, T.R. (2006) Cement flow during impaction allografting: a finite element analysis. J Biomech 39, 493-502. Frei, H., Mitchell, P., Masri, B.A., Duncan, C.P., and Oxland, T.R. (2004) Allograft impaction and cement penetration after revision hip replacement. A histomorphometric analysis in the cadaver femur. J Bone Joint Surg Br 86, 771-6. Frei, H., O'Connell, J., Masri, B.A., Duncan, C.P., and Oxland, T.R. (2005) Biological and mechanical changes of the bone graft-cement interface after impaction allografting. J Orthop Res 23, 1271-9. Fujishiro, T., Nishikawa, T., Niikura, T., Takikawa, S., Nishiyama, T., Mizuno, K., Yoshiya, S., and Kurosaka, M. (2005) Impaction bone grafting with hydroxyapatite: increased femoral component stability in experiments using Sawbones. Acta Orthop 76, 550-4. Fujishiro, T., Nishikawa, T., and Takahiro, N. (2008) Histologic analysis of allograft mixed with hydroxyapatite-tricalcium phosphate used in revision femoral impaction bone grafting. Orthopedics 31, 277. Fujishiro, Y., Hench, L.L., and Oonishi, H. (1997) Quantitative rates of in vivo bone generation for Bioglass and hydroxyapatite particles as bone graft substitute. J Mater Sci Mater Med 8, 649-52. Gie, G.A., Linder, L., Ling, R.S., Simon, J.P., Slooff, T.J., and Timperley, A.J. (1993) Impacted cancellous allografts and cement for revision total hip arthroplasty. J Bone Joint Surg Br 75, 14-21. Gokhale, S., Soliman, A., Dantas, J.P., Richardson, J.B., Cook, F., Kuiper, J.H., and Jones, P. (2005) Variables affecting initial stability of impaction grafting for hip revision. Clin Orthop Relat Res 432, 174-80. Gore, D.R. (2002) Impaction bone grafting for total hip revision. Int Orthop 26, 162-5. 120  Green, D.W., Bolland, B.J., Kanczler, J.M., Lanham, S.A., Walsh, D., Mann, S., and Oreffo, R.O. (2009) Augmentation of skeletal tissue formation in impaction bone grafting using vaterite microsphere biocomposites. Biomaterials 30, 1918-27. Grimm, B., Miles, A.W., and Turner, I.G. (2001) Optimizing a hydroxyapatite/tricalciumphosphate ceramic as a bone graft extender for impaction grafting. J Mater Sci Mater Med 12, 929-34. Halliday, B.R., English, H.W., Timperley, A.J., Gie, G.A., and Ling, R.S. (2003) Femoral impaction grafting with cement in revision total hip replacement. Evolution of the technique and results. J Bone Joint Surg Br 85, 809-17. Hannink, G., Aspenberg, P., Schreurs, B.W., and Buma, P. (2006) High doses of OP-1 inhibit fibrous tissue ingrowth in impaction grafting. Clin Orthop Relat Res 452, 250-9. Hannink, G., Schreurs, B.W., and Buma, P. (2007) No positive effects of OP-1 device on the incorporation of impacted graft materials after 8 weeks: a bone chamber study in goats. Acta Orthop 78, 551-8. Hassaballa, M., Mehendale, S., Poniatowski, S., Kalantzis, G., Smith, E., and Learmonth, I.D. (2009) Subsidence of the stem after impaction bone grafting for revision hip replacement using irradiated bone. J Bone Joint Surg Br 91, 37-43. Heiner, A.D., Callaghan, J.J., and Brown, T.D. (2008) Stability differentials for proximal vs distal fusion of total hip arthroplasty femoral impaction grafts. The Journal of Arthroplasty 23, 921-6. Hisatome, T., Yasunaga, Y., Takahashi, K., and Ochi, M. (2004) Bone remodeling after impacted cancellous allograft in revision hip arthroplasty based on (99m)Tc-MDP bone scintigraphy. Arch Orthop Trauma Surg 124, 52-5. James, L.A., Ibrahim, T., and Esler, C.N. (2004) Microbiological culture results for the femoral head. Are they important to the donor? J Bone Joint Surg Br 86, 797-800. Jeppsson, C., Astrand, J., Tagil, M., and Aspenberg, P. (2003) A combination of bisphosphonate and BMP additives in impacted bone allografts. Acta Orthop Scand 74, 483-9. Karrholm, J., Garellick, G., and Herberts, P. (2007) Swedish Hip Arthroplasty Register – Annual report 2006. Department of Orthopaedics, Sahlgrenska University Hospital. Karrholm, J., Hultmark, P., Carlsson, L., and Malchau, H. (1999) Subsidence of a non-polished stem in revisions of the hip using impaction allograft. Evaluation with radiostereometry and dual-energy X-ray absorptiometry. J Bone Joint Surg Br 81, 135-42. Kligman, M., Rotem, A., and Roffman, M. (2003) Cancellous and cortical morselized allograft in revision total hip replacement: A biomechanical study of implant stability. J Biomech 36, 797-802. Knight, J.L. and Helming, C. (2000) Collarless polished tapered impaction grafting of the femur during revision total hip arthroplasty: pitfalls of the surgical technique and follow-up in 121  31 cases. J Arthroplasty 15, 159-65. Korda, M., Blunn, G., Goodship, A., and Hua, J. (2008) Use of mesenchymal stem cells to enhance bone formation around revision hip replacements. J Orthop Res 26, 880-5. Krupp, R.J., Malkani, A.L., Crawford, C.H., Capello, W.N., Feinberg, J.R., and Baker, D. (2006) Impaction grafting for femoral component revision using a non-polished bead-blasted chrome cobalt stem-average 8 1/2-year follow-up. J Arthroplasty 21, 1180-6. Lamerigts, N.M., Buma, P., Huiskes, R., Schreurs, W., Gardeniers, J., and Sloof, T.J. (2000) Incorporation of morsellized bone graft under controlled loading conditions. A new animal model in the goat. Biomaterials 21, 741-47. Leopold, S.S., Berger, R.A., Rosenberg, A.G., Jacobs, J.J., Quigley, L.R., and Galante, J.O. (1999) Impaction allografting with cement for revision of the femoral component. A minimum four-year follow-up study with use of a precoated femoral stem. J Bone Joint Surg Am 81, 1080-92. Lind, M., Krarup, N., Mikkelsen, S., and Horlyck, E. (2002) Exchange impaction allografting for femoral revision hip arthroplasty: results in 87 cases after 3.6 years' follow-up. J Arthroplasty 17, 158-64. Linder, L. (2000) Cancellous impaction grafting in the human femur: histological and radiographic observations in 6 autopsy femurs and 8 biopsies. Acta Orthop Scand 71, 543-52. Ling, R.S., Timperley, A.J., and Linder, L. (1993) Histology of cancellous impaction grafting in the femur. A case report. J Bone Joint Surg Br 75, 693-6. Mahoney, C.R., Fehringer, E.V., Kopjar, B., and Garvin, K.L. (2005) Femoral revision with impaction grafting and a collarless, polished, tapered stem. Clin Orthop Relat Res 432, 181-7. Majkowski, R.S., Bannister, G.C., and Miles, A.W. (1994) The effect of bleeding on the cement-bone interface. An experimental study. Clin Orthop Relat Res 299, 293-7. Masterson, E.L. and Duncan, C.P. (1997) Subsidence and the cement mantle in femoral impaction allografting. Orthopedics 20, 821-2. Masterson, E.L., Masri, B.A., and Duncan, C.P. (1997a) The cement mantle in the Exeter impaction allografting technique. A cause for concern. J Arthroplasty 12, 759-64. Masterson, E.L., Masri, B.A., Duncan, C.P., Rosenberg, A., Cabanela, M., and Gross, M. (1997b) The cement mantle in femoral impaction allografting. A comparison of three systems from four centres. J Bone Joint Surg Br 79, 908-13. McGee, M.A., Findlay, D.M., Howie, D.W.C.A., Ward, P., Stamenkov, R., Page, T.T., Bruce, W.J., Wildenauer, C.I., and Toth, C. (2004) The use of OP-1 in femoral impaction grafting in a sheep model. J Orthop Res 22, 1008-15. Meding, J.B., Ritter, M.A., Keating, E.M., and Faris, P.M. (1997) Impaction bone-grafting before insertion of a femoral stem with cement in revision total hip arthroplasty. A minimum 122  two-year follow-up study. J Bone Joint Surg Am 79, 1834-41. Mikhail, W.E., Wretenberg, P.F., Weidenhielm, L.R., and Mikhail, M.N. (1999) Complex cemented revision using polished stem and morselized allograft.Minimum 5-years' follow-up. Arch Orthop Trauma Surg 119, 288-91. Morgan, H.D., McCallister, W., Cho, M.S., Casnellie, M.T., and Leopold, S.S. (2004) Impaction allografting for femoral component revision: clinical update. Clin Orthop Relat Res 420, 160-8. Munro, N.A., Downing, M.R., Meakin, J.R., Lee, A.J., and Ashcroft, G.P. (2006) A hydroxyapatite graft substitute redices subsidence in a femoral impaction grafting model. Clin Orthop Relat Res 455, 246-252. Nelissen, R.G., Bauer, T.W., Weidenhielm, L.R., LeGolvan, D.P., and Mikhail, W.E. (1995) Revision hip arthroplasty with the use of cement and impaction grafting. Histological analysis of four cases. J Bone Joint Surg Am 77, 412-22. Nelissen, R.G., Valstar, E.R., Poll, R.G., Garling, E.H., and Brand, R. (2002) Factors associated with excessive migration in bone impaction hip revision surgery: a radiostereometric analysis study. J Arthroplasty 17, 826-33. Nich, C. and Sedel, L. (2006) Bone substitution in revision hip replacement. Int Orthop 30, 525-31. Ohashi, H., Matsuura, M., Ebara, T., Okamoto, Y., and Kou, H. (2009) Factors influencing the stability of stems fixed with impaction graft in vitro. Clin Orthop Relat Res (electronic publication ahead of print). Ornstein, E. (2002) Hip revisions with impacted morselized allograft bone and cement. Acta Orthop Scand Suppl 73, 1-69. Ornstein, E., Atroshi, I., Franzen, H., Johnsson, R., Sandquist, P., and Sundberg, M. (2001) Results of hip revision using the Exeter stem, impacted allograft bone, and cement. Clin Orthop Relat Res 389, 126-33. Ornstein, E., Atroshi, I., Franzen, H., Johnsson, R., Sandquist, P., and Sundberg, M. (2002) Early complications after one hundred and forty-four consecutive hip revisions with impacted morselized allograft bone and cement. J Bone Joint Surg Am 84-A, 1323-8. Ornstein, E., Franzen, H., Johnsson, R., Karlsson, M.K., Linder, L., and Sundberg, M. (2004) Hip revision using the Exeter stem, impacted morselized allograft bone and cement: a consecutive 5-year radiostereometric and radiographic study in 15 hips. Acta Orthop Scand 75, 533-43. Ornstein, E., Franzen, H., Johnsson, R., Stefansdottir, A., Sundberg, M., and Tagil, M. (2003) Hip revision with impacted morselized allografts: unrestricted weight-bearing and restricted weight-bearing have similar effect on migration. A radiostereometry analysis. Arch Orthop Trauma Surg 123, 261-7. 123  Ornstein, E., Franzen, H., Johnsson, R., and Sundberg, M. (2000) Radiostereometric analysis in hip revision surgery--optimal time for index examination: 6 patients revised with impacted allografts and cement followed weekly for 6 weeks. Acta Orthop Scand 71, 360-4. Pekkarinen, J., Alho, A., Lepisto, J., Ylikoski, M., Ylinen, P., and Paavilainen, T. (2000) Impaction bone grafting in revision hip surgery. A high incidence of complications. J Bone Joint Surg Br 82, 103-7. Piccaluga, F., Gonzalez Della Valle, A., Encinas Fernandez, J.C., and Pusso, R. (2002) Revision of the femoral prosthesis with impaction allografting and a Charnley stem. A 2- to 12-year follow-up. J Bone Joint Surg Br 84, 544-9. Robinson, D.E., Lee, M.B., Smith, E.J., and Learmonth, I.D. (2002) Femoral impaction grafting in revision hip arthroplasty with irradiated bone. J Arthroplasty 17, 834-40. Schreurs, B.W., Arts, J.J., Verdonschot, N., Buma, P., Slooff, T.J., and Gardeniers, J.W. (2005) Femoral component revision with use of impaction bone-grafting and a cemented polished stem. J Bone Joint Surg Am 87, 2499-507. Sierra, R.J., Charity, J., Tsiridis, E., Timperley, J.A., and Gie, G.A. (2008) The use of long cemented stems for femoral impaction grafting in revision total hip arthroplasty. J Bone Joint Surg Am 90, 1330-6. Sommerville, S.M., Johnson, N., Bryce, S.L., Journeaux, S.F., and Morgan, D.A. (2000) Contamination of banked femoral head allograft: incidence, bacteriology and donor follow up. Aust N Z J Surg 70, 480-4. Sorensen, J., Ullmark, G., Langstrom, B., and Nilsson, O. (2003) Rapid bone and blood flow formation in impacted morselized allografts: positron emission tomography (PET) studies on allografts in 5 femoral component revisions of total hip arthroplasty. Acta Orthop Scand 74, 633-43. Tagil, M. and Aspenberg, P. (1998) Impaction of cancellous bone grafts impairs osteoconduction in titanium chambers. Clin Orthop Rel Res 352, 231-8. Tagil, M. and Aspenberg, P. (2001) Fibrous tissue armoring increases the mechanical strength of an impacted bone graft. Acta Orthop Scand 72, 78-82. Tagil, M., Jeppson, C., and Aspenberg, P. (2000) Bone graft incorporation - Effects of osteogenic protein-1 and impaction. Clin Orthop Relate Res 371, 240-5. Tagil, M., Jeppsson, C., Wang, J.S., and Aspenberg, P. (2003) No augmentation of morselized and impacted bone graft by OP-1 in a weight-bearing model. Acta Orthop Scand 74, 742-8. Temmerman, O.P., Raijmakers, P.G., Heyligers, I.C., Comans, E.F., Lubberink, M., Teule, G.J., and Lammertsma, A.A. (2008) Bone metabolism after total hip revision surgery with impacted grafting: evaluation using H2 15O and. Mol Imaging Biol 10, 288-93. Notes: fluoride PET; a pilot study Tilley, S., Bolland, B.J., Partridge, K., New, A.M., Latham, J.M., Dunlop, D.G., and Oreffo, 124  R.O. (2006) Taking tissue-engineering principles into theater: augmentation of impacted allograft with human bone marrow stromal cells. Regen Med 1, 685-92. Ullmark, G., Hallin, G., and Nilsson, O. (2002) Impacted corticocancellous allografts and cement for revision of the femur component in total hip arthroplasty. J Arthroplasty 17, 140-9. Ullmark, G. and Linder, L. (1998) Histology of the femur after cancellous impaction grafting using a Charnley prosthesis. Arch Orthop Trauma Surg 117, 170-2. Ullmark, G. and Obrant, K.J. (2002) Histology of impacted bone-graft incorporation. J Arthroplasty 17, 150-7. Ullmark, G., Sorensen, J., Langstrom, B., and Nilsson, O. (2007) Bone regeneration 6 years after impaction bone grafting: a PET analysis. Acta Orthop 78, 201-5. van Biezen, F.C., ten Have, B.L., and Verhaar, J.A. (2000) Impaction bone-grafting of severely defective femora in revision total hip surgery: 21 hips followed for 41-85 months. Acta Orthop Scand 71, 135-42. van de Pol, G.J., Sturm, P.D., van Loon, C.J., Verhagen, C., and Schreurs, B.W. (2007) Microbiological cultures of allografts of the femoral head just before transplantation. J Bone Joint Surg Br 89, 1225-8. van der Donk, S., Buma, P., Verdonschot, N., and Schreurs, B.W. (2002) Effect of load on the early incorporation of impacted morsellized allografts. Biomaterials 23, 297-303. van Doorn, W.J., ten Have, B.L., van Biezen, F.C., Hop, W.C., Ginai, A.Z., and Verhaar, J.A. (2002) Migration of the femoral stem after impaction bone grafting. First results of an ongoing, randomised study of the exeter and elite plus femoral stems using radiostereometric analysis. J Bone Joint Surg Br 84, 825-31. van Haaren, E.H., Smit, T.H., Phipps, K., Wuisman, P.I., Blunn, G., and Heyligers, I.C. (2005) Tricalcium-phosphate and hydroxyapatite bone-graft extender for use in impaction grafting revision surgery. An in vitro study on human femora. J Bone Joint Surg Br 87, 267-71. van Kleunen, J.P., Anbari, K.K., Vu, D., and Garino, J.P. (2003) Massive revision hip arthroplasty: a technique utilizing femoral impaction allografting and collared, textured stems. The Univsersity of Pensylvania Orthopaedic Journal 16, 45-52. Van Kleunen, J.P., Anbari, K.K., Vu, D., and Garino, J.P. (2006) Impaction allografting for massive femoral defects in revision hip arthroplasty using collared textured stems. J Arthroplasty 21, 362-71. Verdonschot, N. and Huiskes, R. (1997) Acrylic cement creeps but does not allow much subsidence of femoral stems. J Bone Joint Surg Br 79, 665-9. Verdonschot, N., van Hal, C.T., Schreurs, B.W., Buma, P., Huiskes, R., and Slooff, T.J. (2001) Time-dependent mechanical properties of HA/TCP particles in relation to morsellized bone grafts for use in impaction grafting. J Biomed Mater Res 58, 599-604. 125  Voor, M.J., Nawab, A., Malkani, A.L., and Ullrich, C.R. (2000) Mechanical properties of compacted morselized cancellous bone graft using one-dimensional consolidation testing. J Biomech 33, 1683-8. Voor, M.J., White, J.E., Grieshaber, J.E., Malkani, A.L., and Ullrich, C.R. (2004) Impacted morselized cancellous bone: mechanical effects of defatting and augmentation with fine hydroxyapatite particles. J Biomech 37, 1233-9. Wang, J.S., Tagil, M., and Aspenberg, P. (2000) Load-bearing increases new bone formation in impacted and morsellized allografts. Clin Orthop Rel Res 378, 274-281. Weidenhielm, L.R., Mikhail, W.E., Wretenberg, P., Fow, J., Simpson, J., and Bauer, T.W. (2001) Analysis of the retrieved hip after revision with impaction grafting. Acta Orthop Scand 72, 609-14. Yao, F., Seed, C., Farrugia, A., Morgan, D., Cordner, S., Wood, D., and Zheng, M.H. (2007) The risk of HIV, HBV, HCV and HTLV infection among musculoskeletal tissue donors in Australia. Am J Transplant 7, 2723-6.  126  APPENDIX 1 MECHANICAL TESTING SET-UPS  127  A1.1 Simulation of hip joint loading Prior to structural testing of the implanted cadaveric femurs (Chapter 2), each specimen was potted distally at 13° of adduction and 0° of flexion (Figure A1.1). The implant was used as a reference for alignment rather than the femurs themselves to eliminate inter-specimen variability due to differences in femoral shaft curvature. The reference axis of the implant was defined by a 2.5 inch bold fastened into a thread hole on the proximal surface of the implant. Specimen alignment was achieved using two laser levels projecting the desired specimen angles in the sagittal plane and frontal plane.  θ Figure A1.1 Specimen potting schematics. Specimen alignment was achieved using two laser levels projecting the desired angles, shown as bold black lines, in the sagittal plane (vertical) and frontal plane (θ=13°). (Right) Custom-built pot, mounted onto a linear guide.  The application of the hip joint contact loads on the cadaveric femur specimens was done with a biaxial servohydraulic testing machine (Figure 2.1). A linear actuator force, F, of -955.5 ± 735.0 N, and a rotary actuator moment, M, of -5.88 ± 8.32 Nm (Figure A1.2) generated the desired craniocaudal and anteroposterior forces, which were described in Chapter 2. The  128  direction of the rotary moment was adjusted between left and right specimens, such that the largest peak anteroposterior force was consistently in the posterior direction. Actuator force, F Moment, M (right specimens) 10 8 6 4 2 0 -2 -4 -6 -8 -10  1600  Force (N)  1400 1200 1000 800 600 400 200 0 0  0.2  0.4  0.6  0.8  Moment (Nm)  Moment, M (left specimens) 1800  1  Time (s) Figure A1.2 Actuator force and moment for one cycle of simulated walking loads. Black line: Actuator force, in compression. Grey lines: Actuator moment for right specimens (solid) and left specimens (dashed).  An attempt was made to also simulate stair climbing loads, with the same hip contact load as walking in the frontal plane but with a higher anteroposterior load of 0.6 times body-weight. However, the higher torsional loading caused fracture of the first two specimens tested (Figure A1.3). The fractures were initiated between the tip of the implant and the potting level, indicating excessive stress concentration in this region. For this reason, the remaining specimens were not subjected to stair climbing loads.  129  Potted level  Figure A1.3 Fracture of two specimens which occurred during simulated stair climbing. Left specimen of femur pair 1 (left). Left specimen of femur pair 2 (right).  A1.2 Measurement of implant motion A motion measurement system was custom-designed for this study to measure the implant motion relative to the bone. The system consisted of a stem reference triangle, which was rigidly attached to the stem, and a femur reference frame, through which six linear variable differential transformers (GCD-121-250, Shaevitz Sensors, Hampton, Virginia) were rigidly attached to the femur (Figure A1.4).  130  location of femur sensor  pin  femur reference frame spacer  stem reference triangle  Figure A1.4 Custom-built motion measurement system components.  The stem reference triangle was attached to a pin, which was fitted into a 5 mm square hole that was machined 5 cm below the implant shoulder on the lateral side by electrical discharge (Figure A1.5). After the stem was implanted in the femur, an 8 mm diameter hole was drilled through the femur, guided by a custom-built steel hole-finder, to expose the hole in the implant (Figure A1.5, left). Prior to the surgical procedures, the hole in the implant was sealed with a small piece of cellophane tape to prevent bone cement from filling the hole during the surgical procedure. After drilling of the femur, the square hole was cleaned with acetone and a 5 mm square pin was glued into it with cyanoacrylate (Figure A1.5, right). The pin was allowed to set in place for at least two hours. The stem reference triangle was then attached to the pin with a set screw and the distance between the triangle’s reference coordinate system (Figure A1.8) and the implant surface, dtriangle, was measured. The junction between the triangle and the pin was strengthened with a drop of polymethyl methacrylate (Lecoset, LECO Corp, St Joseph Michigan).  131  Figure A1.5 Schematics showing the attachment of the stem reference triangle to the implant.  To ensure a consistent positioning of femur reference frame relative to the stem reference triangle, the frame was temporarily bolted onto to the triangle through a spacer (Figure A1.4). The femur reference frame was mounted onto the femur as described in Chapter 2, after which the frame was detached from the stem reference triangle and the spacer was removed. At this point, the frame remained attached rigidly to the femur and the triangle to the stem. The potted specimen was mounted on a linear guide and transferred to the servohydraulic testing machine (Figure A1.6). The sensors were attached to the frame in their respective locations and their positions were adjusted to give initial readings of 0 V, such that their output remained well within their calibrated range, i.e., +/- 1.7 V (Figure A1.7), during testing. To verify that the motion measurement assembly components were rigidly fixed to the stem and femur, the reference frame and triangle were tapped gently, ensuring that the sensor readings returned to zero.  132  linear + rotary actuator load cell specimen  actuator control panel  motion measurement system linear guide  Figure A1.6 Photograph showing a specimen being subjected to simulated walking loads on the servohydraulic testing machine.  The motion of the stem relative to the femur was obtained from the sensor voltages using equations A1.1 to A1.5. First, the LVDT voltages (V1, V2, V3, V4, V5, V6) were converted into sensor displacements (w1, v2, w3, u4, w5, v6): ⎧ w1 ⎫ ⎪v ⎪ ⎪ 2⎪ ⎪⎪w3 ⎪⎪ ⎨ ⎬ = [C1 ⎪u 4 ⎪ ⎪w5 ⎪ ⎪ ⎪ ⎪⎩ v 6 ⎪⎭  C2  C3  C4  C5  ⎧V1 ⎫ ⎪V ⎪ ⎪ 2⎪ ⎪⎪V ⎪⎪ C 6 ]× ⎨ 3 ⎬ ⎪V4 ⎪ ⎪V5 ⎪ ⎪ ⎪ ⎪⎩V6 ⎪⎭  Equation A1.1  where Ci were the LVDT calibration coefficients obtained experimentally by calibrating each sensor with a dial gauge micrometer, i.e., the slopes of the micrometer displacement vs. LVDT voltage curves: C1 = 642.7561 µm/V, C2 = 648.8754 µm/V, C3 = 653.5494 µm/V, C4 = 648.1879 µm/V, C5 = 640.1567 µm/V, C6 = 646.5890 µm/V (Figure A1.7).  133  LVDT 1  LVDT 2  1500  1500  y = 648.88x - 1E-13  1000 500 0 -2  -1  0  1  2  -500 -1000  Dial gauge position (microns)  Dial gauge position (microns)  y = 642.76x - 9E-14 1000 500 0 -2  -1  0  -1500  Sensor output (V)  Sensor output (V)  LVDT 3  LVDT 4  1500  1500  y = 648.19x - 1E-13  1000 500 0 0  1  2  -500 -1000  Dial gauge position (microns)  Dial gauge position (microns)  y = 653.55x - 8E-14  -1  1000 500 0 -2  -1  0  -1500  Sensor output (V)  Sensor output (V)  LVDT 5  LVDT 6  1500  1500  500 0 0 -500 -1000 -1500  Sensor output (V)  1  2  Dial gauge position (microns)  1000  -1  2  -1000  y = 646.59x - 1E-13  y = 640.16x - 5E-14 Dial gauge position (microns)  1  -500  -1500  -2  2  -1000  -1500  -2  1  -500  1000 500 0 -2  -1  0  1  2  -500 -1000 -1500  Sensor output (V)  Figure A1.7 Calibration of each motion sensor against a micron-precision dial gauge.  134  The three-dimensional displacement of the stem reference triangle relative to the femur reference frame (xt, yt, zt, θxt, θyt, θzt) was then obtained from the sensor displacements through the following geometrical relationship (Figure A1.8): ⎧ w1 ⎫ ⎡0 ⎪v ⎪ ⎢0 ⎪ 2⎪ ⎢ ⎪⎪w3 ⎪⎪ ⎢0 ⎨ ⎬ = −⎢ ⎢1 ⎪u 4 ⎪ ⎢0 ⎪w5 ⎪ ⎢ ⎪ ⎪ ⎪⎩ v 6 ⎪⎭ ⎣⎢0  0 1  0  R1  1 0  0  0  0 1 R1  0  0 0  0  0  0 1  0  − R1  1 0  0  0  0 ⎤ ⎧ xt ⎫ − R2 ⎥⎥ ⎪⎪ y t ⎪⎪ 0 ⎥ ⎪⎪ z t ⎪⎪ ⎥×⎨ ⎬ 0 ⎥ ⎪θxt ⎪ 0 ⎥ ⎪θy t ⎪ ⎥ ⎪ ⎪ R2 ⎦⎥ ⎪⎩θz t ⎪⎭  Equation A1.2  where R1 = 55.0 mm and R2 = 60.0 mm.  LVDT 2  LVDT 6  v2  (LVDT 3)  w3  v6  yt  R1  xt LVDT 4  R1  dtriangle (LVDT 1)  w1  R2  R1  (LVDT 5)  w5  R2  ys xs  u4  Pin  Stem  Figure A1.8 Coordinate system of stem and that of the stem reference triangle.  The three-dimensional displacement of the stem relative to the femur reference frame (xs, ys, zs, θxs, θys, θzs) was obtained from that of the stem reference triangle, based on the previously measured distance between the stem surface and the origin of the triangle coordinate system, dtriangle:  135  ⎧ x s ⎫ ⎧ xt ⎫ ⎧ θz t ⎫ ⎪y ⎪ ⎪y ⎪ ⎪ 0 ⎪ ⎪ s⎪ ⎪ t⎪ ⎪ ⎪ ⎪⎪ z s ⎪⎪ ⎪⎪ z t ⎪⎪ ⎪⎪− θxt ⎪⎪ ⎬ ⎨ ⎬ = ⎨ ⎬ + d triangle ⋅ ⎨ ⎪ 0 ⎪ ⎪θx s ⎪ ⎪θxt ⎪ ⎪ 0 ⎪ ⎪θy s ⎪ ⎪θy t ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪⎩ 0 ⎪⎭ ⎪⎩θz s ⎪⎭ ⎪⎩θz t ⎪⎭  Equation A1.3  Finally, the stem motion measurements were transformed to an anatomical coordinate system (Figure A1.9), using Equations A1.4 for right femurs, and A1.5 for the left ones.  proximal lesser trochanter  posterior medial  lateral  distal lateral  medial  valgus rotation  greater trochanter  posterior  femoral head anterior  anterior  flexion  retroversion  Figure A1.9. Illustration of the anatomical coordinate system used to describe threedimensional implant motion. ⎧ posterior translation ⎫ ⎧ x s ⎫ ⎪ lateral translation ⎪ ⎪ y ⎪ ⎪ ⎪ ⎪ s ⎪ ⎪⎪ distal translation ⎪⎪ ⎪⎪− z s ⎪⎪ ⎨ ⎬=⎨ ⎬ ⎪ va lg us rotation ⎪ ⎪ θx s ⎪ ⎪ ⎪ ⎪ θy s ⎪ flexion ⎪ ⎪ ⎪ ⎪ ⎪⎩ ⎪⎭ ⎪⎩ θz s ⎪⎭ retroversion  Equation A1.4  136  ⎧ posterior translation ⎫ ⎧ − x s ⎫ ⎪ lateral translation ⎪ ⎪ y ⎪ ⎪ ⎪ s ⎪ ⎪ ⎪⎪ distal translation ⎪⎪ ⎪⎪ − z s ⎪⎪ ⎬ ⎬=⎨ ⎨ ⎪ va lg us rotation ⎪ ⎪ θx s ⎪ ⎪ ⎪− θy s ⎪ ⎪ flexion ⎪ ⎪ ⎪ ⎪ ⎪⎭ ⎪⎩− θz s ⎪⎭ ⎪⎩ retroversion  Equation A1.5  The motion measurement from each sensor was validated against a dial gauge micrometer, and a maximum error of 2.2 µm was found (Table A1.1). Table A1.1 Validation of each motion sensor against a dial gauge micrometer. Dial (µm) -230.8 -180.8 -130.8 -80.8 -30.8 19.2 69.2 119.2 169.2 219.2 -212.6 -162.6 -112.6 -62.6 -12.6 37.4 87.4 137.4 187.4 237.4 -202.1 -152.1 -102.1 -52.1 -2.1 47.9 97.9 147.9 197.9 -202.1  Sensor (µm) LVDT 1 -231.4 -180.7 -130.0 -80.3 -30.2 19.4 69.6 118.7 168.1 216.9 LVDT 3 -210.4 -160.8 -110.8 -61.4 -12.4 36.3 85.8 135.3 185.3 236.3 LVDT 5 -201.6 -150.4 -102.4 -52.8 -2.2 47.7 96.7 147.9 198.4 -201.6  Absolute error (µm)  Dial (µm)  0.6 0.1 0.8 0.5 0.6 0.2 0.4 0.5 1.1 2.2  -225.6 -175.6 -125.6 -75.6 -25.6 24.4 74.4 124.4 174.4 224.4  2.1 1.8 1.8 1.1 0.1 1.2 1.6 2.2 2.2 1.1  -227.9 -177.9 -127.9 -77.9 -27.9 22.1 72.1 122.1 172.1 222.1  0.5 1.7 0.3 0.7 0.1 0.2 1.2 0.0 0.6 0.5  -196.4 -146.4 -96.4 -46.4 3.6 53.6 103.6 153.6 203.6 -196.4  Sensor (µm) LVDT 2 -227.8 -177.8 -126.5 -75.3 -25.3 24.7 75.6 125.6 174.5 223.9 LVDT 4 -228.7 -178.6 -128.3 -78.9 -27.6 22.6 73.2 122.5 172.7 222.3 LVDT 6 -196.6 -146.1 -95.4 -45.5 3.2 53.7 103.5 153.2 202.4 -196.6  Absolute error (µm) 2.1 2.2 0.9 0.3 0.3 0.3 1.2 1.2 0.2 0.5 0.8 0.7 0.4 1.1 0.3 0.4 1.1 0.4 0.6 0.1 0.2 0.3 1.0 0.9 0.4 0.1 0.2 0.4 1.2 0.2  137  The accuracy of the motion measurement system was also evaluated along the three translational axes by mounting the femur reference frame onto an over-reamed composite femur and subjecting the stem to motion relative to the femur along each axis. The servohydraulic testing machine was used to generate implant motion along the z-axis, and an xy table was used to generate motion along the x and y axes. The motion measurements were validated against a micron-precision dial gauge and maximum errors of 5.8 µm, 2.4 µm and 2.7 µm were found along the x, y, and z axes, respectively (Table A1.2).  Table A1.2 Validation of motion measurement system against dial gauge micrometer.  x-displacement  y-displacement  z-displacement  Dial displacement (µm) -65.0 -111.5 -147.0 -195.0 35.0 79.0 139.0 170.0 -67.5 -118.0 -173.5 -212.0  Motion Measurement (µm) -61.1 -114.7 -152.8 -198.0 34.5 79.9 136.6 168.4 -68.0 -120.7 -173.9 -211.4  Absolute error (µm) 3.9 3.2 5.8 3.0 0.5 0.9 2.4 1.6 0.5 2.7 0.4 0.6  138  APPENDIX 2 HISTOMORPHOMETRIC ANALYSIS METHODS  139  A2.1 Preparation of histology slides Prior to cross-sectioning, the remaining soft tissue was removed as much as possible with a scalpel, pliers and sand paper. The stem was removed by tapping it proximally with a hammer, and the stem-shaped canal was filled with polymethyl methacrylate (Lecoset, Leco Corp.) mixed with green or blue food colorant (not red). The specimen was aligned and mounted on the diamond saw table (model 310 CP, Exakt Apparatebau, Norderstedt, Germany), and a permanent marker was used to make a small mark on the outside of femur, indicating the proximal side of each cross section. This mark will later be dissolved during ethanol dehydration, but it was useful initially for shape matching the cross sections - the proximal cross-sections were easy to match given their taper, but the distal ones would have been be less obvious without a mark. The specimen was cut into 6 mm thick cross sections. The cross-sections were shape-matched and labeled, while being careful not to damage the fragile graft bed surfaces. Starting at the first proximal cross-section where the entire stem area was present, the cross-sections were labeled: e.g., 1065L-3: donor identification number 1065, left femur, third slide from the most proximal cross-section, etc. The cross-section label was marked on the distal surface of each cross-section with a pencil, on either cortical bone or cement. Alternating cross-sections were prepared used for histomorphometric analysis. The crosssections were set in formalin and dehydrated in ethanol solutions as described in Chapter 2. The total duration of the dehydration process was four days – any longer than that was found to cause some swelling and dissolution of the bone cement. After dehydration, the cross sections were dried in an oven at 60°C for one hour, and allowed to cool down for a few minutes. The dried cross-sections were placed proximal surface down in plastic moulds and embedded in resin (Buehler EpoThin, Lake Bluff, Illinois). The resin (powder) was mixed with the hardener (liquid) with the appropriate ratio as instructed by the manufacturer until it was uniform, after which it was poured over the cross-sections. Each cross-section was lifted and lowered with tweezers to dislodge the larger air bubbles, and 140  placed in a vacuum chamber. A vacuum of 25 in. Hg was applied, held for a couple of minutes, and released slowly, i.e., at a rate of approximately 0.5 in.Hg/s. The cross section was lifted and lowered again with tweezers to dislodge the newly appeared air bubbles. The vacuum process was repeated until no air bubbles remained and the cross-section was thoroughly embedded in resin. The specimens were labeled again, this time with pencil on a small piece of paper which was inserted in the resin facing away from the cross-section, along the mould wall. The moulds were covered with a lid to protect from dust or debris, and the resin was allowed to set overnight. The resin-impregnated cross-sections were processed into histology slides by trained technicians (Jesse Chen and Caron Fournier). One slide was obtained from the proximal surface of each cross-section, and it was ground to a thickness of 200 µm.  A2.2 Calcium staining of histology slides Stain preparation: • Mix  together the following in a beaker:  -  2 g alizarin red S dye (powder)  -  100ml distilled water  • Insert  a pH meter.  • Slowly • Stop  add/mix 0.5% ammonium hydroxide.  when pH reaches between 4.1 and 4.3.  Slide staining procedure: • Rinse  slide with distilled water  • Immerse • Rinse • Set  slide in Alizarin red stain solution  gently with distilled water  aside to air dry  141  A2.3 Slide photography for histomorphometric analysis digital camera  ruler  slide  green cardboard  microscope stand digital camera  cardboard box  slide  ruler  light box  Figure A2.1 Slide photography set-up for histomorphometric analysis. (Top) Set-up for measurement of cement contact with endosteum and cement area. (Bottom) Set-up for measurement of graft porosity.  142  APPENDIX 3 GRAFT SHEAR AND COMPRESSION TESTING  143  The custom-built shear and compression testing apparatuses used for the graft characterization experiments (Chapter 3) are shown in Figures A3.1 and A3.2.  Instron 8874  Fc actuator/load cell piston spacers  graft mould  Figure A3.1 Compression testing apparatus. Schematics of compression testing assembly on the servohydraulic testing machine (left). Mould with drainage holes (right).  compression actuator load cell piston top mould cylinder  Instron 8874  Fc graft shearing actuator Fs  compression actuator load cell  shear actuator / load cell bottom mould cylinder  piston top mould spacers bottom mould shearing piston  load cell  linear guide  linear guide  Figure A3.2 Shear testing apparatus. Schematics (left) and photograph (right) of shear testing assembly on the servohydraulic testing machine.  144  Low impaction force (approximately 300 N) 350  Impaction force (N)  300 250 200 150 100 50 0 0  0.5  1  1.5  2  2.5 Time (s)  3  3.5  4  4.5  5  3.5  4  4.5  5  4  4.5  5  High impaction force (approximately 600 N) 700  Impaction force (N)  600 500 400 300 200 100 0 0  0.5  1  1.5  2  2.5 Time (s)  3  Moderate impaction force (approximately 900 N) 1200  Impaction force (N)  1000 800 600 400 200 0 0  0.5  1  1.5  2  2.5 Time (s)  3  3.5  Figure A3.3 Forces applied during simulated impaction for (top) low, (middle) moderate, and (bottome) high impaction forces.  145  APPENDIX 4 FINITE ELEMENT ANALYSIS MESH AND STEM MOTION RESULTS  146  A4.1 Finite element mesh: (left) intact, and (right) under load. Note that the displacements were amplified by a factor of two for this illustration.  147  δendosteum  δgraft/composite δmantle δstem-cement  Stem Cement mantle Graft-cement composite Graft  Figure A4.2 Migration and micromotion decomposed into motion occurring at the stem-cement and endosteum interfaces (δstem-cement and δendosteum, respectively); motion as a result of shear deformation of the cement mantle (δmantle); and motion due to shear deformation of the graft and/or graft/cement composite regions (δgraft/composite).  148  Profile A stem-cement interface graft/graft-cement composite 4000  stem-cement sliding  cement mantle endosteum interface stem-cement bonded  Migration (microns)  3500  Stem  3000 2500 2000  Cement mantle  1500 1000 500 0 300  600  900  300  600  900  Graft-cement composite  Impaction force (N) stem-cement interface graft/graft-cement composite 350  stem-cement sliding  cement mantle endosteum interface stem-cement bonded  Graft  Micromotion (microns)  300 250 200 150 100 50 0 300  600  900  300  600  900  Impaction force (N)  Figure A4.3 Implant migration (top) and micromotion (bottom) predicted with the FE model (Chapter 4), for femoral impaction allografting constructs with no cementendosteum contact, but having 25% cement penetration into the graft bed (Profile A).  149  Profile B stem-cement interface graft/graft-cement composite 3000 stem-cement sliding  cement mantle endosteum interface stem-cement bonded  Stem  Migration (microns)  2500 2000 1500  Cement mantle  1000 500 0 300  600  900  300  600  900  Graft-cement composite  Impaction force (N)  stem-cement interface graft/graft-cement composite  Micromotion (microns)  250  stem-cement sliding  cement mantle endosteum interface  Graft  stem-cement bonded  200 150 100 50 0 300  600  900  300  600  900  Impaction force (N)  Figure A4.4 Implant migration (top) and micromotion (bottom) predicted with the FE model (Chapter 4), for constructs with no cement-endosteum contact, but having 50% cement penetration into the graft bed (Profile B).  150  Profile C stem-cement interface graft/graft-cement composite 1600  stem-cement sliding  cement mantle endosteum interface stem-cement bonded  Migration (microns)  1400  Stem  1200 1000 800  Cement mantle  600 400 200 0 300  600  900  300  600  900  Graft-cement composite  Impaction force (N) stem-cement interface graft/graft-cement composite  cement mantle endosteum interface  100  Micromotion (microns)  90  stem-cement sliding  Graft  stem-cement bonded  80 70 60 50 40 30 20 10 0 300  600  900  300  600  900  Impaction force (N)  Figure A4.5 Implant migration (top) and micromotion (bottom) predicted with the FE model (Chapter 4), for constructs with no cement-endosteum contact, but having 75% cement penetration into the graft bed (Profile C).  151  Profile D stem-cement interface graft/graft-cement composite 450  cement mantle endosteum interface  endosteum sliding  endosteum bonded  Migration (microns)  400  Stem  350 stemcement sliding  300 stemcement sliding  250 200 150 100  Cement mantle  stemcement bonded  stemcement bonded  900  600  300  900  600  300  900  600  300  900  600  0  300  50  Graft-cement composite  Impaction force (N) stem-cement interface graft/graft-cement composite 45  endosteum sliding  40 Micromotion (microns)  cement mantle endosteum interface  Graft  endosteum bonded  35 30 25 20  stemcement sliding  15 10  stemcement bonded  stemcement sliding  stemcement bonded  900  600  300  900  600  300  900  600  300  900  600  0  300  5  Impaction force (N)  Figure A4.6 Implant migration (top) and micromotion (bottom) predicted with the FE model (Chapter 4), for constructs with 11% cement-endosteum contact (Profile D).  152  Profile E stem-cement interface graft/graft-cement composite  cement mantle endosteum interface  300 endosteum sliding  endosteum bonded  Migration (microns)  250  Stem  200 150 100  stemcement sliding  50  stemcement bonded  stemcement sliding  stemcement bonded  Cement mantle  900  600  300  900  600  300  900  600  300  900  600  300  0  Graft-cement composite  Impaction force (N) stem-cement interface graft/graft-cement composite  Micromotion (microns)  35  cement mantle endosteum interface  endosteum sliding  Graft  endosteum bonded  30 25 20 15  stemcement sliding  10  stemcement bonded  stemcement sliding  stemcement bonded  5 900  600  300  900  600  300  900  600  300  900  600  300  0 Impaction force (N)  Figure A4.7 Implant migration (top) and micromotion (bottom) predicted with the FE model (Chapter 4), for constructs with 24% cement-endosteum contact (Profile E).  153  Profile F stem-cement interface graft/graft-cement composite 225  cement mantle endosteum interface  endosteum sliding  endosteum bonded  Migration (microns)  200  Stem  175 150 stemcement bonded  125 100 75  Cement mantle stemcement sliding  stemcement sliding  50  stemcement bonded  900  600  300  900  600  300  900  600  300  900  600  0  300  25  Graft-cement composite  Impaction force (N) stem-cement interface graft/graft-cement composite  cement mantle endosteum interface  Graft  30 endosteum bonded  25 20 15  900  600  300  900  600  300  900  600  300  0  stemcement bonded  stemcement sliding  stemcement bonded  900  5  600  stemcement sliding  10  300  Micromotion (microns)  endosteum sliding  Impaction force (N)  Figure A4.8 Implant migration (top) and micromotion (bottom) predicted with the FE model (Chapter 4), for constructs with 36% cement-endosteum contact (Profile F).  154  Profile G stem-cement interface graft/graft-cement composite  cement mantle endosteum interface  200 endosteum sliding  endosteum bonded  Stem  150 125 stemcement bonded stemcement sliding  300  900  600  300  900  600  300  0  300  25  stemcement bonded  stemcement sliding 900  50  900  75  Cement mantle  600  100  600  Migration (microns)  175  Graft-cement composite  Impaction force (N) stem-cement interface graft/graft-cement composite  cement mantle endosteum interface  25 20 15 10  stemcement sliding  stemcement bonded  900  600  300  900  600  300  900  600  300  900  600  stemcement sliding  5 0  Graft  endosteum bonded  stemcement bonded  300  Micromotion (microns)  endosteum sliding  Impaction force (N)  Figure A4.9 Implant migration (top) and micromotion (bottom) predicted with the FE model (Chapter 4), for constructs with 49% cement-endosteum contact (Profile G).  155  Profile H stem-cement interface graft/graft-cement composite  cement mantle endosteum interface  175 endosteum sliding  endosteum bonded  Stem  125 100 stemcement sliding  stemcement bonded  stemcement bonded  50  300  900  600  300  900  600  300  900  600  300  25 0  Cement mantle  900  stemcement sliding  75  600  Migration (microns)  150  Graft-cement composite  Impaction force (N) stem-cement interface graft/graft-cement composite  cement mantle endosteum interface  25 20 15 10  stemcement bonded  stemcement sliding  Graft  900  600  300  900  600  300  900  600  300  900  600  stemcement sliding  5 0  endosteum bonded  stemcement bonded  300  Micromotion (microns)  endosteum sliding  Impaction force (N)  Figure A4.10 Implant migration (top) and micromotion (bottom) predicted with the FE model (Chapter 4), for constructs with 61% cement-endosteum contact (Profile H).  156  Profile I stem-cement interface graft/graft-cement composite  cement mantle endosteum interface  150 endosteum sliding  endosteum bonded  Stem  100 stemcement sliding  50  stemcement bonded  stemcement sliding  stemcement bonded  300  900  600  300  900  600  300  900  600  300  25 0  Cement mantle  900  75  600  Migration (microns)  125  Graft-cement composite  Impaction force (N) stem-cement interface graft/graft-cement composite  cement mantle endosteum interface  20  15  endosteum bonded  stemcement bonded  stemcement sliding  stemcement bonded  10  Graft 5  900  600  300  900  600  300  900  600  300  900  600  0  stemcement sliding  300  Micromotion (microns)  endosteum sliding  Impaction force (N)  Figure A4.11 Implant migration (top) and micromotion (bottom) predicted with the FE model (Chapter 4), for constructs with 80% cement-endosteum contact (Profile I).  157  stem-cement interface  cement mantle  graft/graft-cement composite  endosteum interface  3000  µendosteum = 0.61  Migration (microns)  2500  µendosteum = 1.0  2000 1500 1000 500 0 300  600  600  stem-cement interface  cement mantle  graft/graft-cement composite  endosteum interface  300  µendosteum = 0.61  250  Migration (microns)  900 300 Impaction force (N)  900  µendosteum = 1.0  200 150 100 50 0 300  600  900 300 Impaction force (N)  600  900  Figure A4.12 Sensitivity analysis – effect of endosteum friction coefficient, µendosteum, for profile B (top) and profile E (bottom). Shown are the results for constructs where the stemcement interface was sliding. 158  stem-cement interface  cement mantle  graft/graft-cement composite  endosteum interface  2500  νgraft = 0. 2  Migration (microns)  2000 1500  νgraft = 0.4  1000 500 0 300  600  900  300  600  900  Impaction force (N) stem-cement interface  cement mantle  graft/graft-cement composite  endosteum interface  300  νgraft = 0. 2  νgraft = 0.4  Migration (microns)  250 200 150 100 50 0 300  600  900  300  600  900  Impaction force (N)  Figure A4.13 Sensitivity analysis – effect of graft Poisson’s ratio, νgraft, for profile B (top) and profile E (bottom). Shown are the results for constructs where the stem-cement interface was sliding.  159  APPENDIX 5  UBC RESEARCH ETHICS BOARD CERTIFICATES OF APPROVAL  160  161  162  163  APPENDIX 6  GLOSSARY  164  adduction: action by which the femur is drawn towards the body axis anterior: situated towards the front of the body anteroposterior: along an axis directed from the front towards the back of the body arthroplasty: joint replacement surgery aseptic: without infection cancellous bone: part of a bone having a porous structure cement mantle: region surrounding a cemented femoral stem that is occupied by cement; in this thesis, cement mantle refers to the region of pure cement surrounding the stem that does not include the regions of cement-graft composite cortex (femoral): the outer layer of a bone, in this case the cortical shell making up femoral canal cortical bone: parts of a bone having a dense non-porous structure craniocaudal: along an axis directed from the head towards the feet diaphysis: the shaft of a long bone, in this case the femur distal: situated away from the point of attachment; e.g., distal femur is the part of the femur that is located towards the knee, therefore distal stem migration refers to the migration of the stem towards the knee endosteum: membrane lining the medullary cavity of a bone; in this thesis endosteum is used to refer to the inner surface of the femoral canal (adjective: endosteal) condyle (femoral): rounded prominence at the distal end of the femur flexion/extension: rotation of the stem relative to the femur in the median plane, i.e., the plane that divides the body into left and right halves; see illustration in Appendix 1 histomorphometric analysis: quantitative study of the microscopic structure of a tissue; in this case femurs and/or graft specimens iatrogenic fracture: fracture that is caused by the surgical procedure in vitro: in an artificial environment outside a living organism; e.g., in cadaveric or artificial femurs (antonym: in vivo) in vivo: within a living organism; e.g., inside a living animal incorporation (of bone graft): process by which the graft becomes a viable part of the bone into or onto which it was implanted; i.e., the host femur initial stability (of the stem): describes the initial postoperative resistance to subsidence of the stem within the bone construct (synonym: primary stability) instability (of the hip): partial or complete joint dislocation intramedullary canal: canal that is located inside a bone, in this case the femur lateral: located away from the median plane, i.e., the plane that divides the body into left and right halves; see illustration in Appendix 1 lytic defect: bone defect caused by osteolysis; i.e., degeneration of bone tissue through disease  165  medial: situated towards the median plane (antonym: lateral) mediolateral: along an axis directed from the medial side towards the lateral side medullary cavity: marrow cavity inside the shaft of a long bone metaphysis: proximal femur morphology: study of structure or shape of a living thing such as an organ; in this case the femur-implant construct (adjective: morphological) morsellized cancellous bone (MCB): cancellous bone graft that has been morsellized into particles with a bone mill neo-medullary canal: intramedullary canal that has been re-shaped surgically periosteal surface: fibrous outer layer of a bone, to which muscles attach posterior: situated towards the back (antonym: anterior) prophylactic: preventive; e.g., aiming to prevent fracture proximal: situated towards the point of attachment; e.g., proximal femur is the part of the femur that is located towards the hip proximodistal: along the long axis of the femur, directed towards the knee radiolucent: almost entirely invisible in radiographs remodeling (of the graft): formation of an oriented trabecular structure within the graft that has adapted to the directions of loading retroversion/anteversion: rotation of the stem relative to the femur in the transverse plane, i.e., about the long axis of the femur; see illustration in Appendix 1 stem: femoral implant; these terms have been used interchangeably throughout this thesis subsidence (of the stem): non-reversible gradual stem displacement relative to the femur (synonym: migration) trabecular bone: see cancellous bone valgus/varus rotation: rotation of the stem relative to the femur in the coronal plane; i.e., the  plane that divides the body into front and back halves; see illustration in Appendix 1  166  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0070873/manifest

Comment

Related Items