Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

An in vitro biomechanical study of impaction allografting for revision total hip arthroplasty Robinson, Marcus Calem 2003

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

Item Metadata

Download

Media
831-ubc_2003-0294.pdf [ 35.72MB ]
Metadata
JSON: 831-1.0079037.json
JSON-LD: 831-1.0079037-ld.json
RDF/XML (Pretty): 831-1.0079037-rdf.xml
RDF/JSON: 831-1.0079037-rdf.json
Turtle: 831-1.0079037-turtle.txt
N-Triples: 831-1.0079037-rdf-ntriples.txt
Original Record: 831-1.0079037-source.json
Full Text
831-1.0079037-fulltext.txt
Citation
831-1.0079037.ris

Full Text

AN IN VITRO BIOMECHANICAL STUDY OF IMPACTION ALLOGRAFTING FOR REVISION TOTAL HIP ARTHROPLASTY by M A R C U S C A L E M ROBINSON B . A . S c , The University of British Columbia, 2000  A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T OF THE REQUIREMENTS FOR THE DEGREE OF M A S T E R OF APPLIED SCIENCE in THE F A C U L T Y OF G R A D U A T E STUDIES (Department of Metals and Materials Engineering)  We accept this thesis as conforming to the required standard  The University of British Columbia April 2003 © Marcus Calem Robinson, 2003  U B C Rare Books and Special Collections - Thesis Authorisation Form  Page 1 of 1  In p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t of the requirements f o r an advanced degree at the U n i v e r s i t y of B r i t i s h Columbia, I agree t h a t the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and study. I f u r t h e r agree t h a t p e r m i s s i o n f o r e x t e n s i v e copying of t h i s t h e s i s f o r s c h o l a r l y purposes may be granted by the head of my department or by h i s or her r e p r e s e n t a t i v e s . I t i s understood t h a t copying or p u b l i c a t i o n of t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l not be allowed without my w r i t t e n p e r m i s s i o n .  Department The U n i v e r s i t y of B r i t i s h Columbia Vancouver, Canada  http://www.library.ubc.ca/spcoll/thesauth.html  4/22/2003  Abstract The impaction allografting procedure for the treatment of failed hip reconstructions has shown promising results, but a number of intraoperative and postoperative concerns exist. A fundamental understanding of the biomechanical characteristics of the construct would help in providing surgical guidelines to address the issues of stem subsidence, femoral fracture, and resource usage. To this end, an in vitro model was developed to compare the structural response of the construct as a function of the material composition (i.e. 100% morsellized bone, 100% cement, and a combination of morsellized bone and cement as used in the traditional impaction allografting procedure), distal plug fixation (i.e. fixed orfreedistal plug), and stem surface finish (i.e. polished or polished and coated with mould release). To replicate clinical conditions while eliminating confounding variables, the model system comprised aluminum tubing to represent the femur and an axisymmetric singletaper stem. One orthopaedic surgeon performed all procedures in a total of 25 specimens. Successive uniaxial cyclic compression was applied to the models to simulate 1-BW (i.e. body weight), 2-BW, and 3-BW loading for a total of 1000 cycles. Applied force, axial stem displacement, strain at the outside of the tubes were recorded to calculate construct stiffness, subsidence, and hoop and axial strain. At the conclusion of each test, the stems were pulled out of the tubes to gain further insight into the relationship among material properties, loading, and interfacial shear strengths. The structural behaviour of the impaction allografting (IA) construct was closer to that of the 100%o cement (CE) construct than that of the 100% morsellized bone (MB) construct. The IA and CE constructs did not show substantial stem subsidence and hence exhibited predominantly axial strain resultingfromhigh shear stresses at the stem-cement interface. Conversely, the MB specimens showed excessive stem subsidence and consequent increase in the hoop:axial strain ratio with morsellized bone compression. There was no change in the structural behaviour of the specimens when a mould release-coated stem was substituted for the polished stem. With the distal plug removed, the IA and CE specimens slipped at the tube-material interface, suggesting that the distal plug or some other form of external or geometric support is required for structural integrity.  Table of Contents ABSTRACT  ii  T A B L E OF CONTENTS  iii  LIST OF TABLES  v  LIST OF FIGURES  vi  ACKNOWLEDGEMENTS 1.0  viii  INTRODUCTION  1  1.1  REVIEW OF CLINICAL SERIES  1.2  REVIEW OF MATERIALS AND BONE REMODELLING RESEARCH  20  1.3  SUMMARY  30  METHODS  32  1.1.1 1.1.2 1.1.3 1.1.4 1.1.5 1.1.6 1.1.7 1.1.8 1.2.1 1.2.2 1.2.3 1.2.4  2.0 2.1  Introduction to the Structure and Properties of Bone The Material Properties of Bone and Other THA Materials The Material Properties of Morsellized Bone (MB) Bone Graft Viability, Resorption, and Morbidity in RTHA with IA  MATERIALS PREPARATION  2.1.1 2.1.2 2.1.3 2.1.4 2.1.5 2.1.6 2.2  45  Specimen Preparation Structural Testing Image A nalysis Pull-out Tests Statistical Analysis  45 49 55 56 57  RESULTS  3.2.1 3.2.2 3.2.3 3.2.4  58  INSIDE SURFACE FINISH TESTS STRUCTURAL TESTING  General Observations Load-Displacement Image Analysis Pull-out Tests  21 23 24 26  32 34 34 36 41 43  TEST PROCEDURE  3.1 3.2  1 4 5 8 9 12 15 18  32  Prosthesis Development Cement Preparation Bone Graft Preparation Aluminum Tubing Design Strain Gauge Preparation Instrument Preparation  2.2.1 2.2.22.2.3 2.2.4 2.2.5 3.0  1  Introduction to Total Hip Arthroplasty (THA) The Indications for Revision Total Hip Arthroplasty (RTHA) The Procedure for RTHA with Impaction Allografting (IA) The Goals of Revision Total Hip Arthroplasty Performance Measurement Standards The Performance of RTHA with IA - complications andfailure mechanisms Prosthesis Subsidence Mechanisms Critical Variables Affecting the Performance of RTHA with IA  58 60  ; '•'  60 63 72 73  3.2.5 4.0  Load-Strain  76  DISCUSSION  82  4.1 INSIDE SURFACE FINISH 4.2 STRUCTURAL TESTING 4.2.1 Limitations 4.2.2 General Observations 4.2.3 Load-Displacement 4.2.4 Image Analysis 4.2.5 Pull-out Tests 4.2.6 Load-Strain 4.2.7 Effect of Coated Stem 4.3 CLINICAL RELEVANCE 4.4 SUMMARY 5.0  82 83 84 85 86 94 94 96 101 102 104  CONCLUSIONS AND RECOMMENDATIONS  106  REFERENCES  108  APPENDIX A: TESTING MACHINE AND STRAIN GAUGE CALIBRATION  114  APPENDIX B: STRAIN GAUGE DATA  116  iv  List of Tables Table 1 Common component and fixation variations in Total Hip Arthroplasty 4 Table 2 R T H A complications and failure scenarios found in recent studies 13 Table 3 Reported possible causes of subsidence 17 Table 4 Component, technique, and patient variables associated with R T H A with impaction allografting on the femoral side 19 Table 5 Indicative values of Young's moduli and static strengths for most materials and interfaces used in T H A reconstruction (Huiskes & Verdonschot 1997) 24 Table 6 Radiographic evaluation of R T H A with IA: comparison of two series 29 Table 7 Subject information for the bone used in the current study 35 Table 8 Before and after conditions of specimens used in the inside surface finish tests. Specimen 5 was not tested because the M B lost all structural integrity after the distal plug had been removed 38 Table 9 Testing matrix 46 Table 10 Load history assumed for implanted hip prostheses (Baleani et al 1999) 52 Table 11 Specimen conditions 63 Table 12 Effect of coating prosthesis with a silicon-based mould release spray (IA constructs only) 71 Table 13 Mean hoop:axial strain ratio (standard deviation in parentheses) 81  v  List of Figures Figure 1 Schematic section of a cemented Charnley prosthesis (Huiskes & Verdonschot 1997) 3 Figure 2 Femoral bone-graft-cement-stem construct after impaction allografting (Frei 2001) 8 Figure 3 Proposed subsidence mechanisms for the impaction allografting construct 17 Figure 4 Typical construct of a long bone (Anatomy & Physiology 1993) 22 Figure 5 Exeter stem typically used in the impaction allografting technique. Note the centralizer covering the distal tip of the stem [Howmedica (http://www.revisionhip.com/frames/restexe.html)] 33 Figure 6 Photograph of morsellized bone chips (left) produced by Lere Bone Mill 35 Figure 7 Prepared specimens for the "inside surface finish" test. Three specimens contained 100% cement. Five specimens contained 100% morsellized bone at various levels of impaction 37 Figure 8 Test apparatus for "inside surface finish" testing (left). Schematic of test apparatus (right). A gap was left in the bottom of the tube through which the specimen could move 39 Figure 9 Aluminum tube used to represent the femoral diaphysis (left). Note location of rosettes A , B, and C. Rosette D was placed 180 degrees from rosette B. Right: Polished steel stem used to represent Exeter stem 42 Figure 10 Custom impaction allografting instruments. From left: proximal cement seal, proximal tamp, distal impactor, proximal impactor (phantom stem). Front: guide wire 44 Figure 11 Jig for specimen fixation in the testing machine: top (left) and base (right) 45 Figure 12 Left: Aluminum tubes used for impaction allografting models. Photograph shows guide wire in place. Centre: Schematic of photo. Right: Schematic of distal impaction 47 Figure 13 Left: Creation of the neo-medullary canal via impaction with the proximal impactor. Centre: Schematic of proximal impaction. Right: Completed neo-medullary canal 47 Figure 14 Left: Retrograde filling of the neo-medullary canal. Right: Schematic of cement filling 48 Figure 15 Left: Finished impaction allografting tube model. Right: Schematic of impaction allografting tube model 49 Figure 16 (A) Hip joint forces determined from an in vivo instrumented prosthesis. (B) and (C) Distribution of the bending stresses at the medial side of the frontal plane of the prosthesis (Huiskes & Verdonschot 1997) 51 Figure 17 Impaction allografting model loaded into the servo-hydraulic test machine for cyclic compressive loading 53 Figure 18 Representation of the loading conditions applied to the tube models 53 Figure 19 Schematic of test set-up. Test specimens were loaded into the load frame and subjected to cyclic compressive loading for a total of 1000 cycles at 1 Hz. 55 Figure 20 Schematic of Pull-out test 56 Figure 21 A typical load-displacement plot for the inside surface finish tests 58 Figure 22 Push-out force for each of the eight specimens tested: three 100% cement (CE) specimens and eight 100% morsellized bone (MB) specimens 59 Figure 23 The tube-cement and tube-morsellized bone measured interfacial shear strengths 60 Figure 24 A : Enhanced radiograph of a typical impaction allografting model. The gamma level was decreased so that the cement was distinguishable from the morsellized bone. B: Original radiograph. 61 Figure 25 Cement penetration through the morsellized bone to the tube wall 62 Figure 26 A : Representative illustration of the loading pattern. B: Representative resultant stem displacement when subjected to loading pattern on right 64 Figure 27 Typical stem displacement curves for impaction allografting (IA), 100% cement (CE), and 100% morsellized bone (MB) 65 Figure 28 The stiffness of the model constructs as a function of material composition: impaction allografting (IA), 100% cement (CE), and 100% morsellized bone (MB) 66 Figure 29 The stiffness of the M B construct as a function of density of morsellized bone used to construct the model. Straight lines represent "best fits" for each data set 67 Figure 30 Overall subsidence of the model constructs at the conclusion of testing 68 Figure 31 The subsidence of the morsellized bone (MB) constructs as a function of the mass of bone used to construct the specimen 69  vi  Figure 32 Schematic of testing with the distal plug removed. A : Initial conditions. B: The IA and C E constructs slipped at 2*BW loading. C: The M B construct did not slip even at 3*BW loading 70 Figure 33 Radiograph of a typical M B specimen taken before and after testing. Left: Before testing. Right: After testing the stem had subsided without disruption of the morsellized bone-tube interface.72 Figure 34 Typical force-displacement plots obtained in pull-out tests 73 Figure 35 Mean pull-out force for the three tested material compositions: impaction allografting (IA), 100% morsellized bone (MB), and 100% cement (CE) 74 Figure 36 Pull-out mechanisms. A : The stem pulled out of M B specimens. B: The IA and C E constructs pull-out as a cohesive unit. C: A sample of IA pull-out 75 Figure 37 Sample plot of hoop (A) and axial (B) strains measured at a particular rosette location 77 Figure 38 Detail view of hoop and axial strain measurements from Figure 37 at the beginning of the third loading regime. Applied force is also plotted .' 79 Figure 39 Mean hoop strain at the third loading regime for the three material compositions as a function of rosette location 80 Figure 40 Mean axial strain at the third loading regime for the three material compositions as a function of rosette location 80 Figure 41 Before and after radiograph of IA subsidence outlier. Left: Before testing. Right: After testing construct has failed at the tube-morsellized bone interface 91 Figure 42 Comparison of relative trends obtained in this study (shown on the left side of each bar) with those presented by Berzins et al (shown on the left side of each bar) 92 Figure 43 Load transfer via a straight-tapered cone pushed into a cylindrical counterpart. A : The shear stress at the bonded interface can equilibrate the applied force. B: At a smooth, press-fitted interface, equilibrium relies on the vertical component of compressive interface stress. For slightly tapered cones, a significant amount of subsidence must occur in order for the compressive stress required for equilibrium to develop. (Huiskes & Verdonschot 1997) 97  vii  Acknowledgements I would like to acknowledge the support of the Composites Group and the Department of Metals & Materials Engineering, the Division of Orthopaedic Engineering Research, and the Department of Orthopaedics. Thanks to Simon, Dominic, and Carolyne for their technical help with the project. Thanks to Elvis, Hanspeter, and Anthony for their technical and philosophical advice throughout my time as a graduate student. Thanks to Ka-Hay, Eric, Erik, Neil, and Morris for their continued support. Thanks to Lars Frederiksen and Susan Baer for making the process gratifying. Special thanks to my supervisors, Drs. Goran Femlund and Tom Oxland for providing an outstanding environment for learning. Both Drs. Femlund and Oxland were instrumental in my development as a student and in my personal growth. Thank you for your counsel and support throughout the years.  viii  1.0 Introduction 1.1  Review of Clinical Series  The following section provides a review o f the clinical aspects o f revision total hip arthroplasty with impaction allografting (IA). The section begins with a brief introduction to total hip arthroplasty followed by a more in-depth look at the indications for revision surgery, the surgical procedure, and a synopsis o f the complications associated with revision surgery as presented in recent literature.  1.1.1  Introduction to Total H i p Arthroplasty ( T H A )  Total hip arthroplasty ( T H A ) is a highly successful surgical procedure to replace the hip joint with artificial components that is performed between 500,000 and one million times per year (Huiskes & Verdonschot 1997). T H A is considered to be effective in relieving pain and restoring full or partial mobility o f the hip joint. Indications such as osteoarthritis, rheumatoid arthritis, congenital deformities, and post-traumatic syndromes may be treated with T H A (Huiskes & Verdonschot 1997).  Osteoarthritis ( O A ) refers to a disease of joint cartilage, associated with secondary changes in the underlying bone, which may ultimately cause pain and impair the function of the affected joint (Oxford Reference: Concise Medical Dictionary). O A involves the breakdown o f the articular surface o f weight-bearing joints, causing an inflammatory response. Arthroplasty, or joint replacement, may be employed after conservative treatment methods have failed.  1  Rheumatoid arthritis (RA) is a less common indication for T H A as it is a systemic disease leading to chronic inflammation that typically involves aches and pains of smaller joints. Congenital deformities of the skeleton (acetabulum, pelvis, or femur) or pelvic soft tissue may result in irregular hip-joint wear that may cause pain or impair function. Surgical intervention with T H A may be required to correct deformities. Finally, trauma to the hip such as falling on one's side may require T H A if the articular surface has been significantly compromised or if the fracture is unable to heal satisfactorily by conventional methods.  Depending on the clinical presentation of the patient, one of several T H A methods may be used for treatment. In general, the orthopaedic surgeon exposes the proximal femur and performs an osteotomy at the femoral neck. The acetabulum and femoral medullary canal are prepared for the implant. The implant consists of a femoral stem, femoral head, acetabular implant, and acetabular liner (Figure 1). After performing a trial reduction, the final components are implanted and aligned.  2  o — — lateral, mewSial - — o  ;  Figure 1 Schematic section of a cemented Charnley prosthesis (Huiskes & Verdonschot 1997)  The development of the THA components and the surgical procedure has progressed gradually since the introduction of the Charnley device in the early 1960's to an extent that there are currently several options for surgeons. Today, very specific indications demand very specific combinations of implant components and surgical procedure. Table 1 presents some of the more common component and fixation variations. However, considerable debate still exists as to the ideal implant for any given indication. Further complicating matters, variables such as individual surgeon preference, hospital economics and politics, and vendor relations may affect the final implant selection criteria.  3  Table 1 Common component and fixation variations in Total Hip Arthroplasty  Variations Stem material shape surface finish coating  titanium, Co-Cr, stainless steel collar/collarless, straight/tapered/double-tapered, short/long polished/roughened none/hydroxyapatite coating/porous coating/proximal coating  Head material diameter  Co-Cr, stainless steel, ceramic 22 - 32 mm  Acetabulum material  UHMWPE, ceramic, Co-Cr, stainless steel  Technique stem fixation cement/cementless (press-fit/porous ingrowth) liner fixation press-fit/retaining ring acetabular fixation cement/cementless (pressfit/porousingrowth/macrolock screw) exposure osteotomy  1.1.2  lateral/posterolateral/anterolateral/posterior routine/transtrochanteric/extended osteotomy  The Indications for Revision Total Hip Arthroplasty (RTHA)  While state-of-the-art in total hip arthroplasty is debatable, nearly all of the procedures commonly practised provide a significant improvement in the quality of life of patients with debilitating hip pathologies. Regardless of the particulars of the THA procedure, the overall effect is that 90% of patients experience a successful THA for at least 10 years (Huiskes & Verdonschot 1997).  Revision of a total hip arthroplasty (RTHA) is required after clinical failure of the joint complex. For those who do experience failure of the total hip replacement, the cause of  4  failure may be dislocation, wear, component failure, loosening, or infection. A misaligned or subsided stem is prone to dislocation by impingement with excessive joint movement or loading. After the first dislocation, recurrent dislocations at less severe hip loads are likely. Wear of the polyethylene liner releases tiny particles that migrate to the component interfaces and trigger an immunoresponse. The response is the release of macrophage, or scavenger, cells that dissolve the affected bone (osteolysis), resulting in implant loosening. Mechanical failure of the implant components is rare, but the effects of many years of cyclic loading may have an effect on the structural integrity of the components. Loosening may occur due to osteolysis, but may also occur as a result of loss of interface integrity (often by either overloading or by stress shielding). A n infected hip implant causes pain and is generally removed and replaced. A l l of these failure scenarios are possible indications for revision surgery.  1.1.3  The Procedure for R T H A with Impaction Allografting (IA)  The severity of bone loss associated with osteoporosis, osteolysis, and implant extraction is often such that it is very difficult to obtain a stable construct in a revision hip arthroplasty. The term stable is often used in the clinical setting to denote a structure that does not require medical intervention nor is at risk for requiring medical attention. This does not necessarily imply structural rigidity as designated by the engineering community. In challenging cases such as these, the orthopaedic surgeon has a number of options, including treatment as a primary, cemented or uncemented reconstruction, and using a longer stem and distal fixation. Revision with a cemented stem and revision with the cementless stem technique have been unsuccessful in treating segmental femoral  5  defects and extreme femoral bone loss (Mohler & Cowin 1998; Barba & Paprosky 1998). With the cemented component procedure, there is no biological healing and there may be bone resorption due to stress shielding. Furthermore, cement fixation in a damaged proximal femur is not ideal, there is some question as to the long-term effectiveness of cement fixation, and re-revision is difficult (Rosenberg 1998). The cementless technique, while providing biological fixation by promoting bone ingrowth, has been limited by femora with poor bone quality and geometrical inconsistencies. Without adequate initial fixation and appropriate prosthesis contact, the environment is not conducive to bone ingrowth. Similarly, cementless femoral revisions with cancellous grafting have shown early failures and are not generally performed (Leopold et al 2000).  In an attempt to integrate the benefits of both the cemented and cementless solutions, Gie and colleagues (Gie et al 1993) outlined a procedure for the femoral component known as impaction allografting (IA) that provides an initially stable construct and the opportunity for bony reconstitution. Impaction allografting is most appropriate for revision of the femoral component and severe bone loss from the proximal femur, ectatic proximal diaphysis, thin cortices, and a large diameter canal (Leopold et al 1999).  The impaction allografting procedure for RTHA involves the creation of a bone foundation in severely defective femora using morsellized bone and then implanting a new hip stem with acrylic (PMMA) cement. Morsellized allograft bone (MB) chips are impacted into the medullary canal to fill cavitary defects or to assist complex reconstructions. Approximately three femoral heads are ground in a bone mill to produce  6  particles of 3-5 mm diameter (Haddad & Duncan 1999), although some inconsistency has been reported. Brodt (Brodt et al 1998) found that typically 80% of the bone graft (by weight) was in the range of 0.42 - 3.2 mm while Tanabe (Tanabe et al 1999) found that at least 50% of the bone graft was less than 2 mm for 4 typical bone mills. Generally trabecular bone is used and a slurry is produced comprising bone chips, marrow (fat + water), and water. After the primary prosthesis is removed, the medullary canal may be reamed and the distal plug is positioned. The MB slurry is inserted and impacted with distal and proximal impactors. The proximal impactors are cannulated phantom stems that are oversized to provide the necessary space for cement. Impaction occurs several times as the surgeon works proximally from the distal plug and concludes with more precise impaction employing a series of proximal tamps. Bone cement is injected into the neomedullary canal and pressurised to infiltrate the MB. Finally, the revision stem is inserted and reduced. Figure 2 shows a schematic of the RTHA with impaction allografting. The patient may mobilise the limb within days of the surgery, but full weight bearing may not occur until 3 months post-operatively. Refer to Gie (Gie et al 1993), Leopold (Leopold et al 1999), and van Biezen (van Biezen et al 2000) for further description of the standard impaction allografting technique. Recently, some attempt has been made to refine the process by employing a radial impaction technique rather than the traditional axial impaction (Stulberg 2000). The aim of this technique is to provide a simplified impaction technique and more consistent impacted graft.  7  Figure 2 Femoral bone-graft-cement-stem construct after impaction allografting (Frei 2001)  1.1.4  The Goals of Revision Total Hip Arthroplasty  Pain, loss of mobility, and osteolysis are often associated with a failed hip implant. Furthermore, the extraction of the primary hip implant and the natural ageing process of the patient contribute to the depletion of femoral bone quality. Poor bone quality increases the risk of femoral fracture and makes revisions difficult. Consequently, the goals of revision total hip arthroplasty (RTHA) are to relieve pain and to restore joint function. This can be achieved by providing immediate stable fixation and by promoting the restitution of bone stock. Restitution of bone stock facilitates re-revision surgeries, strengthens the femur, and may aid in providing secondary stem fixation.  The combination of the cement and morsellized bone used in the impaction allografting technique may provide both a structural and biological solution. The Exeter system for impaction allografting (based on primary cemented reconstructions) is thought to succeed because, with an extremely low surface roughness, the stem has the ability to subside  8  within the cement mantle and into the distal centralizer (Fowler et al 1988). Masterson et al (Masterson et al 1997) confirmed that the Exeter system aimed to maintain compression of the MB by cold flow of the cement during gradual subsidence of the prosthesis. With subsidence, Gie, Fowler, et al (Gie et al 1990) found the Exeter stem in the primary reconstruction to impart radial forces on the cement mantle and stabilize the construct. Huiskes and Verdonschot (Huiskes & Verdonschot 1997) have confirmed that cement creeps, but concede that "Acrylic cement.. .does not allow much subsidence of femoral stems". Moreover, no in vitro studies of impaction allografting have been able to replicate the amount of subsidence reported clinically (Malkani et al 1996; Berzins et al 1996). Finally, there is some question as to the reliability and accuracy of the reported measures of subsidence.  1.1.5  Performance Measurement Standards  The performance of RTHA with impaction allografting may be clinically quantified by means of histological analysis of biopsies or retrievals, radiographic analysis, radiostereometric analysis, DEXA, or clinical outcome measures.  Histology is the study of the structure of tissues by means of special staining techniques combined with light and electron microscopy (Oxford Reference: Concise Medical Dictionary). This invasive technique allows the clinician to evaluate the nature and structure of the tissue surrounding the implant.  9  A radiograph is an image produced on a film or photographic plates by directing X-rays through the body (Oxford Reference: Concise Medical Dictionary). The radiograph allows the clinician to grade the cementing technique and to look for evidence of stem loosening, subsidence, changes in implant position, interfacial break-down, osteolysis, and remodelling (Leopold et al 1999).  In radiostereometric analysis (RSA), small tantalum balls are implanted on the stem and the greater and lesser trochanters at the time of surgery and radiographs are taken at various time intervals. This allows the clinician to determine implant migration to an accuracy of 100 micrometers (Huiskes & Verdonschot 1997).  Dual energy x-ray absorptiometry, or DEXA, is a 2D scanning technique generally used to determine bone mineral density or bone mass. Two concentrated x-ray beams are passed through an object and the intensity is measured on the other side. The amount of x-ray energy that is blocked by bone varies with the density of the bone. As a 2-D technique, DEXA does not account for the "depth" of the bone and larger diameter bones will report a higher density than smaller bones of equal density. DEXA can be used to determine bone mass in vivo to examine bone remodelling (Huiskes & Verdonschot 1997).  The Harris Hip Score is an example of one. of several quality-of-life rating systems used by clinicians. In this interview-style system, the patient is awarded a score based on pain,  10  function, absence of deformity, and range of motion. This method of analysis is often the most convenient to obtain because it only requires a telephone call or letter to the patient.  A final method of reporting the success of RTHA is survival analysis. In this method, the number of patients that do not need a re-revision after a stated period of time is assessed statistically.  Studies have been conducted to examine the reliability and repeatability of the radiographical method of performance quantification. Radiographic procedures are not standardised and tend to be inherently imprecise because of the 3D to 2D conversions. Radiographic analysis can also be difficult and highly dependent on the individual performing the analysis (van Biezen et al 2000). Furthermore, Tagil & Aspenberg (Tagil & Aspenberg 2001) showed that impacted bone appears radiographically to remodel, but histological analyses revealed a heterogeneous region with a mixture of living and dead bone. Linder (Linder 2000) examined histological versus radiographical analysis and found that radiographic criteria used for primary THA do not necessarily apply to revision THA. Radiolucent lines in a primary hip reconstruction indicate that fibrous tissue orfibrocartilagehas resorbed and replaced the bone closest to the implant. However, in RTHA with impaction allografting, the presence of radiolucent lines indicates viable tissue closest to the graft, but the absence of radiolucent lines says little about the viability of the soft tissue layer.  11  1.1.6  The Performance of RTHA with IA - complications and failure mechanisms  Impaction allografting has shown promising clinical results with complications including stem subsidence and femoral fracture, and is a challenging, resource intensive procedure. Since the presentation of the impaction allografting technique in 1993, several studies have used one or more assessment methods to determine the performance of RTHA with impaction allografting. In 1999, Leopold and colleagues (Leopold et al 1999) studied 29 revision hips and found that the impaction allografting technique provided satisfactory clinical and radiographic results in the intermediate term with a 92% survivorship at 6 years and the mean Harris Hip Score increasing from 54 to 87. Linder (Linder 2000) concurred with the results of Leopold and colleagues, finding good functional results, good radiographic results, and few radiolucent lines in 14 revised hips. In fact, most clinical series reviewing impaction allografting for revision hip surgery give positive results - there is an associated decrease in pain and increase in mobility. In a review of the literature to date, Leopold and colleagues (Leopold et al 2000) found that femoral impaction allografting with cement is generally good and has shown to restore bone stock. Even severely damaged femora can be successfully treated by impaction allografting (van Biezen et al 2000).  As with any other revision surgery, however, the success of impaction allografting is generally worse than that of primary THA because the endosteal femoral surface is smooth (providing poor interface fixation) and because of the severity of bone loss (Brewster et al 1999). And while the short and intermediate-term success of impaction allografting is encouraging, there is some concern about the long-term durability of the  12  construct (Leopold et al 2000). Further, the radiographic success has been questioned because bone healing is unpredictable and the interpretation of radiographs is inferred, rather than absolute (Linder 2000).  There are several common causes of RTHA failure and many surgical complications. Even advocates of impaction allografting concede that high complication rates make the procedure attractive for only very specific indications (Masri 2001, Leopold et al 1999). A cursory list of studies of RTHA with impaction allografting is presented in Table 2. An "x" indicates causes of failure or surgical complications that are mentioned in a particular paper. Complications that are marked with an "o" were observed, but not considered to be a clinical concern.  Table 2 RTHA complications and failure scenarios found in recent studies Perioperative Availability  Cost  Piccaluga et al 2002  Postoperative  Infection  Surgical difficulty  X  X  Femoral fracture X  Nonunion/ misalignment X  Dislocation  Subsidence  X  X  Voor et al 2000  X  Knight & Helming 2000  X  X  van Biezen et al 2000  X  X  Stulberg 2000  X  X  X  X  X  X  Ornstein et al 2000  X  Leopold etal 1999  x  Duncan etal 1998  X  X  X  0  X  X  0  X  0  X  Keene etal 1998  X  Masterson et al 1997 Ornstein etal 1997  X X  Medingetal 1997 Masterson & Duncan 1997 Eldridgeetal 1997  X  X X X  Slooffet al 1996  X  Ling 1996  X  Eltingetal 1995  X  X  0  X  X  Franzen et al 1995  0  0  Gie etal 1993  X  0  13  0 0  X  0  Around the time of the surgery, the main concerns with impaction allografting include donor bone availability, cost, infection, surgical difficulty, and femoral fracture. With at least three femoral heads required for each femur, the current impaction allografting procedure requiring morsellized allograft bone places a significant demand on an already anaemic bone supply. Both disease transfer and infection are serious concerns when performing impaction allografting, although current screening procedures make disease transfer rare. The extra steps involved by cementing and impaction make impaction allografting a difficult procedure. It is often technically challenging to obtain consistent cement and MB mantles and it can be physically demanding to achieve the ideal compaction level. Furthermore, cerclage wiring is often required to allow the femur to withstand impaction. The act of impacting the allograft into place imparts considerable stress on the severely defective femora, often resulting in intraoperative fractures (see Table 2).  Postoperatively, complications and failure mechanisms include non-union, repeated hip dislocation, and stem subsidence. Trochanteric non-union (in the case of trochanteric osteotomy) may be a result of incorrect surgical alignment, but may be exacerbated by the settling of the MB with loading. As with primary hip dislocation, revision hip dislocation is generally caused by excessive joint moments or normal joint moments in combination with a subsided or loose stem. Finally, subsidence refers to the distal movement of the stem with respect to the femur and is a common source of failure (Elting et al 1995, Eldridge et al 1997, Franzen et al 1995). Severe stem subsidence is  14  associated with thigh pain, hip dislocation, and component revision (Leopold et al 1999). However, there is considerable debate as to whether subsidence is advantageous, detrimental, or benign (Leopold et al 2000). A small amount of subsidence (1-2 mm) seems to be obligatory to have the stem "settle"; a medium amount of subsidence (1-10 mm) seems to be acceptable; and a large amount of subsidence (>10 mm) seems to be unacceptable (Karrholm et al 1999, Nivbrant et al 1999).  The extent to which the allograft incorporates or remodels is critical to the stability of the construct and will be discussed in more detail in the following section.  In summary, there are both perioperative and postoperative concerns that may limit the attractiveness of RTHA with impaction allografting. Perioperatively, femoral fracture appears to be the biggest concern, followed by surgical difficulty, and complications with the bone graft itself. Postoperatively, subsidence appears to be the leading cause of failure. Finally, graft incorporation and femoral dislocation are also prominent areas of concern.  1.1.7  Prosthesis Subsidence Mechanisms  The precise subsidence mechanism is not clear. Researchers have attributed implant subsidence to the failure of the stem-cement interface, cement mantle creep or cold flow, cement mantle fracture, failure of the cement-morsellized bone interface, irrecoverable deformation of the morsellized bone, and the failure of the morsellized bone-cortex interface (Figure 3, Table 3). Leopold (Leopold et al 1999) theorised that the subsidence  15  mechanism may be cold flow or fracture of cement mantle. Ornstein (Omstein et al 2000) suggested that migration might be due to creep, bone resorption, or graft compression. Haddad & Duncan (Haddad & Duncan 1999) reported that some series have shown radiographic evidence of cement mantle fracture, perhaps leading to subsidence and Linder (Linder 2000) assumed that the initial migration of the implant was due to "after-compaction" of the graft and that the gradual reduction in migration is due to the fibrous reinforcement of the graft. In a study of the stem-cement interfacial debonding of a primary THA, Yoon and colleagues (Yoon et al 2001b) found that constructs with smooth, tapered, and unbonded stems experienced compressive failure of the cement mantle. In contrast, the roughened or pre-coated stem systems that were rigidly bonded failed in shear at the cement-bone interface.  Previous researchers have not explicitly measured motion at each of the five possible locations (Figure 3), so it is difficult to assess the relative motion at each location. In addition, inadequacies in measurement techniques have compounded the difficulty in assessing subsidence location. For example, it was shown earlier that there is controversy regarding the amount of subsidence of the Exeter stem within the cement. Also, the role of the distal plug has not been examined. It is possible that the entire construct subsided as a consequence of distal plug subsidence and, without adequate reference points, that subsidence may have been mistakenly reported as stem subsidence with respect to the cement mantle.  16  Table 3 Reported possible causes of subsidence.  Cement-stem debond/no bond Piccaluga et al 2002 Leopold et al 2000 Ornstein et al 2000 van Biezen et al 2000 Karrholmetal 1999 McLaren etal 1998 Masterson et al 1997 Meding etal 1997 Masterson & Duncan 1997 Malkani et al 1996 Verdonschot & Huiskes 1996 Elting etal 1995 Franzenetal 1995 Gie etal 1993 Harris 1992  Cement creep  Cement fracture  Cement-graft failure  Graft packing/ crushing/resorption  Graft-cortex failure  X  x  X X  X X X  X X  X  X  X X X X  X  X X  X  X  X  X  X  X  X  X  X  X X X  X X  X  X  Stem-cement interface Cement creep/failure  Cement-morsellized bone interface Morsellized bone yielding  Morsellized bone-tube interface failure  Figure 3 Proposed subsidence mechanisms for the impaction allografting construct.  17  X  1.1.8  Critical Variables Affecting the Performance of RTHA with IA  As outlined earlier, there are several complications and causes of failure associated with impaction allografting, and while the IA technique is fairly well established, inconsistencies in published results may be attributed to any number of variables. The aim of this section is not to present an exhaustive literature review to determine the relationship between variables and results, but to point out that it is often not known why inconsistent results exist. Indeed, while researchers may speculate on those variables that affect IA performance, the extent of the individual and combined effects of these variables is not known.  Table 4 outlines the possible variations in RTHA with IA in terms of components, patients, and surgical technique. The stem options used for femoral revisions are similar to those for primary hip replacement in terms of stem material, size, shape, surface finish, and coating. The cement and the allograft are also considerations when impaction allografting is used. The allograft is perhaps the most studied component of RTHA with IA. Researchers have speculated on the effects of altering the graft source (origin on skeleton), type (cancellous or cortical), pre-treatment (i.e. preloading) (Giesen et al 1999), fat and water content (Voor et al 2000), final porosity (Rice et al 1988), grading (size, shape, density) (Brodt et al 1998, Tanabe et al 1999, Brewster et al 1999), and adding binder (Speirs 2001) or particles to the MB (Brewster et al 1999). It is thought that optimising the properties of the allograft will diminish the occurrence or magnitude  18  of implant subsidence and loosening. These variables will be examined in more detail in the following section.  Table 4 Component, technique, and patient variables associated with RTHA with impaction allografting on the femoral side  R T H A Femoral Variations Components stem material  titanium, Co-Cr steel, stainless steel  shape  collar/collarless, straight/tapered/double-tapered, short/long, size  surface finish  polished/roughened  coating  none/hydroxylapetite coating/porous coating/proximal coating  cement  type, viscosity, mantle thickness, penetration  allograft  source, type, pretreatment, fat and water content, grading (size, shape, density), addition of binder or particles mantle thickness, porosity, irradiation  distal centralizer  present/absent  distal plug  fixation strength  Patient postoperative care  magnitude and frequency of weightbearing, amount of mobilization  selection  degree of bone stock deficiency  femur  taper angle, inside surface finish  Technique fixation  cement/cementless (press-fit), cementing technique  impaction ostectomy  routine/RIG, amount of force, level of distal impaction routine/transtrochanteric/extended osteotomy  Patient selection and postoperative care may also affect the results presented by researchers. Considerable variation exists in patient inclusion or exclusion criteria, and while some studies extensively detail the criteria (Leopold et al 1999, Ornstein et al 2000, Stulberg 2000, van Biezen et al 2000), some studies do not (Linder 2000). This could have a profound effect on the results of the series. Furthermore, because of the time-  19  dependent nature of the revision hip construct, the magnitude and frequency of weightbearing and the amount of mobilisation will have long term repercussions on the success of the surgery.  The surgical technique in impaction allografting has a more significant effect on the outcome of the surgery than it does for primary hip replacements. As outlined previously, RTHA with IA is rarely performed without cement and it is thought imperative to achieve a consistent cement mantle regardless of the cementing technique (Haddad & Duncan 1999). The amount of force exerted to impact the allograft may also be vital to the outcome of the RTHA surgery. Both Brewster (Brewster et al 1999) and Tanabe (Tanabe et al 1999) found that the morsellized bone stiffness increased with compaction energy, but did not correlate a stiffer MB to a more stable construct or a more successful surgery.  1.2  Review of Materials and Bone Remodelling Research  Research has been conducted on both the material properties of the components used in impaction allografting and the biological implications of IA, but there has been limited study on the structural properties of the IA construct. If one were to approach the failure scenarios associated with impaction allografting from a purely mechanical standpoint, the solution would be difficult to achieve but straightforward: provide a stable environment for the implant where stresses are transferred through the RTHA components such that neither material nor interfacial failure occurs. However, there is a biological component to hip arthroplasty that dramatically affects the structural properties of the joint over time.  20  The following section provides an introduction to the properties of bone in general and those of materials used in impaction allografting in particular.  1.2.1  Introduction to the Structure and Properties of Bone  Bone is anisotropic, heterogeneous, viscoelastic, and adaptive. It is a living tissue that provides protection of internal organs, structural support for the body, attachment sites for muscles and ligaments, storage of salts (calcium), and a continuous supply of new blood cells. Generally, 60% of the total weight of bone comprises inorganic components, typically Calcium and Phosphorus in the form of calcium hydroxylapatite. Type 1 collagen accounts for approximately 30% of the total weight, while the remainder is water. The relative density, p*, distinguishes cortical bone (p* > 0.7) from cancellous (0.05 < p* < 0.70). Blood cells and blood platelets are formed between the spicules of cancellous bone and in the medullary cavity. The architecture of both cancellous and cortical bone varies from a random collagen alignment to a highly oriented structure depending on the function of the bone. The result is anisotropic material properties.  The femur, a typical long bone, comprises an inner layer (endosteum) and an outer layer (periosteum) that are separated by Haversian canals that run longitudinally along the length of the bone (Figure 4). The periosteum is a layer of dense connective tissue that covers the surface of a bone. The outer layer of the periosteum is extremely dense and contains a large number of blood vessels. The inner layer is more cellular in appearance and contains osteoblasts and fewer blood vessels. The ratio of cortical to trabecular bone varies depending on the location. Proximally, there is only a thin cortical shell  21  surrounding the spongy region that makes up the femoral head, neck, and trochanters. More distally, the cortex increases thickness to reach a maximum at the longitudinal midpoint. Concurrently, the cancellous bone thins until there is a medullary canal throughout the diaphyseal region.  Lacuna  Conelitui;  Figure 4 Typical construct of a long bone (Anatomy & Physiology 1993)  The combination of anisotropy and heterogeneity in bone leads to distinctive mechanical properties. Increases in elastic modulus and yield strength accompany an increase in bone density (Cellular solids: structure & properties 1988). However, there is also an associated decrease in the plastic collapse region that follows yielding. Gibson and Ashby (Cellular solids: structure & properties 1988) also reported that the elastic  22  modulus of bone varies with the cube of bone density, while the compressive strength varies with the square of bone density. These properties also vary with loading rate.  Fundamental to the understanding of the replacement hip is to understand the dynamic nature of bone. The strain-adaptive bone remodelling theory suggests that bone reacts to overloading by altering its structure. This is manifested by new bone formation/loss in the form of alignment of trabeculae, increased density, increased cross-sectional area, a redistribution of density, increased periosteal width, decreased endosteal diameter, or increased cortical width. The cell that is responsible for the formation of bone is known as an osteoblast. The formation of bone takes place in three stages by the action of osteoblasts. A meshwork of collagen fibres is deposited in connective tissue, followed by the production of a cementing polysaccharide. Finally the cement is impregnated with minute crystals of calcium salts. The osteoblasts become enclosed within the matrix as osteocytes (bone cells) (Oxford Reference: Concise Medical Dictionary). In a similar manner, bone responds to underloading by resorption.  1.2.2  The Material Properties of Bone and Other THA Materials  The properties of metals, ceramics, polymers, and biological tissues typically used in total hip arthroplasty have been well characterised. Huiskes and Verdonschot (1997) summarized the Young's Moduli and static strength for several of these materials (Table 5). It was reported earlier that bone is anisotropic and viscoelastic. More specifically, the Young's Modulus of cortical bone is 15-20 GPa in the axial direction, but only 11.5 GPa in the transverse direction (Hayes & Bouxsein 1997). Furthermore, bone exhibits  23  mild viscoelastic behaviour as the tensile strength and modulus increases with strain rate (Cellular solids: structure & properties 1988).  Table 5 Indicative values of Young's moduli and static.strengths for most materials and interfaces used in THA reconstruction (Huiskes & Verdonschot 1997). C o C r alloy Titanium Acrylic cement (PMMA) UHMWPE Cortical bone  Young's modulus Static strength 800-1000 G P a under tension 200-220 G P a 800-1500 G P a under tension 100-130 G P a 100 M P a under compression 2-3 G P a 25-40 M P a under tension 1 GPa 20-30 M P a under tension 20-50 M P a under tension 15-20 G P a 150-200 M P a under compression 500-1500 M P a 3-10 M P a under compression  Cancellous bone Fibrous tissue 1 MPa Metal-acrylic cement interface  5-8 M P a under shear 5-10 M P a under tension 30-50 M P a under shear 2-4 M P a under shear 7-10 M P a under tension  Hydroxylapatite-bone interface Acrylic cement-bone interface  1.2.3  The Material Properties of Morsellized Bone (MB)  The characterisation of morsellized allograft bone is still in development and the procedures are not yet standardised. Brodt (Brodt et al 1998) performed triaxial compression tests to 30% axial strain at 5 confining pressure levelsfrom0.276 to 0.552 MPa. The stress-strain curves showed two linear regions - the first initially stiff, followed by a more compliant region. The average Young's Modulus for the pre-crush phase and the Poisson's Ratio were 100 MPa (51.67 - 152.4 MPa) and 0.2 (0.14 - 0.25) respectively. It is important to note that the apparent axial moduli of both regions were linear functions of the confining pressure and that the particle size had no effect on material properties.  24  Tanabe (Tanabe et al 1999) performed quasistatic uniaxial compression tests on impacted MB from 4 different bone mills. The stiffness (at 0.2 strain) was much lower than that of a triaxial confined compression test and ranged from ~ 0.25 MPa to 3.75 MPa. Of particular note was the conclusion that the stiffness increased with both the number of compactions and with a broad distribution of particle sizes.  Considering the graft as a porous, permeable solid filled with fluid under dynamic loading, Giesen (Giesen et al 1999) performed confined compression creep tests of impacted (pre-conditioned) MB. This study found that the mean permeability was 8.82-10' m /Nsandthe compressive modulus was 38.7 MPa. The theory hypothesised 12  4  by Giesen and colleagues was that the total strain is a summation of the strain from fluid exudation (£f) and the strain from particle settling (s ). The Sf was found to correlate well s  with the quasilinear biphasic model delineated by Mow et al (Mow et al 1980). This suggests that morsellized allograft bone can be thought of as a visco-elastoplastic material due to the nonrecoverable deformation associated with the settling of the bone particles.  The effects of fibrous ingrowth on the mechanical properties of the graft were studied by Tagil & Aspenberg (Tagil & Aspenberg 2001). Compression tests of recently impacted grafts and grafts that had been penetrated by fibrous tissue growing in between the graft trabeculae were performed. The result was that the compressive strength nearly doubled after 4 weeks of ingrowth (1.7 MPa versus 2.9 MPa).  25  Voor and colleagues (Voor et al 2000) performed uniaxial confined compression tests to study influence of fat and water content on the properties of MB. At 1.09 MPa stress, the strain was 30.9% and the confined modulus was 8.0 MPa. The rate of consolidation (the 5  2  speed at which the load is shifted from the liquid to solid phase) was 2.2-10" cm /s and the steady-state creep rate was 1.9%. Better properties were achieved with lower fat content (23.1%, 9.6 MPa, 3.4-10" cm /s, 0.9%), but there were no significant differences 3  2  in mechanical properties when only the water content was reduced. Brewster (Brewster et al 1999) examined the graft properties as a function of grading, normal load, and compaction. MB properties increased with increased normal load, shear strain (strain hardening), and compaction energy. Resistance to shear was a function of particle friction and interlocking. The shear strength of MB at maximum strain (9.5%) and 95 kPa normal stress was 120.36 kPa.  1.2.4  Bone Graft Viability, Resorption, and Morbidity in RTHA with IA  Morsellized bone used in impaction allografting may become viable, incorporated, remodelled, resorbed, or it may become morbid and inert. A graft is said to be viable when it is capable of living an independent existence. Often, a viable graft is one that has undergone some amount of trabecular or cortical healing. This is often the fundamental goal with regards to the biological success of the operation. More specifically, repair and incorporation or remodelling is referred to as healing. That is, trabecular remodelling is the formulation of a trabecular network extending from the endosteal cortex into the cement along the directions of the predominant stresses. In this case, the graft has  26  changed from morsellized slurry into a pattern of trabeculae running obliquelyfromthe endosteum to the cement along the principal stress lines. Alternatively, trabecular incorporation is the trabecularization of the graft without any distinctive orientation, or any change in the graft that is not remodelling. Cortical healing is the thickening of the cortex. A thinned out or defective cortex regaining normal cortical structure and thickness is said to have healed. Finally, resorption is the loss of substance through physiological or pathological means. In the case of bone, when underloading is experienced, osteoclasts work to resorb calcified bone.  In reviewing the histological reports of impaction allografting retrievals and biopsies, Haddad (Haddad & Duncan 1999) reported that researchers found bone to remodel in three zones: an inner zone (bone cement, fibrous tissue, necrotic trabeculae, and some remodelling), middle zone (viable trabecular bone), and outer zone (viable cortex and periosteum). In general, new bone formation follows vascular invasion. Brewster (Brewster et al 1999) noted that MB allowed for revascularization and incorporation into the host bone, but could be inhibited by impaction. Linder (Linder 2000) found that a viable cortical shell formed around the grafted area, but that some of the graft remained even after eight years. This graft was embedded in dense fibrous tissue, forming a composite structure that has been characterised by Tagil & Aspenberg (Tagil & Aspenberg 2001). Thus, with time it is possible to have cortical healing, trabecular incorporation, trabecular remodelling, resorption, no change, or a heterogeneous mixture of living and necrotic tissue. The neomedullary contents then become viable trabecular  27  bone, a composite of graft particles and fibrous tissue, necrotic graft with no vascular invasion, and a fibrous tissue membrane (Linder 2000).  The fibrous tissue membrane has inferior mechanical properties (Table 5) and may result from excessive stem micromotion. Small interfacial movements also contribute to abrasive wear, cement fracture, and interface breakdown (Karrholm et al 2000). However, there may be some advantage with micromotion. Elting (Elting et al 1995) found that bone formation requires cells that will form bone, a scaffold for osteoconduction, a bony bed, a blood supply, stability, continual micro-mechanical forces (via micromotion), and bone morphogenic and growth factors. It appears that while some micromotion is desirable, stem movement of less than 150 um is ideal for avoiding fibrous tissue development (Pilliar et al 1986).  Researchers have reported comprehensive investigations into bone remodelling in hips reconstructed with impaction allografting and have found variable results. In a study of Gruen zones in 21 IA hips, van Biezen (van Biezen et al 2000) found that 4% showed no change, 6% showed resorption, 50% showed trabecular remodelling, 6% showed trabecular incorporation, 33% showed cortical repair and trabecular remodelling, 1% showed cortical repair and incorporation, and 25% were unable to be studied due to radiographical limitations (Table 6). In a similar study of 14 hips, Linder (Linder 2000) found that 36% showed cortical healing, 7% showed trabecular remodelling, 64% showed trabecular incorporation, and 14% showed no change. Care must be applied to the comparison of the two studies because one reports findings as a percentage of zones,  28  while the other reports as a percentage of hips. However, some interesting trends can be noted in the comparison. First, there was a higher incidence of "no change" in the Linder study, corresponding to a shorter follow-up duration. This is consistent with the timedependent nature of bone incorporation and remodelling. Further, the results for trabecular healing are consistent with the belief that bone graft first incorporates into the host bone and then remodels according to the principal strains. Results from the van Biezen study suggest that there was sufficient time for most of the bone grafts to remodel, while there has been significant incorporation in the Linder study but perhaps not sufficient time for remodelling. Again however, care must be exercised when generating conclusions from these studies because of the significant effect of confounding variables on the results. For example, the van Biezen study used only severely defective femora, but the inclusion criteria for the Linder study may have been different. In summary, there appears to be tremendous variability in the bone remodelling capabilities associated with the impaction allografting technique (Weidenhielm et al 2001), which is further exacerbated by differences in inclusion criteria, surgical technique, and clinical assessment (Leopold & Rosenberg 2000). Table 6 Radiographic evaluation of RTHA with IA: comparison of two series.  Number of Hips Min Follow-up Duration (months) Mean Follow-up Duration (months) Max Follow-up Duration (months) Results  No change Resorption Trabecular Remodelling Trabecular Incorporation Cortical Healing Unable to be Studied  Van Biezen et al (2000) 21 41 60 85  Linder (2000) 14 3 29 99  Percentage of Zones  Percentage of Hips  4% 6% 50% 6% 33% 25% of zones  14% not reported 7% 64% 36% 7% of hips  29  1.3  Summary  The previous sections showed that there is a need for improvement of the impaction allografting procedure and outlined the mechanical and biological components that need to be addressed when attempting to rectify existing complications and failure scenarios. Confounding variables and measurement inconsistencies have made it difficult to isolate key variables affecting the structural response of IA constructs. The goal of revision hip surgery is to have immediate rigid fixation of the implant, restoration of bone stock, and positive bone remodelling. Critical to its success, the impaction allografting technique has an immediate structural foundation and a long-term biological foundation that may enhance the structural rigidity of the system. The technique relies on cement for its primary stability and morsellized bone for cortical restitution. Previous investigations have generally been limited to clinical reviews and research on material properties. The research on material properties tended to focus on the mechanical behaviour of the MB or cement, but did not correlate findings to the structural response of the construct. Further, the effect of distal plug fixation and stem surface finish has not been studied.  Objectives  The IA technique is promising, but reviews show a number of surgical complications and failure scenarios, of which subsidence is a fundamental concern. These reviews generally provide insight into the performance of the procedure under varying conditions, but only attempt to theorise on the underlying cause of failure. Furthermore, the method by which  30  load is transferred from the stem to the cortex is not clear. Therefore, the objectives of this research are to:  1. Develop an in vitro model of the impaction allografting construct. 2. Determine the effect of material composition, distal plug fixation, and stem surface finish on: a. Structural stiffness (micromotion) b. Stem Subsidence 3. Establish the location of substantial subsidence. 4. Find the strain state as a function of material composition, distal plug fixation, stem surface finish, and cranial-caudal position.  This research examined the structural aspect of the femoral side of the IA complex while maintaining consideration for the biological component of the procedure. Although biologic healing was not accounted for, structural responses that may affect cortical repair were considered (i.e. micromotion). By examining morsellized bone as part of a system comprising cement, bone, and prosthesis, the composite may be optimised in the future so as to provide adequate initial implant stability, an ideal environment for bone regrowth, and a conduit to effectively distribute hip reaction forces and moments to prevent femoral bone resorption. This will help to diminish the postoperative failure scenarios of subsidence, nonunion, aseptic loosening, and dislocation.  31  2.0 Methods 2.1  Materials  Preparation  The following section describes the development of an in vitro axisymmetric model to meet the research objectives outlined in Chapter 1. The model was of the form outlined by Yoon et al (2001). However, while Yoon and colleagues examined the failure response of a stem-cement-"bone" structure, this research explored the structural responses of a stem-cement-morsellized bone-"bone" structure instrumented with strain gauges.  Although the axisymmetric model does not exactly represent the clinical conditions, it provided the ability to isolate variables when examining their effect on the system. This type of model has been used in other research work (Grimm et al 2003; Gross & Abel 2001; Huiskes 1980) and the dimensions of the model are similar to those published in previous studies: 30 mm outside diameter (OD), 5 mm cortex, and 140 mm length (Gross & Abel 2001; Hedia et al 1996; Yoon et al 1989; Huiskes & Boeklagen 1989). Care was taken to ensure that both the model and the surgical instruments were comparable to their real-life counterparts.  2.1.1  Prosthesis Development  The stems used in this study were modeled after those most commonly used in vivo. In an illustrative case of the impaction grafting technique, Mikhail and Weidenhielm maintain that, "The implant we use is forged from cobalt-chromium alloy, double tapered  32  and polished." (Mikhail & Weidenhielm 1998). The double tapered, highly polished, CoCr stem is known as the Exeter stem (Stryker Howmedica, Allendale, NJ, USA) and is regularly used by orthopaedic surgeons at many centres (Leopold et al 1999), including Vancouver Hospital. Drawings from the Howmedica website for a standard stem (offset = 44 mm, head diameter = 26 mm, length = 160 mm) were used to dimension the model stem: length = 160 mm (plus an additional 28 mm non-tapered proximal section), distal diameter = 4 mm, proximal diameter = 25.4 mm. Whereas the Exeter stems are double tapered at 4° and 2°, the stem models in this study had an axisymmetrical circular cone shape with a uniform 4° taper angle. The model stems were fabricated from a plaincarbon steel and polished to 2000 grit (Ra = 0.19 urn). The proximal non-tapered section was tapped to allow fixation to the load frame. Distal centralizers are used clinically to ensure central positioning and to provide a space for the stem to subside without point loading on the distal cement mantle (Figure 5). Due to limited availability, distal centralizers were not used on all test specimens.  Figure 5 Exeter stem typically used in the impaction allografting technique. Note the centralizer covering the distal tip of the stem [Howmedica (http://www.revisionhip.com/frames/restexe.html)].  33  2.1.2  Cement Preparation  Surgical Simplex P, a radiopaque bone cement (methyl methacrylate liquid combined with a mixture of polymethyl methacrylate, methyl methacrylate-styrene-copolymer and barium sulfate powder) from Stryker Howmedica (Allendale, NJ, USA), was used in all trials. For each trial, an experienced orthopaedic surgeon prepared the cement in a manner outlined by the manufacturer in the "Instructions for Use" bulletin distributed with the cement package. The entire contents of the powder packet and liquid vial were emptied into a mixing assembly (Prism II Vacuum Mixing Cartridge, DePuy, Warsaw, USA) and stirred thoroughly for a total of 1 - 2.5 minutes. While some ambiguity exists in the instruction bulletin, the cement manufacturer stated that, "The correct working consistency of the product.. .is best determined by the experience of the surgeon." Every effort was made to maintain inter-trial consistency and to fully emulate the clinical procedure.  2.1.3  Bone Graft Preparation  Eight fresh-frozen femoral heads and fifteenfresh-frozenfemoral condyles were used in the preparation of the bone graft (see Table 7 for specimen data). Generally, femoral heads are preferred for use in the impaction allografting (IA) procedure (Haddad & Duncan 1999; Gie et al 1993), but femoral condyles are composed of similar bone architecture and are endorsed for use in the procedure (Mikhaii & Weidenhielm 1998). Most soft tissue was removed and the specimens were cut into approximately 15 x 15 x 40 mm sections with a band saw. The sections were passed through a bone mill (Lere  34  Bone Mill, DePuy, Inc.) once with a course blade and a second time with a fine blade. In a previous study, Speirs (Speirs 2001b) found that this method produced particles ranging in size from 0.6 to 4.0 mm which is within the range observed in previous studies outlined in Section 1.2.3. The bone chips (See Figure 6) were then sealed in a plastic container and frozen until used.  Table 7 Subject information for the bone used in the current study. lit. (cm) Wt. (kg) Smoker  Specimen ID  Part  Side  Sex  Age  Race  1015  Femur Head  Left  Mat  59  Caucasian  183  64  1016  Femur Head  RighC  Male  70  Caucasian  175  61  Male  84  Caucasian  190  59  Unknown  Mat  85  Caucasian  168  73  Unknown Unknown  Unknown Yes  1017  Femur Head  Left  1019  Femur Head  RighC  1019  Femur Head  Left  M a Ic  85  Caucasian  168  73  1022  Femur Head  Left  Female  70  Caucasian  160  53  Unknown  1023  Femur Head  Right  Female  80  Caucasian  158  60  Unknown Unknown  1089  Femur Head  Left  Male  73  Unknown  Unknown  Unknown  1019  Femur Condyle  Right  Male  85  Caucasian  168  73  1 Jnknown  1019  Femur Condyle  Left  Male  85  Caucasian  168  73  Unknown  1021  Femur C o n d y l e  Right  Male  90  Caucasian  165  70  Unknown  1022  Femur Condyle  Right  Female  70  Caucasian  160  53  Unknown  1023  Femur Condyle  Left  Female  80  Caucasian  158  60  Unknown  1024  Femur Condyle  Left  Male  84  Caucasian  170  65  Unknown  1031  Femur C o n d y l e  Left  Male  72  Caucasian  188  59  Unknown  1033  Femur Condyle  Left  Male  99  Unknown  175  65  Unknown  1034  Femur Condyle  Left  Male  71  Unknown  175  65  Unknown  1034  Femur C o n d y l e  Right  Male  71  Unknown  175  65  Unknown  1036  Femur Condyle  Left  Male  93  Unknown  155  54  Unknown  1068  Femur Condyle  Left  Male  68  Caucasian  180  120  Unknown  1 069  Femur Condyle  Right  Male  67  Caucasian  180  120  Unknown  1089  Femur Condyle  Left  Male  73  Unknown  Unknown  Unknown  Unknown  1094  Femur Condvle  Left  Male  77  Unknown  178  85  Unknown  Figure 6 Photograph of morsellized bone chips (left) produced by Lere Bone Mill.  35  2.1.4  Aluminum Tubing Design  To replicate the structural characteristics of the femur, yet eliminate many potential confounding anatomic variables, aluminum tubing was used to represent the diaphysis of the femur, (Yoon et al 2001a). The inside surface of the tubing was roughened and the tubing was dimensioned to match the properties of cortical bone (section 2.1.4.2).  2.1.4.1 Inside Surface Finish Tests The inside surface of the aluminum cylinders was machined to mimic as closely as possible clinical conditions. A push-out study with morsellized bone (MB) and PMMA cement was conducted to determine the appropriate surface finish for the inside surface of the aluminum cylinders such that the cylinders provided interfacial shear strengths similar to those observed for cadaveric bone by Frei (Frei et al 2003b). Frei (2003b) also found that the morsellized bone density varied along the length of the femur. Therefore, the morsellized bone specimens were prepared with a variety of densities.  Eight aluminum tubes (3.34 cm OD x 2.84 cm ID) were cut into lengths of 5.27 cm and the inside surfaces were roughened with a surgical high-speed cutter (MIDAS-REX M2, L9217) (Paprosky et al 2001). Each tube was fitted with a 1.27 cm thick polyethylene disc (2.84 cm OD) that sat flush with the bottom of the tube. Two packages of cement (approximately 90 mL when mixed) were prepared as indicated above and three tubes were filled in a retrograde fashion using a cement gun (Mikhail & Weidenhielm 1998). The cylinder opening was sealed with a rubber gasket and pressure was maintained with  36  the cement gun until the cement had cured. The specimens were allowed to cool to room temperature and the mass of each specimen was recorded before and after adding cement.  In a fourth tube, the surgeon added and impacted the morsellized bone, just as he would fill the distal aspect of the femoral canal in the clinical impaction allografting procedure. In the other four tubes, the surgeon added approximately 17 g, 21 g, 26 g, and 33 g of MB and impacted as necessary to make the bone flush with the top of the cylinder (Figure 7). The actual masses of material in each tube are presented in Table 8. The after-testing mass of the specimens sometimes varied considerably from the beforetesting mass because liquid (water and marrow) was allowed to escape from the cylinders.  Figure 7 Prepared specimens for the "inside surface finish" test. Three specimens contained 100% cement. Five specimens contained 100% morsellized bone at various levels of impaction.  37  Table 8 Before and after conditions of specimens used in the inside surface finish tests. Specimen 5 was not tested because the MB lost all structural integrity after the distal plug had been removed. •  Specimen  1 2 3 4 5 6 7 8  *  Conditions  Pressurized cement Pressurized cement Pressurized cement MB - maximum impaction MB-no impaction MB - light impaction MB - finger pressure MB - preferred impaction  After Testing  Prior to Testing Mass of material (g)  Density (g/cm )  Mass of material (g)  Density (g/cm )  28  1.1  28  1.1  27  1.1  27  1.1  28  1.1  28  1.1  33  1.3  24  0.90  17  0.67  -  -  27  1.1  19  0.77  26  1.0  13  0.59  32  1.3  25  0.96  3  3  Each polyethylene disc was removed to allow for a 1.27 cm gap through which the cement or MB could move when load was applied. Specimens in this study were tested with an 8874 tabletop fatigue frame paired with the FastTrack 8800 multi-channel axis controller (INSTRON, Canton, MA, USA). The 8874 is capable of a 100 mm stroke, ± 135 rotation, ± 100 Nm torque, a dynamic maximum load of ± 25 kN, and a static 0  maximum load of ± 33.7 kN. After load frame calibration and balancing (see Appendix A), the specimens were mounted on its base with an alignment fixture. A 2.54 cm OD steel tool was constructed to push the material through a distance of 1.0 cm. A positioncontrolled load was applied at 1 mm/s (Figure 8), and load and position data were recorded at 50 Hz.  38  • ys  /—  Applied Displacement  Plunger  Morsellized Bone  Figure 8 Test apparatus for "inside surface finish" testing (left). Schematic of test apparatus (right). A gap was left in the bottom of the tube through which the specimen could move.  2.1.4.2  Tubing Dimensions  Axisymmetric aluminum tubing was used to represent the diaphysis of the femur. Cristofolini et al measured the mid-diaphysis region of eight femurs and found the outer diameter to range from 25 to 35 mm (Cristofolini et al 1996). The tubes used in this study had an outside diameter of 33 mm. In a study of the changes in femur dimensions following non-cemented hip arthroplasties, Robinson and colleagues (Robinson et al 1994) recorded cortex thicknesses from 5 to 11 mm along the distal half of the implant. Given that indications for impaction allografting include significant cortical thinning, but that minimal reaming is done on the medullary canal, a reasonable estimate for the femoral bone thickness is 8.5 mm. Mechanics of materials analysis suggests that to match the structural stiffness of cortical bone, the wall thickness of the aluminum tubing should be 2.5 mm:  39  If only circumferential stiffness is considered (i.e. the tube is loaded radially from the inside), the loading on the tube is analogous to that of a pressure vessel. For a cylindrical pressure vessel, the stress in the circumferential direction, a, is given by: (1)  a = pr/t  where p is the internal pressure, r is the inner radius, and t is the wall thickness of the vessel. Given Hooke's Law for linearly elastic materials (E = a/s), it follows that, s = pr/tE  (2)  Thus, for two tubes to experience the same strain with identical pressure, r,/t,E, = r /t E 2  2  2  (3)  Alternatively, if only purely axial compression of the tube is considered, the displacement, 8, of the structure is given by, 8 = PL/AE  (4)  where P is the applied load, L is the length of the specimen, A is the sectional area, and E is the modulus of the material. As before, for similar behaviour, A,E!=A E 2  2  (5)  Because cortical bone is an anisotropic material, the moduli in the axial and hoop directions are 20.0 GPa and 11.5 GPa, respectively. Therefore, when both the circumferential and axial stiffness criteria are satisfied simultaneously, a 35 mm OD diaphyseal bone with an 8.5 mm wall thickness can be represented by an aluminum tube (E = 69.0 GPa (Mechanics of Materials)) with a 33 mm OD and a 2.5 mm wall thickness. Finally, it should be noted that the aluminum tubing used in this study was entirely free  40  of cavitary defects that are routinely seen in vivo. Thus, some caution should be taken when applying the results of the study to the clinical setting (Ling 1997).  2.1.5  Strain Gauge Preparation  Four strain gauge rosettes were applied to the outer surface of each aluminum tube. Three strain gauges in a 45° or 60° pattern (strain rosette) allow calculation of normal strain in the x and y directions and shear strain in the xy direction (Mechanics of Materials) (Figure 9). The principal in-plane strains and the maximum in-plane shear strain at the strain gauge location can then be determined. Four strain gauges rosettes were applied to each tube to measure the strain as a function of loading. The rosettes used were CEA-13-125UR-120 (Micro-Measurements Group, Raleigh, NC, USA ). These are general-purpose CEA-series 45° single-plane rosettes with compact geometry. The exposed solder tab area is 2.0 x 1.5 mm. The selection of the strain gauges was based on:  •  Gage Length: 0.125 inches (most common)  •  Gage Pattern: UR planar rectangular rosette (rosettes are used when there is a multiaxial strain field and the principal strain directions are unknown)  •  Gage Series: CEA (standard foil and backing combination)  •  Resistance: 120 Ohms (compatible with equipment)  •  STC number: 13 (close to the aluminum thermal expansion coefficient of 13 ppm/°F or 22 ppm/°C)  41  This type of strain gauge is recommended for temperatures between -75 and 175 °C, strain ranges of ± 3%, and fatigue life of ± 1500 ustrain for 10 cycles. Each rosette was 5  applied as outlined by the manufacturer in Instruction Bulletin B-127-12 and Bulletin 309D. Three rosettes were placed at 5.04, 8.88, and 13.0 cm from the distal end of the tube, respectively. The fourth rosette was applied at the same position, but 180° from the middle rosette (see Figure 9). One metre of 3-wire lead wire was soldered to the rosette terminals as outlined in Bulletin 309D. Finally, the tube-strain gauge system was calibrated prior to each round of testing (Appendix A).  Figure 9 Aluminum tube used to represent the femoral diaphysis (left). Note location of rosettes A, B, and C. Rosette D was placed 180 degrees from rosette B. Right: Polished steel stem used to represent Exeter stem.  42  2.1.6  Instrument Preparation  Custom surgical instrumentation (distal plug, guide wire, proximal and distal impactor, proximal tamp, phantom stem, proximal seal, jig fixtures) was developed for use with the impaction allografting tube models. To represent the distal plug used in vivo, solid polyethylene discs (height = 20 mm, diameter = 28.5 mm) were press-fit into each Aluminum tube until flush with the bottom edge of the tube. There are several variations on the geometry of the distal plug, but its critical feature is to provide an immovable base for the duration of the impaction procedure. In fact, in Master Techniques in Orthopaedic Surgery (1998), Mikhail and Weidenheilm use a winged shape and explicitly stated that, "it is essential that the plug does not subside during the later packing/tamping process" (Mikhail & Weidenhielm 1998). To this end, surgeons often enhance the fixation of the plug by adding external guide wires through the femur. In the tube model used in this study, the plug was artificially constrained against the testing machine base such that no translation of the plug was possible.  A custom 40 cm guide wire was built from a 4.75 mm diameter steel rod. The distal 3 mm of the guide wire was tapered to facilitate entry into the hole in the distal plug. A 5 mm diameter hole was pre-drilled 10 mm deep into the centre of the plug to allow for placement of the guide wire. Custom instrumentation was also developed to represent the impactors and tamps used in the impaction allografting procedure. The distal impactor was a 20 cm steel rod with a 28 mm diameter and doubled as a slap-hammer for the proximal impactor. The proximal impactor was a 2 mm oversized phantom stem constructed from 25 mm diameter steel rod. Its overall length was 19 cm, consisting of a  43  constant cross-section region for the most proximal 45 mm, and then tapering down to 6 mm at the distal tip (4° taper angle). Both impactors were cannulated with a 5 mm hole to enable the instruments to slide over the guide wire. The proximal tamp, used for final graft impaction when the phantom is in place, was created from aluminum rod and polyethylene tubing (see Figure 10).  Figure 10 Custom impaction allografting instruments. From left: proximal cement seal, proximal tamp, distal impactor, proximal impactor (phantom stem). Front: guide wire.  A proximal seal was created as an addition to the cement mixing kit. A piece of 25 mm thick foam (closed-cell) was shaped to fit into the proximal end of the Aluminum tube (Figure 10). The seal was cannulated for the cement dispenser and contained a slit to allow exudation of fat, marrow, and excess cement.  44  The fixture to hold the test specimens in the testing machine consisted of a base (20 mm thick Aluminum disc with a 90 mm outside diameter and a 29 mm centre hole) and a top piece (12 mm Aluminum plate, 100 mm square). The base had a set-screw to secure the tube within the base and the top piece was drilled to mate to the testing machine plunger (see Figure 11).  Figure 11 Jig for specimen fixation in the testing machine: top (left) and base (right).  2.2  Test Procedure  The following section outlines the specimen preparation, structural testing, image analysis, and pull-out tests.  2.2.1  Specimen Preparation  An experienced orthopaedic surgeon (R.M.Dominic  Meek M B C h B BSc M D FRCS)  performed  25 procedures on the tube models, varying the material composition, distal plug fixation, and stem surface coating. That is, materials included either a 100% morsellized graft (MB) construct, a 100% cement (CE) construct, or a mixture of both materials as used in  45  a standard impaction allografting procedure (IA). After altering the material composition, the specimens were left with the distal plug fully fixed within the tube, or had the plug completely removed from the tube. Finally, the stems on three specimens were coated with a silicon-based mould release in addition to the regular 2000 grit polishing. The mould release coating served to prevent bonding between the cement and the prosthesis. Table 9 displays the conditions of all 25 specimens. The three variable conditions examined in this study represent the extreme boundaries possible. As the first iteration of the impaction allografting tube models, it was important to establish the limits reached by variables that are fundamental to the structural behaviour of the construct.  Table 9 Testing matrix.  Polished Stem /: Polished Stem / No Plug Fixed Plug  Coated Stem / Fixed Plug  Coated Stem / No Plug  Impaction Allografting (IA)  6  3  2  1  100% Cement (CE)  3  3  0  0  100%) Morsellized Bone (MB)  4  3  0  0  For both the IA and MB conditions, the aluminum tube was affixed to the preparation table with a clamp and the guide wire was threaded into the distal plug (Figure 12). A l l gauges and wiring were shielded from over-splash of the graft with plastic sheeting. The surgeon alternately ladled graft into the tube and impacted the graft with the distal impactor until there was a 25 mm distal graft mantle (Figure 12).  46  Figure 12 Left: Aluminum tubes used for impaction allografting models. Photograph shows guide wire in place. Centre: Schematic of photo. Right: Schematic of distal impaction.  The proximal impactor was subsequently used to create a neo-medullary canal within the tube that was the size of the phantom stem (see Figure 13). The proximal tamp was used in combination with the proximal impactor to achieve a consistently high level of impaction even in the most proximal region of the construct. The surgeon's preferred impaction force was used for all procedures.  Figure 13 Left: Creation of the neo-medullary canal via impaction with the proximal impactor. Centre: Schematic of proximal impaction. Right: Completed neo-medullary canal.  47  Following graft impaction, the stem was introduced into the canal to complete the M B construct, while cement was added to the IA construct before stem insertion. One package of cement was prepared as outlined above and injected into the neo-medullary canal in a retrograde fashion (Figure 14). After filling the canal, the proximal end of the tube was closed with a foam seal while pressure was maintained with the cement gun. Five minutes following addition of the binder to the P M M A monomer powder, the stem was slowly inserted into the construct and held until the cement set (approximately 9 minutes). The final constructs consisted of a 1-5 mm graft mantle around the stem and, in the case of the IA procedure, a theoretical 2 mm cement mantle (Haddad & Duncan 1999). Figure 15 illustrates the final impaction allografting model.  Cement  Figure 14 Left: Retrograde filling of the neo-medullary canal. Right: Schematic of cement filling.  48  Figure 15 Left: Finished impaction allografting tube model. Right: Schematic of impaction allografting tube model.  For the 100% cement construct, two packages of cement were mixed in the cement cartridge prior to retrograde introduction into the aluminum cylinder. As in the IA case, pressure was maintained on the cement and the stem was inserted at the five-minute mark. After the models were constructed, each specimen was weighed and the final position of the stem with respect to the tube was recorded.  2.2.2  Structural Testing  Successive uniaxial cyclic compression was applied to the tube models at 1 • body weight (BW) for 250 cycles, 2»BW for 250 cycles, and 3»BW for 500 cycles. These tests were designed to determine the structural response of the constructs. Force-displacement data  49  were applicable to micromotion and subsidence (research objective 2), while the forcestrain data related to the construct strain state (research objective 4).  In vivo  loading of the THA structure arises from hip joint reaction forces and moments  caused internally by muscle contraction or externally by the transmission of the ground reaction force and can be calculated analytically or measured experimentally. Analytically, hip joint reaction loads are derived from kinematic data (limb segment position, velocity, and acceleration), kinetic data (ground reaction force), and anthropometric data (limb segment mass, length, centre of gravity, etc.) of an instrumented subject. In a typical gait cycle, the resultant load is characterised by a peak at heel strike and another at toe-off. Using the inverse dynamics approach with a linksegment model, several researches have calculated the resultant hip joint force and found it to be between 3.0 and 5.8 times body weight (Bergmann et al 1993).  Instrumented prostheses have allowed researchers to measure in vivo loading at the hip joint. Verdonschot et al (Verdonschot et al 1991) described the variation of the hip joint loads throughout a gait cycle (1.1 seconds). They found very low anterior-posterior (AP) forces, up to 1»BW medial-lateral (ML) forces, and up to 3»BW cranial-caudal (CC) forces (all from a point at the center of the femoral head). Figure 16 illustrates typical loading patterns recorded at the hip joint for an in vivo instrumented prosthesis. Bergmann and colleagues (Bergmann et al 1993) also recorded in vivo prosthesis loading and found mean peak reaction forces ranged from 2.8'BW to 4.8»BW for level walking at various speeds.  50  Figure 16 (A) Hip joint forces determined from an in vivo instrumented prosthesis. (B) and (C) Distribution of the bending stresses at the medial side of the frontal plane of the prosthesis (Huiskes & Verdonschot 1997).  Although in vivo loading at the femoral head is a multi-axial combination of compression, torsion, and bending, researchers often resolve the resultant forces and moments at the hip joint into simple axial point loads. In general, a common rule-ofthumb used throughout the biomechanics community is that the femoral head experiences a reaction compressive force approximately three times the subject's body weight while walking. In 1989 Yoon (Yoon et al 1989) used a 3000 N point load to test model hip arthroplasties, while Gross (Gross & Abel 2001) applied a similar load at an 11° angle to the vertical plane. In models of implanted hip prostheses, Baleani and colleagues (Baleani et al 1999) assumed axial joint reaction forces of 1.2 - 5.2 • BW for normal activities occurring 35,000 - 2,500,000 cycles per year (see Table 10). With regards to the failure modes associated with the impaction allografting procedure, simple axial  51  compression was the most relevant. It is the single largest load component and directly contributes to subsidence. The successive loading regimes applied in this study represent a condensed version of a patient's postoperative progression from aided walking to full weight bearing.  Table 10 Load history assumed for implanted hip prostheses (Baleani et al 1999) Activity Walking  '  Jogging  Cycles/year  Specific activity/speed  Supposed occurrence (%")  M a x load  2.5* 1 0  L e v e l walking/1 km h"  20  282  L e v e l walking/3 km h"  1  60  324  L e v e l walking/5 km h"'  20  429  L e v e l j o g g i n g / 5 km  50  484  30  496  J o g g i n g upstairs  10  51 5  J o g g i n g downstairs  10  384  W a l k i n g upstairs/slow  20  333  Walking  70  356  W a l k i n g upstairs/fast  10  386  6.4* 1 0  6  5  1  1  L e v e l j o g g i n g / 5 km"  Ascending  Stairs  D e s c e n d i n g Stairs  4.2* 1 0 _ 4  3.5* 1 0  4  2  upstairs/normal  W a l k i n g downstairs/slow  20  374  W a l k i n g downstairs/normal  70  387  W a l k i n g downstairs/fast  •0  432  S itting/rising  7.2* 1 0  4  R i s i n g from a chair  100  123  Jolting  ] .8* 1 0  3  Stum bling  100  720  The occurence of activities within 1 year for active patients. The annual occurence is roughly into specific activities. T h e last c o l u m n reports the m a x i m u m  (%BW)  separated  hip joint load c o r r e s p o n d i n g to each activity.  Once prepared, specimens were mounted into the load frame for structural testing (Figure 17) with shims to correct for any stem misalignment. The strain gauges were nulled and an initial 50 N compressive base load was applied to the specimen. Wavemaker Editor™ software was used to apply the following load: relative ramp to 0.5-BW; cycle from 01-BW for 250 cycles; relative ramp to 1BW; cycle from 0-2-BW for 250 cycles; relative ramp to 1.5BW; cycle from 0-3BW for 500 cycles. Each cyclic regime was applied at a frequency of 1 Hz. Although body weight was assumed to be 75 kg, limitations in the  52  I N S T R O N feedback control limited the cycling regimes to approximately 70% o f the desired values. Figure 18 shows a representation o f the applied loading conditions.  500 Cycles"  -2500  Figure 18 Representation of the loading conditions applied to the tube models.  53  A dedicated measurement PC was used to record applied axial load, axial prosthesis displacement, and tube strain as a function of time. All data were collected at 5 Hz and synchronized with a digital trigger. Load and displacement were recorded via a load cell and linear variable displacement transducer (LVDT) mounted on the INSTRON. The load cell (SensorData M211-113, Sterling Heights, MI) measured force in the vertical direction and was used in both measuring and controlling the system. The LVDT was built into the load frame and measured the vertical position of the crosshead. A custom data acquisition software application was developed using Lab VIEW 6.1 (National Instruments, Austin, TX, USA) to read, process, and record strain gauge data obtained through two National Instruments SCXI-1520 8-channel universal strain gauge input modules. The Wheatstone bridge was configured in a quarter-bridge format with a 5 V excitation voltage. The software application included shunt (100 000 Ohm resistor to simulate 568.72 microstrain), gain adjust, and null features. Figure 19 illustrates the relationship among all components of the impaction allografting model constructs including strain gauging and loading.  54  i Bone Femur ug »e Rosette  Displacement, and Strain Figure 19 Schematic of test set-up. Test specimens were loaded into the load frame and subjected to cyclic compressive loading for a total of 1000 cycles at 1 Hz.  2.2.3  Image Analysis  To establish the location of subsidence (research objective 3), radiographs of each specimen were taken before and after each test. Two specimens were shot on each film with a calibration piece consisting of aluminum tubing and morsellized bone situated between the specimens to confirm consistency between radiographs. Each radiograph was taken with identical settings: 84 cm between the specimen and the source, 10 M A , 70 kVolts Peak (X-ray beam energy), 2.5 mAs (milliAmpere • seconds). The radiograph films were developed and then digitized with a Nikon Coolpix 950 digital camera (Nikon Corporation, Tokyo, Japan). The photographs were taken in a darkened room with no flash to maintain consistent exposure. Each file was of the JPEG format at a size of 1200 x 1600 pixels. The resolution was 300 x 300 dpi with a bit depth of 24. After a centre-  55  weighted average metering, the exposure was locked to F number = F/3.4 and 1/86 seconds exposure time.  2.2.4  Pull-out Tests  Following completion of the structural testing, each specimen underwent a pull-out test (Figure 20). Though this method had not been described in previous literature, it was performed to give further insight into interfacial shear strengths. The specimens were loaded into the testing machine as in the structural tests and subjected to a positioncontrolled tensile load acting on the stem at a displacement rate of 1.5 mm/s for 50 mm. Load, displacement, and time data were recorded at 5 Hz and the mode of separation was noted.  1.5 mm/s  a  Steel Stem  •  P M M A cement  • •  Figure 20 Schematic of Pull-out test.  56  Morsellized Bone Aluminum Femur Polymer Plug  2.2.5  Statistical Analysis  Statistical analyses were performed with STATISTICA 5.5 (StatSoft Inc., Tulsa, OK, USA). To determine the effect of the distal plug on stem subsidence, a 1-way ANCOVA with fixed effects was performed with plug fixation as the independent variable and subsidence as the dependent variable. Construct material (IA, CE, or MB) and the presence of the distal centralizer were held as covariates to negate their effect on the findings. To determine the effect of the distal plug on stiffness, a 2-way ANCOVA with fixed effects and 1 repeated measure was performed. The repeated measure was the 3 loading regimes in which stiffness was measured. Material and distal centralizer were again covariates.  To determine the effect of material composition on stiffness, a 2-way ANCOVA with fixed effects and 1 repeated measures factor (3 loading regimes) was performed. Only the centralizer was held as a covariate while material composition was the independent variable. Similarly, a 1-way ANCOVA with fixed effects was performed to identify the effect of material composition on subsidence. For analyses with significant main effects, a post hoc comparison was performed using the Tukey HSD for unequal N . A p-value less than 0.05 was considered significant.  57  3.0 Results 3.1  Inside Surface Finish Tests  Push-out tests were conducted to determine the cement-tube and morsellized bone-tube interfacial strengths. In these position-controlled experiments, the force required to overcome the interfacial bonding between the aluminum tube and the studied material typically had three distinct phases: an initial linear region; a nonlinearly increasing region; and then a region where the force plateaued or slightly decreased (Figure 21). Each specimen tested displayed this characteristic behaviour except specimens 5, 6, and 7, which showed an approximately constant load between 0 and 5 N during the test. The force at the end of the linear region and the peak force at the initiation of the plateau region for each of the eight specimens tested are shown in Figure 22.  ^ •o  ra  Jj  2000 1800 1600 1400 1200 1000 800 600 400 200 0 0.00  2.00  4.00  6.00  8.00  10.00  D i s p l a c e m e n t (mm)  Figure 21 A typical load-displacement plot for the inside surface finish tests.  58  12.00  P u s h - o u t Force 4000 3500  • Linear Force Peak (N)  3000  • Force Plateau Peak (N)  2500 o 2000 o u. 1500 1000 500  0 CE  CE  CE  MB  MB  MB  (1.1)  (1.1)  (1.1)  (.67)  (1.0)  (1.1)  MB  MB  (1.3) (1.3)  S p e c i m e n Material (Density in g/cm 3) A  Figure 22 Push-out force for each of the eight specimens tested: three 100% cement ( C E ) specimens and eight 100% morsellized bone (MB) specimens.  To calculate the interfacial shear strength between the material and the roughened aluminum tubing, the peak force of the load-displacement curve was divided by the area of the tube-material interface. The term interfacial shear strength refers to the nominal interfacial shear strength and denotes an average strength over the applicable area. Figure 23 shows that the tube-cement interface shear strength was approximately one order of magnitude larger than that of the morsellized bone-tube interface. With a surface area of 35.7 cm , the cement specimens had a mean interfacial shear strength of 2  0.66 MPa, while the MB specimens had a mean strength of 0.06 MPa. Specimens 5, 6, and 7 were not included in the calculations because the force required to displace the material at 1 mm/s was consolidating the material rather than displacing the entire structure.  59  0.7  Cement  Morsellized Bone  Figure 23 The tube-cement and tube-morsellized bone measured interfacial shear strengths.  3.2  Structural Testing  The following section presents the results of the compressive and tensile tests performed on the tube models.  3.2.1  General Observations  Although the model impaction allografting instruments and procedure were designed to produce a uniform 2 mm cement mantle, the resultant cement mantle was rather inconsistent. Figure 24 shows a modified radiograph of a typical impaction allografting model construct. The radiograph image has been enhanced with a decrease in the gamma value to darken the image and emphasize contrast in the lighter areas. By decreasing the gamma value from unity to 0.1, the contrast is enhanced such that only the cement is visible in the cement-bone graft amalgam. When comparing the enhanced radiograph with the original (Figure 24a and Figure 24b, respectively), the difficulty in distinguishing the cement from the bone graft in a standard radiograph is apparent. When  60  the same enhancement was applied to the M B and C E constructs, the morsellized bone was not visible (in the case of the M B constructs), but the cement was visible (in the case of the C E constructs).  Figure 24 A: Enhanced radiograph of a typical impaction allografting model. The gamma level was decreased so that the cement was distinguishable from the morsellized bone. B: Original radiograph.  61  The above images have shown the inconsistency in the cement mantle; regions of both abundant cement and deficient cement are apparent in the enhanced radiographs. In fact, visual examinations of the exterior of the model IA construct after they had been removed from the aluminum cylinder often revealed regions where the cement had penetrated the bone graft to the cylinder wall (Figure 25). Because the areas of the regions of full cement penetration were very small compared to the overall area, this phenomenon was difficult to detect on enhanced radiographs. Moreover, the twodimensional nature of radiographs reduced the chances of detection. Though not measured quantitatively, there were also inconsistencies between specimens.  Figure 25 Cement penetration through the morsellized bone to the tube wall.  The mass and stem angle for each specimen are reported in Table 11. The stem angle was defined as the angle between the cross-sectional plane of the stem and the horizontal plane. That is, if the surgeon implanted the stem exactly as designed, the stem angle would be 0°. In reality, however, the alignment of the stem was generally off by 1 or 2 degrees.  62  Table 11 Specimen conditions.  Stem Composite Bone Specimen Specimen Cement Material ;'-'Mass:(g)v; Mass (g) Mass (g) Angle (°) # • :  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25  3.2.2  IA IA IA IA IA CE CE MB MB MB IA MB IA CE CE CE MB IA IA MB IA IA CE MB IA  97 100 111 95 110  73 68 50 54 64 101 87 92 80 86 80 93 95 102 69 88 88 79 .  77 99  2.0 1.0 0.7 1.0 0.7 1.8 1.7 0.7 0.3 0.8 1.3 1.0 0.2 2.7 0.5 0.5 0.8 1.2 0.7 1.0 0.0 0.5 1.0 1.0 0.5  Load-Displacement  When a loading pattern such as that exhibited in Figure 26a was applied to the impaction allografting models, the constructs behaved in a manner similar to that illustrated in Figure 26b.  63  Effect of Material Composition Whereas Figure 26b is merely a representation of the resultant displacement curve, Figure 27 shows the actual data obtainedfromthree specimens that were typical of all specimens in their respective material composition groups (IA, MB, and CE). General inspection of these curves revealed three significant findings. First, when compressive loads were applied at 1»BW, 2»BW, and 3»BW, corresponding regions of successively increasing displacement were observed in the displacement-time curve. Second, due to the scale on the time and displacement axes, the cyclic nature of the displacement data was manifested as the apparent "thickness" of the lines in Figure 27. Finally, within each loading regime, although the peak load remained constant, there was a distinct increase in the displacement of the specimens. This may be referred to as "structural creep", but does not necessarily imply material creep.  Figure 26 A: Representative illustration of the loading pattern. B: Representative resultant stem displacement when subjected to loading pattern on right.  64  -25 -\ 0  1  1  1  1  1  1  200  400  600  800  1000  1200  Time (s)  Figure 27 Typical stem displacement curves for impaction allografting (IA), 100% cement (CE), and 100% morsellized bone (MB).  The properties of the displacement-time curve outlined above are indications of the recoverable and non-recoverable structural characteristics of the models. For each individual cycle, as load was applied and released, the stem displaced and nearly returned to its original position. In the course of one cycle, the change in force divided by the change in displacement was defined as the stiffness of the construct. Subsequent mentions of the term stiffness refer to the stiffness at the 100 cycle of a particular th  loading regime. Another property, subsidence, was used to quantify the non-recoverable displacement that the construct undergoes. Subsidence was defined as the change in displacement from the original unloaded position to the final unloaded position.  65  Figure 28 shows the stiffness of the model constructs as a function of material composition. The mean values of stiffness were plotted for each loading regime and error bars indicate one standard deviation. Throughout the entire test, the mean stiffness of the 100% cement (CE) constructs was greater than that of either the impaction allografting (IA) constructs (p = 0.005) or the 100% morsellized bone (MB) constructs (p = 0.012). There was also a significant interaction effect between material composition and loading (p < 0.0001). The CE constructs showed a trend toward decreasing stiffness with increasing load such that by the third loading regime, the difference in stiffness between the CE constructs and the MB constructs was not statistically significant (p = 0.57). The IA constructs showed a similar decreasing-stiffhess trend, although with stiffness values nearly half those of the CE constructs. In contrast, the MB constructs increased stiffness with increasing load to such an extent that they were nearly as stiff as the CE constructs in the third loading regime.  Figure 28 The stiffness of the model constructs as a function of material composition: impaction allografting (IA), 100% cement (CE), and 100% morsellized bone (MB)  66  Figure 29 shows the correlation between the density of graft used in the MB constructs and the construct stiffness throughout testing. The stiffness of the MB construct in the first loading regime (i.e. 1-BW) increased linearly with MB density (R = 0.36) (p = 2  0.16). The same trend was observed in the second (R = 0.42) (p = 0.12) and in the third 2  loading regimes (R = 0.63) (p = 0.032). No such relationships were observed for either the IA constructs or the CE constructs.  18000 A 1 ' B W * 2 * B W • 3*BW 16000  14000  E £ in m |t •4-»  12000 R = 2  10000 R =0  tn  m S  2  8000 6000  R =0 2  4000 0.6  0.7  0.8  0.9  1  1.1  1.2  1.3  1.4  Density (g/cm 3) A  Figure 29 The stiffness of the MB construct as a function of density of morsellized bone used to construct the model. Straight lines represent "best fits" for each data set.  The subsidence of the model constructs as a function of material composition is plotted in Figure 30. The MB constructs subsided considerably more than either the IA (p = 0.0002) or C E (p = 0.0005) constructs. While the subsidence of the IA and C E constructs 67  were somewhat similar (p = 0.88), they would have been much closer but for one outlier in the IA dataset. One specimen in the IA group subsided 7.77 mm, causing an increase in the mean IA subsidence from 1.21 mm to 2.03 mm.  30.00 25.00 21.85  | 20.00 OIA  o  c  0) •D  • MB  15.00  • CE  "55 •§ 10.00 2.03 0.24  5 .0 00 0 0.  Figure 30 Overall subsidence of the model constructs at the conclusion of testing.  The density of the morsellized bone in the MB constructs was correlated with overall subsidence of the construct, (R = 0.63) (p = 0.033), decreasing linearly with increasing bone density (Figure 31). Similar relationships were not detected for either the IA or CE constructs.  68  30  5  0-1 0.6  ,  ,  ,  ,  ,  ,  ,  1  0.7  0.8  0.9  1  1.1  1.2  1.3  1.4  Density (g/cm 3) A  Figure 31 The subsidence of the morsellized bone (MB) constructs as a function of the mass of bone used to construct the specimen.  Effect ofDistal Plug Fixation The fixation of the distal plug did not have a statistically significant effect on either the stiffness (p = 0.25) or subsidence (p = 0.34) of the M B specimens, nor on the stiffness of the IA and CE constructs. However, in all IA and CE constructs, the removal of the distal plug did have an effect on the subsidence of the entire construct with respect to the aluminum tube. This phenomenon is illustrated in Figure 32. Figure 32A shows the initial conditions of the loading setup. When cyclic loading at l ' B W was applied, all three model constructs behaved as they would with the distal plug in place. At 2-BW (Figure 32B), the IA and CE constructs slipped at the tube-material interface until constrained by the base of the testing machine, while the stem in the M B construct continued to subside within the material itself. At 3-BW (Figure 32C), the tube-material interface of the 100% M B construct maintained its integrity as all observed subsidence resulted from movement of the stem into the material. This behaviour was consistent for all specimens tested.  69  Figure 32 Schematic of testing with the distal plug removed. A: Initial conditions. B: The IA and C E constructs slipped at 2*BW loading. C: The MB construct did not slip even at 3«BW loading.  Effect of Stem Surface Finish  Coating three of the IA model prostheses with a silicon-based mould release to simulate the surface of the highly-polished Exeter stem did not significantly alter any of the 70  properties measured in this study (Table 12). The mean subsidence of the constructs decreased from 2.44 mm to 0.81 mm when using the mould release. However, with one major outlier (7.77 mm) in the polished stem group, and with so few samples, this effect was not a direct result of any real differences in stem treatments. Similarly, the small sample size precluded realizing any differences between the stem treatments with regards to stiffness (l'BW, 2»BW, and 3«BW) and pull-out force (see section 3.2.4). Furthermore, both the polished stem group and the coated stem group constructs experienced substantial slippage at the tube-morsellized bone interface at 2*BW when they underwent testing with the distal plug removed.  Table 12 Effect of coating prosthesis with a silicon-based mould release spray (IA constructs only).  Subsidence (mm)  Polished Stein  Polished & Coated Stem  (2000 grit)  (Si-based mould release)  Mean  Std. Dev.  2.44  2.64  10740.9  8365.4  N= '  Mean  Range  "N =  6  0.81  0.64 - 0.98  2  3871.9  9  13559.0  12944.914674.3  3  4166.4  9  8851.9  7152.89741.9  3  3 3  1«BW Stiffness (N/mm) 2-BW Stiffness (N/mm) 3*BW Stiffness (N/mm) Pull-out Force (N)  7144.3  3192.1  9  6360.1  5505.06809.0  515  174  8  678  538-818  71  3.2.3  Image Analysis  The nature of subsidence was studied from the pre- and post-test radiographs taken of the specimens. Figure 33 is representative of the behaviour of the MB constructs. On the left hand side is the radiograph of the specimen shortly after it was constructed. The image on the right hand side is the specimen following compression testing. It is apparent from visual inspection that the tube-material interface remained intact, although the stem subsided approximately 20 mm.  Figure 33 Radiograph of a typical MB specimen taken before and after testing. Left: Before testing. Right: After testing the stem had subsided without disruption of the morsellized bone-tube interface.  72  The stem subsidence of the IA and CE constructs was always less than 2 mm except in one case. The resolution of the imaging system did not allow the detection of the details necessary to determine the source of non-recoverable deformation in these constructs.  3.2.4  Pull-out Tests  Effect of Material Composition  Figure 34 shows typical force-displacement curves of the pull-out behaviour for the three different material compositions. The IA constructs displayed an initial rapid rise in force followed by a fairly linear gradual decrease. A similar initial rapid rise was observed in the C E constructs, but after the initial force peak was a more constant, but inconsistent, section. The M B constructs were distinguished by an initial rapid rise in force followed by an equally rapid decrease until the force reached near zero.  Figure 34 Typical force-displacement plots obtained in pull-out tests.  73  The pull-out forces for each specimen were averaged within each material group (Figure 35). The IA construct pull-out force was the lowest, followed by the MB and the CE constructs. A high standard deviation was calculated for the CE constructs because the force required to pull out two of the six specimens was much larger than the rest. Specimens 16 and 23 had pull-out forces of 2648 and 2710 N respectively, while the mean pull-out force for the remaining four specimens was 596 N .  2500 -,  1  2000  a> 1500 p  Figure 3 5 Mean pull-out force for the three tested material compositions: impaction allografting (IA), 100% morsellized bone (MB), and 100% cement (CE).  The mechanism for pull-out varied according to the material composition of the model constructs. For the MB constructs, the stem pulled out on its own, leaving the impacted bone completely intact (Figure 36a). For the CE and IA constructs, the stem and material  74  behaved as a single, solid unit (Figure 36b). Figure 36c shows the result of a typical IA specimen after the pull-out test. While the CE constructs pulled out as a solid cylinder, the IA constructs always broke distal to the distal centralizer such that there was approximately 1 cm of impacted bone remaining in the aluminum tube after the pull-out test. The behaviour was consistent for all 25 specimens.  Figure 36 Pull-out mechanisms. A : The stem pulled out of M B specimens. B : T h e I A a n d C E constructs pull-out as a cohesive unit. C : A sample of IA pull-out.  75  Effect of Stem Surface Finish  In terms of pull-out failure type and pull-out forces, the coated stem and polished stem groups were not substantially different (Table 12). In the pull-out test, both specimen groups behaved as a single, cohesive material: the stem, cement, and morsellized bone remained as one entity.  3.2.5 Load-Strain There was large variability in the strain gauge data and the results presented reflect trends in the data after a number of data sets were omitted. The reasons for the large variability is not fully understood but is believed to be caused by the slight asymmetry of the constructs due to the manual impaction technique. Each rosette was oriented along the aluminum tube such that strain gauges were at +45°, 90°, and -45° to the longitudinal axis (0°). The 90° and 0° directions were designated hoop and axial directions, respectively. Figure 37 shows data from one representative specimen over the course of testing. The plots are blurred due to the compression of the data on the time axis. The hoop strains,  Sh p,  strains, e  were calculated using standard strain transformation equations:  ax  iai,  00  were measured directly from the 90° strain gauge, while the axial  Yxy = 2eb  - (ea + s )  (6)  c  where a, b, and c are the directions -45°, 90°, and +45°. Then, to find the axial strain component (45° counterclockwise from the +45° gauge), e iai ax  = (£a + 6c) / 2 - (e - e ) • cos (2-45°) / 2 a  c  76  • sin (2-45°) / 2  (7)  Although the readings at many rosettes resembled the curve above, some data sets deemed to be abnormal in magnitude, offset, range, or attenuation were omitted for the subsequent summary plots. The strain readings from Rosette D (the rosette placed 180° from Rosette B) were removed from consideration, as were all the data from specimen 5.  77  Plots of the hoop and axial strain for all locations on all specimens are presented in Appendix B.  Effect of Material Composition and Cranial-Caudal Position The peak strains at the beginning of the third loading regime were compared as they represent the maximum strain state of the in vivo construct at the beginning of full weight-bearing. Hence, the first ten peak strains of the third loading regime at each rosette were averaged and referred to as the "strain" of a particular specimen at a specific location. The IA and C E constructs that were tested without a distal plug all experienced substantial slippage at the material-tube interface at 2*BW and therefore data from these specimens were not applicable to the study. Figure 38 shows the hoop and axial strains, and the applied force for a typical rosette. For each cycle, as the load was applied, the magnitude of both the axial and hoop strains increased. In the data set shown in Figure 38, it is interesting to note that on release of the load, the axial strain became positive behaviour occasionally observed in other specimens and at other rosette locations. It was common for the axial strain to be negative (compressive) and the hoop strain to be positive (tensile). Of the 48 data sets that were considered (3 rosettes on 16 specimens), 48 (100%) of the rosettes had negative axial strain, while 42 (88%o) of the rosettes had positive, hoop strain. Also, 7 of 7 (100%) of the IA constructs and 2 of 2 (100%>) of the C E constructs showed an increase in axial strain magnitude at each location from proximal to distal, but only 1 of 7 (14%>) of the MB constructs demonstrated this trend.  78  2  o  I c  s  •£ (hoop)  •e (axial)  - 1 0 0 0  g  - 1 2 0 0  £  Force (N)  Figure 38 Detail view of hoop and axial strain measurements from Figure 37 at the beginning of the third loading regime. Applied force is also plotted.  The strains at each location were averaged among material composition (Figure 39 and Figure 40). The hoop strains of the M B constructs were greater than that of the IA or C E constructs at all locations, but there was considerable variation in the mean M B data. Within each set of constructs, the magnitude of the mean axial strain was greater than the magnitude of the mean hoop strain except at the most proximal location (Table 13). In 14 of 16 (88%) specimens at the most distal location (Rosette A), the magnitude of the axial strain was greater than that of the hoop strain. Similarly, 15 of 16 (94%>) specimens at Rosette B demonstrated larger axial strain than hoop strain, while only 4 of 16 (25%) exhibited this behaviour at the most proximal location (Rosette C).  79  250  O* 200 o 150  • IA  c  "s 100  CO Q. O O X  B  MB  •  CE  J 50  0 A (distal)  B  C (proximal)  Rosette Location Figure 39 Mean hoop strain at the third loading regime for the three material compositions as a function of rosette location.  Rosette Location A (distal)  B  C (proximal)  1  Figure 40 Mean axial strain at the third loading regime for the three material compositions as a function of rosette location.  80  Table 13 Mean hoop:axial strain ratio (standard deviation in parentheses).  Material  Impaction Allografting (IA) 100% Morsellized Bone (MB) 100% Cement (CE)  Rosette A (distal)  Rosette B  Rosette C (proximal)  0.70 (0.6)  0.40 (0.1)  1.40 (0.6)  0.49 (0.3)  0.33 (0.4)  2.40 (2.8)  0.46 (0.2)  0.45 (0.01)  1.18(0.08)  Effect of Stem Surface Finish  No differences were noted in the hoop and axial strain states for specimens tested with coated and polished stems, respectively.  81  4.0 Discussion 4.1  Inside Surface Finish  To achieve a model with good fidelity, it is important to obtain an interfacial shear strength between the aluminum tube and the material similar to that in vivo.  The shear strength values of 0.66 and 0.06 MPa for the cement-tube and MB-tube interfaces obtained in this study are close to those found in previous studies on cadaveric femora. Huiskes and Verdonschot (Huiskes & Verdonschot 1997) reported the static strength of the cement-bone interface in shear to be 2 - 4 MPa and Mann et al (Mann et al 1999) found the cement-bone interfacial shear strength to be 2.25 MPa. The assessment of morsellized bone-cortical bone interfacial strength is more convoluted because of the varying nature of the morsellized bone graft with impaction and over time with loading. In an in vitro study on impaction allografting in eight cadaveric femora, Frei and colleagues (Frei et al 2003b) found the mean cement-bone interfacial strength to be between 1.1 and 1.4 MPa, whereas that of the initial MB-bone interface was found to be approximately 0.3 MPa.  The previously reported values are higher than those found in this study, which may be due to cancellous bone interlocking, femoral geometry variations, and the amount of morsellized bone impaction. Previous studies of this interface may reflect the presence of cancellous bone and femoral defects that enhance the mechanical interlocking of the interface. However, the tube models developed for this investigation replicated the common clinical situation during impaction allografting in the revision scenario, where 82  there is no cancellous bone. The aluminum tubes had constant diameter and were defectfree. In addition, the results presented here may be lower than those achieved in a full in vitro model because full impaction was not achieved in the MB models. Instead, impaction was performed solely with the distal impactor. In practice, however, both distal and proximal impactors are used and the proximal impactor, by nature of its shape, provides higher radial forces in the bone graft which manifest as normal forces at the MB-tube interface.  Finally, although attempts were made to uniformly vary the density of the MB specimens, it was difficult to create specimens between 0.67 and 1.0 g/cm . In 3  attempting to lightlyfinger-packspecimen 7 to 21 g, the surgeon overshot the target by 5 g. This suggests, and is corroborated anecdotally, that graft impaction is more of a step function than a continuous function. That is, impacted graft appears to exist in three distinct phases: not impacted (-0.67 g/cm ), impacted (-1.1 g/cm ), or well impacted 3  3  (-1.3 g/cm ). Both the surgeon's preferred impaction force and his maximum force seem to reside in the well-impacted regime. Also, the density of the specimens changed as a function of the amount of liquid allowed to drain from the graft. Table 8 shows that once the distal plugs were removed and the specimens underwent loading, both the masses and densities changed dramatically.  4.2  Structural  Testing  The experimental models in this study were developed to answer questions fundamental to the structural behaviour of the in vivo impaction allografting construct.  83  There are  limitations of the models with regards to biology, surgical procedure, geometry, materials, and loading. However, once validated, the models provide the means to identify and isolate variables relevant to the success of the impaction allografting construct and understand the system's load transfer mechanisms. The boundaries of the system were established by studying models with 100% cement and 100% morsellized bone graft in addition to the standard impaction allografting composition. Similarly, the present study investigated the impact of the distal plug on the structural integrity of the system at the limits of distal plug strength and examined the effects of prosthesis surface coating. The following section provides a discussion of the model limitations, the rationale for the experiments, a comparison of the results to previous literature, and presents a more comprehensive exploration of the results obtained in the structural tests.  4.2.1 Limitations The stiffness and subsidence results may have been affected by the inability to exactly replicate the Exeter impaction allografting system. The model IA constructs developed in this study did not substantially subside into the cement mantle as designed by the Exeter group. There were two potentially significant discrepancies between the Exeter system and the models developed for this study. First, not all specimens were fitted with a distal centralizer. Without the centralizer, the stem is thought not to have room in which to subside. However, there were very little, if any, differences between specimens with the centralizer and those without and all statistics were performed with the centralizer as a covariate. Second, the model prosthesis (mean Ra = 0.19; standard deviation = 0.019LUTI) was significantly rougher than the Exeter prosthesis (Ra = 0.03;  84  [Mitchell 2002]) and may have bonded to the cement. However, when the stem was introduced into the neomedullary canal, more than five minutes had elapsed since the cement had been mixed. At that time, the cement had exceeded the doughy, or tacky, phase of the cure cycle and was nearing the set time. Therefore, cement-stem bonding was unlikely. Furthermore, in an RSA study on the subsidence of non-polished stems used in RTHA with IA, Karrholm (Karrholm et al 1999) concluded that the effect of the stem surface on subsidence is minor.  Morsellized bone density was measured and correlated to stiffness (Figure 29) and subsidence (Figure 31), but this measurement only provided an average density throughout the construct. It did not account for local variances in density, nor did it predict the amount of cement penetration into the graft. The success of the impaction allografting arthroplasty may be affected by the graft porosity and the cement penetration into the graft. Both of these variables in turn may be related to graft density.  4.2.2  General Observations  Schematics of the impaction allografting technique generally present a distinct, consistent heterogeneous structure: prosthesis, 2 mm cement mantle, 5 mm morsellized bone mantle, and cortical bone. To examine the validity of this clinical working-assumption, radiographs were taken of each specimen following impaction allografting. The intraand inter-specimen inconsistencies found in the cement mantle in this study, consistent with those reported by Masterson et al (Masterson et al 1997), do not support representation of the IA structure in the idealized manner outlined above. These  85  variations may have contributed to variability in the load-displacement and load-strain results. A less substantial cement mantle may elicit stiffness and subsidence behaviour more similar to the MB constructs and intra-specimen inconsistencies in the cement mantle may affect local strain readings. However, the subsidence mechanism and pullout force were unaffected by cement mantle variations. In both cases, the cement mantle was sufficient to cause the stem-cement-allograft construct to behave as a cohesive unit rather than 3 separate materials with distinct interfaces. This agrees well with the hypothesis presented by Ling (Ling et al 1993): "the composite behaves more like a bone-coated implant than a solid implant inserted into bone".  Although Ling (Ling 1996) stated that allowing the cement to reach the endosteal surface of the femur indicated that impaction had been defective, the surgeon performing the procedures in this study felt that impaction was not only adequate, but may have surpassed that of in vivo cases. Nevertheless, cement was observed to have reached the inside surface of the aluminum tubes, consistent with a recent study by Frei (Frei et al 2003a). Morsellized bone particle size, cement viscosity, and cement pressurization may also affect the degree of cement interdigitization, but these variables were all controlled to replicate intraoperative conditions.  4.2.3 Load-Displacement The structural stiffness and subsidence of the impaction allografting models were determined as a function of material composition and distal plug fixation because excessive component mobility may be detrimental to the success of the impaction  86  allografting revision procedure. Mobility comprises recoverable displacement, or micromotion,  and non-recoverable displacement, or subsidence. Throughout this study,  stem micromotion (measured in mm) was normalized with respect to force and reported as stiffness (measured in N/mm). Clinically, the term subsidence is often used synonymously with the terms migration and settling and the reconstructed joint is considered unacceptable if the stem has displaced more than 10 mm. While subsidence has proven to be a concern clinically, it was thought prudent to also study micromotion because of its biological implications.  Effect of Material Composition on Stiffness and  Subsidence  It was expected that the stiffness of the model constructs would follow the trend of the respective moduli of the individual materials. Speirs (Speirs 2001a) found the modulus of morsellized bone under confined compression to be between 6 and 11 MPa, though even slight variations in the testing protocol can lead to variable results due to the viscoelastic nature of the material (Voor et al 2000; Giesen et al 1999). The modulus of PMMA cement is 2-3 GPa (Table 5). While the 100% CE construct was the stiffest construct in each loading regime, it was not overwhelmingly stiffer and certainly not 300 times stiffer as the material moduli suggest. Furthermore, the MB construct became nearly as stiff as the CE construct by the third loading regime. Hence, the model IA construct exhibited the largest micromotion at 3»BW loading, while the model CE and MB constructs were nearly equal in micromotion at the same load.  87  The behaviour of the model MB constructs may be influenced by the geometry and load history of the structure. With loading and subsequent consolidation, the confined morsellized bone particles may behave as a unit of cancellous bone rather than individual particles. In a cadaveric study of the morphology of impaction allografting, the relative density of the graft after impaction was 50% (Frei et al 2003b). This value implies that well impacted and confined bone particles have a similar make-up as cancellous bone (5 - 70%) relative density) and will behave accordingly. Therefore, when predicting the structural stiffness of constructs with different material compositions, it may be more appropriate to use the stiffness of cancellous bone (0.5 - 1.5 GPa) rather than that of morsellized bone. This value provides a more realistic comparison to constructs fabricated with PMMA cement ( 2 - 3 GPa) (Figure 28). Figure 29, relating the density of graft and the structural stiffness, also supports the argument that the modulus of morsellized bone particles increases with increased MB consolidation.  An increase in interface motion with increased force may explain the progressive decrease in stiffness observed in the CE and IA constructs. The stiffness, or micromotion, measurements were based on the stem motion with respect to the aluminum tube. Because interfacial interference inhibits material motion, the displacement of the stem within the construct would be less than if the same force was applied solely to the material. Therefore, when the tube-material interface broke down at 2-BW, an increase in micromotion was observed.  88  It was anticipated that the constructs tested in this study would subside according to published in vivo reports. Subsidence reported in impaction allografting series has varied considerably, ranging from 0 to more than 10 mm (Karrholm et al 1999), but has generally been less than 4 mm. Ling (Ling 1997) reported that earlier impaction allografting procedures without the use of cement were performed and found to have marked subsidence. Also, the cemented primary construct is generally considered to have very low subsidence. The measured subsidence values summarized in Figure 30 support the trends outlined in the literature: unacceptably high subsidence for MB constructs, low subsidence for IA constructs, and very low subsidence for CE constructs.  With regards to the MB constructs, subsidence was most likely to occur as a result of the stem sliding along the stem-morsellized bone interface and subsequently compressing the bone chips. This theory was supported by radiographs (Figure 33) and the relationship between density of morsellized bone and the amount of subsidence (Figure 31). Also, Brewster (Brewster et al 1999) found that the morsellized bone particles themselves did not break, but rearranged to form a more compact structure. Therefore, it is likely that MB subsidence was due to morsellized bone consolidation.  The 100% cement constructs showed very little subsidence and radiographical evidence was insufficient to identify the location of subsidence, but it is possible to speculate on the location of IA subsidence. Because the stem did not subside into the cement in the CE construct, it is unlikely that it did so in the IA construct. This precludes stem-cement interface failure and cement creep/fracture from being highly influential on the  89  subsidence of the stem. Perhaps the length of testing (1000 cycles) and the test temperature (room temperature) were insufficient for exposing the creep mechanism of the cement. Furthermore, the cement-morsellized bone interface was shown to have such thorough interdigitization as to prevent interfacial failure. Finally, the morsellized bone was shown to readily consolidate and the morsellized bone-tube interface was shown to have low interfacial shear strength. This suggests that for the IA constructs, subsidence likely occurred as a result of morsellized bone compression and slippage at the morsellized bone-tube interface. This theory is supported radiographically by the one specimen in the IA construct group that had substantial subsidence (7.7 mm). Figure 41 shows the apparent morsellized bone-tube interface failure that accompanied morsellized bone consolidation. It appears that the stem did not subside within the cement. Rather, the entire stem-cement unit (including the cement-morsellized bone mixture) slid down the tube wall. In order to achieve this displacement, the graft must have consolidated. It is significant that this specimen was the surgeon's first procedure. Perhaps there was a learning curve and the morsellized bone in this specimen was not compacted as well as subsequent constructs.  90  Figure 41 Before and after radiograph of IA subsidence outlier. Left: Before testing. Right: After testing construct has failed at the tube-morsellized bone interface.  To summarize, the construct subsidence and stiffness were plotted on a continuum (Figure 42) and trends compared to a related study by Berzins et al (Berzins et al 1996). Berzins et al examined the initial stem migration and micromotion in cadaveric femurs using the impaction allografting technique (IA), a primary uncemented hip reconstruction (UC), and a cemented construct (CE). Although the magnitudes of the results presented in the two studies are not equal, the trends relate well. This is promising for the validity 91  of the models developed in the current study because it suggests that the results of the tube models can be transferred to cadaveric models and, by extension, in vivo situations. Berzins et al, however, were unable to assess the location of subsidence in the structure. The tube models allow identification of the effect of different variables on the IA construct by limiting confounding variables.  Migration / S u b s i d e n c e  B e r z i n s (1996)  Micromotion  Current Work  B e r z i n s (1996)  Current Work  Figure 42 Comparison of relative trends obtained in this study (shown on the left side of each bar) with those presented by Berzins et al (shown on the left side of each bar).  Effect of Distal Plug  Fixation  Current impaction allografting literature, while acknowledging the importance of the distal plug during impaction, has neglected to examine the postoperative effect of distal  92  plug fixation. If the prosthesis does subside into the cement mantle, generating sufficient radial forces to stabilize the construct, one would expect that the distal plug has little bearing on the structural integrity of the construct. This premise was supported by the results of structural testing with the distal plug removed. None of the IA and CE constructs exhibited marked subsidence and none maintained structural integrity without the distal plug at 3»BW loading. These specimens all slipped at the material-tube interface. Conversely, no difference was observed between MB specimens with and without the distal plug. Considering these results, it is plausible that stem subsidence does indeed sufficiently increase radial forces. Also, the IA and CE construct behaviours suggest that the presence of cement effectively consolidated the stem-material system into a continuous functional unit.  The effect of stem subsidence on the material-tube interface is apparent when examining the push-out shear strengths. Assuming that all the force applied to the stem was purely compressive (lateral area = 0.0114 m ), the shear strength of the tube-material interface 2  can be estimated. The material-tube interface of IA and CE constructs failed at approximately 1300 N. Therefore, the interfacial shear strength was 0.11 MPa. The morsellized bone-tube interface did not fail at 2000 N , so the shear strength was at least 0.18 MPa. In comparison, earlier push-out strength tests for the tube-cement and tubemorsellized bone interfaces were measured at 0.66 and 0.06 MPa, respectively. Thus, there was a substantial increase in radial, or normal, forces as a result of stem subsidence.  93  The results presented here suggest that the fixation of the distal plug is not only integral to the impaction process, but may also be critically important in preventing subsidence. There is some evidence of the stem and composite (cement and morsellized bone) subsiding relative to the femur (see Table 3) in vivo, although the position of the distal plug was not measured in these studies. The role of the distal plug may also be extended to the effect of geometry on the system. A tapered aluminum tube would have a similar effect on blocking composite motion as a fixed distal plug. Therefore, if subsidence of the composite is to be prevented without regard for stem subsidence, either the distal plug fixation or the geometry of the tube (femur) must be adequate.  4.2.4  Image Analysis  The difficulty in distinguishing the cement from the morsellized bone and the relatively low subsidence values observed in the IA and CE constructs suggest that radiographic evaluation may not be sufficient to extract diminutive details from the impaction allografting models. Franzen et al reported that radiographs are inadequate for reliably measuring prosthetic subsidence of less than 4 mm (Franzen et al 1995). In addition, in a multi-centre study of 189 revision procedures, experienced surgeons could not distinguish the cement mantle from the impacted allograft in 10.6 - 18.9% of Gruen zones (Masterson etal 1997).  4.2.5  Pull-out Tests  Pull-out tests were performed on all specimens to gain more insight into the effect of material composition on interfacial strengths. Although tests such as these had not  94  previously been performed, it was possible to speculate on the outcome of the tests. Because of the surface roughness of the stem and the aluminum tube, it seemed reasonable to assume that the stem would be pulled out of the construct. There was some question as to the replication of the highly polished surface of the Exeter stem, but the stems used in the model were much less rough than the inside surface of the aluminum tubes (roughened with a high speed cutter). Assuming that little or no chemical bonding occurred because of the retarded introduction of the cement into the tube, the interfaces must have resisted motion purely due to static and kinetic friction. The fact that the IA and CE constructs failed at the material-tube interface suggests that the normal forces at the stem-cement interface were substantially larger than those at the material-tube interface. This may have been possible given the large prosthesis stresses developed from cement cure shrinkage as reported by Roques (Roques et al 2002). In addition, sufficient normal reaction forces may have been caused by slight stem subsidence in the CE and IA structures.  The cohesive nature of the IA structure was also described by Gie et al (Gie et al 1993) in a retrieval of an impaction allografting stem. This cohesiveness and the relatively low interfacial shear strength between morsellized bone and the aluminum tube resulted in a low pull-out force for the IA structures. With a lateral area of 0.0114 m , the calculated 2  interfacial shear strength was 0.05 MPa. Similarly, the interfacial shear strength of the CE constructs was 0.11 MPa. With regards to the MB constructs, the morsellized bonetube interface did not fail, so the interfacial shear strength was at least 0.08 MPa.  95  4.2.6 Load-Strain To validate the theory that stem subsidence substantially increases radial forces in the impaction allografting construct and to provide a generalized overview of the stresstransfer mechanisms, strain gauges were applied to the external surface of the aluminum tubes. When examining load transfer characteristics of the tube models, it is necessary to consider several concepts: load transfer by means of interfacial shear stress, equilibrium considerations for composite structures with non-constant cross-sections and different interfacial properties, hydrostatic loading of a pressure vessel, and elastic deformation of an axially loaded member. Each concept will be described and then compared to the load-strain results to identify the appropriate load transfer mechanisms.  Modelling a non-tapered femur-prosthesis system, Huiskes and Verdonschot illustrated the concept of load transfer via interfacial shear stress. They described load transfer regions as a function of cranial-caudal position along the length of the femur (Huiskes & Verdonschot 1997). On the proximal and distal ends, there was an increase in interfacial shear stress and a subsequent increase or decrease in bone loading. Therefore, depending on the length of these end regions, some strain increase at Rosettes A and C may be expected.  Huiskes and Verdonschot also modeled load transfer via a cone pushed into a cylinder for different interfacial conditions. Interfacial conditions are often thought of as bonded, loose with Coulomb friction, or loose without friction. Bonded interfaces are ideal for transferring stress through the development of shear stress; some shear stress is  9 6  developed through the friction mechanism; and no shear stress develops in the case of frictionless, unbonded surfaces. Figure 43 illustrates the difference in stress transfer mechanisms for a simplified hip model where friction is present and absent. When a prosthesis is bonded to the femur, load transfer occurs primarily by shear stress (Figure 43a). In contrast, the unbonded stem requires subsidence for equilibrium to be maintained. With subsidence, substantial compressive stress is created at the interface to equilibrate the applied load (Figure 43b) (Huiskes & Verdonschot 1997). Note that the stresses developed in the structures to equilibrate the applied load are not constant along the length of the specimen as the force carried by each component changes as the area of load sharing changes.  Figure 43 Load transfer via a straight-tapered cone pushed into a cylindrical counterpart. A: The shear stress at the bonded interface can equilibrate the applied force. B: At a smooth, press-fitted interface, equilibrium relies on the vertical component of compressive interface stress. For slightly tapered cones, a significant amount of subsidence must occur in order for the compressive stress required for equilibrium to develop. (Huiskes & Verdonschot 1997)  97  Because the morsellized bone is fluid-filled and contained, both hydrostatic stress and axial elastic deformation may contribute to the overall strain state of the tube models. For a thin-walled cylinder under hydrostatic pressure, the stress (a) is given by:  o- oop = Pr/t h  a axial = Pr/2t O radial ~0  (8)  (9) (10)  where P is the applied pressure, r is the inside radius, and t is the thickness of the cylinder. If Hooke's Law is applied to these equations,  Ehoop - (O hoop " V O axial " V O radial)/E  (11)  = Pr(l-v/2)/Et Eaxial = ( C axial " V O hoop " V O radial)/E  (12)  = Pr(l/2-v)/Et where v is the Poisson's ratio of the tube.  For an axially loaded tube, a hoop = 0  a  axia  i = -F/27irt  O radial = 0  (13)  (14) (15)  Applying Hooke's Law gives, Ehoop =  vF/2E7irt  (16)  £axial =  -F/2E7rrt  (17)  98  Overall, if both hydrostatic and axial loading are considered, Ehoop (total)  —  Shoop (hydrostatic)  ^hoop (axial-load) , -  £axial (total)  —  ^axial (hydrostatic)  ^axial (axial-load)  -  "  0  8)  (19) Effect of Material Composition and Cranial-Caudal  Position  The relationships in equations 18 and 19 can be used to help assess the nature of the strains recorded in testing. Though both hydrostatic and axial loading conditions for the tube models yield positive hoop strain, hydrostatic loading predicts positive axial strain whereas axial loading predicts negative axial strain. Most specimens showed positive hoop strain and negative axial strain. This suggests that the specimens were influenced more by axial loading.  Furthermore, axial loading theory, in accordance with equilibrium considerations, predicts that axial strain will increase along the length of specimen (in the distal direction). Both the IA and CE constructed behaved this way, but the MB constructs were more inclined to have equal strain along the length of the tube. This suggests that the MB constructs had a significant hydrostatic loading component.  The stress transfer theory discussed above for stems that have subsided (unbonded, frictionless interface) illustrates that large radial forces are developed with subsidence. These radial forces are manifested as hoop strain in thin-walled vessels. In accordance with this theory, the MB hoop strains were larger than either the IA or CE strains. It is  99  reasonable to conclude from this that the MB constructs behaved as in Figure 43b, while the IA and CE constructs behaved as in Figure 43a.  If the constructs behaved as pressure vessels, the magnitude of the hoop:axial strain ratio would be greater than one, and the opposite would be true for an axially loaded tube. More specifically, from equations 4 and 5 above, the  Sh o :£axiai 0  P  Poisson's ratio of 0.3). Similarly, from equations 9 and 10, the  ratio is 4.26 (for a £hoo :£axiai P  ratio is 0.3.  The results suggest that the specimens behaved more as axially loaded tubes at rosette locations A and B, and behaved more as pressure vessels at location C. Further, if only pure axial compression is applied to the tube, the predicted strain in the axial direction is -120 ps, while that in the hoop direction should be 40 pe. Nearly all strain readings at all locations were less than these values, suggesting that combined loading effects were present.  Once again, caution should be exercised when interpreting the strain gauge results presented in this study. The high variability in the data indicates that when attempting to produce identical conditions, the loading conditions were more complicated than anticipated. The disagreement of strain readings at rosette locations B and D indicated that symmetry was not satisfied. Specimens loaded in compression are especially sensitive to uneven loading, which may have been aggravated by angled stem placement or a small misalignment in the loading jig. Baring that, there may have been confounding factors not accounted for.  100  Regardless of the source, variability in strain observed in vitro may manifest as clinicallyobserved variability in bone remodelling. The strain adaptive bone remodelling theory proposes that bone remodels in direct relation to strain stimuli. The variability in bone remodelling reported in numerous clinical series may be partly explained by the highly complicated and variable nature of the structural state of strain.  4.2.7 Effect of Coated Stem As a precaution against stem-cement bonding, a coating was applied to the stems in three specimens. It was thought that if indeed the model stems were bonding to the cement, this effect would be exposed with a near frictionless and unbonded surface. The siliconbased mould release applied to the prosthesis is routinely used in the polymers and composites industry as an effective coating to prevent polymers from bonding to a manufacturing tool. Given the effectiveness of this product, there is no reason to suspect that the PMMA cement bonded chemically to the steel stem. However, no difference was detected in the behaviour of constructs with and without the stem coating. This suggests that the subsidence mechanism was independent of stem surface finish. This suspicion is in agreement with previous observation by Karrholm et al (Karrholm et al 1999) in an in vivo study that found that the effect of the stem surface on subsidence was minor and further suggests that had a highly polished Exeter stem been used in the current study, the results would not have been affected.  101  4.3  Clinical  Relevance  The impaction allografting technique for revision of total hip replacements is resource intensive and has had variable clinical results, with common instances of subsidence and intraoperative femoral fracture. Table 2 shows that subsidence is considered one of the more prominent failure scenarios. Furthermore, other failure mechanisms or complications, including thigh pain, hip dislocation, and re-revision may be associated with subsidence (Leopold et al 1999).  Intraoperative or early postoperative femoral fractures have been reported in nearly every published impaction allografting series. The extreme forces used in the impaction procedure often aggravate deficient femora and result in longitudinal fractures that require cerclage wiring support and additional healing time.  The RTHA with IA procedure is technically demanding and time-consuming as tremendous care and effort must be taken at each stage of the procedure. Intraoperative complications like wiring to prevent femoral fracture only add more time to the procedure. Furthermore, the supply of bone graft is limited and has the possibility of transmitting disease.  The issues of subsidence, femoral fracture, and excessive resource usage may be alleviated respectively with an alteration in material composition, a decrease in impaction, and the development of a suitable synthetic material. The tube model tests showed that a structure comprising 100% cement provided the least subsidence.  102  However, the use of this structure negates the beneficial bone remodelling effects associated with the use of morsellized bone graft. Hence, if morsellized bone is to be used in the material composition, stiffness and subsidence testing advocates the use of cement in the structure. On its own, morsellized bone does not provide substantial initial structural support. It is only when MB has been impacted and combined with cement that the initial construct is adequate for initial hip joint loading (Stulberg 2000).  Furthermore, data presented in the current study for the tube models indicate that longterm stem mobility will only be decreased with either initial stem subsidence or via artificial constraint (distal plug fixation). However, the tube models showed that immediate stem subsidence is unlikely with the use of cement. In vivo, the conditions may be different because the stem is thought to subside into the cement mantle and femoral geometry also contributes to artificially constrain the construct. Despite these differences, the clinical outcome is likely to improve if the construct is thoroughly constrained by the femoral geometry or by means of a securely fixed distal plug. Finally, to maintain construct cohesion, the cement must substantially interdigitate the morsellized bone. This study concurs with Karrholm's (Karrholm et al 1999) thoughts on decreasing subsidence with improved grafting technique and stiffer graft material. A much more substantial gain, however, may be realized with the development of a suitable synthetic alternative to morsellized bone that does not need to be impacted and can be easily introduced in the medullary canal.  103  It is apparent that intraoperative femoral fracture is a direct consequence of the impaction of morsellized bone, but the IA procedure is plagued by incompatible design goals. The frequency of fracture may be eased with a decrease in impaction forces, but an increase in subsidence is associated with a decrease in MB density. In addition, an increase in impaction may hinder osteoconduction. Therefore, the development of a viable alternative to morsellized bone, one that does not require impaction, may be a better solution.  The development of a suitable synthetic material to ease concerns about resources is initially promising. Recently, researchers have begun studying the effects of additives on the mechanical characteristics of morsellized allograft bone. Speirs demonstrated improved material properties by mixing cement with bone particles (Speirs 2001a). Haddad (Haddad & Duncan 1999) suggested that adding biomaterials to MB or substituting the MB completely might be the way of the future. Finally, Brewster (Brewster et al 1999) found that MB strength increased with the addition of bioglass particles. This suggests that other nonresorbable materials (i.e. hydroxylapatite granules or titanium granules) may be employed for revision surgery.  4.4  Summary  The 100% morsellized bone constructs experienced a combination of axial and hydrostatic loading. Radiographical evidence showed that morsellized bone-stem interface failure and graft consolidation allowed stem subsidence to occur. With  104  subsidence, there was an associated increase in stiffness, an increase in interfacial shear strength, and an increase in radial forces (hoop strains).  There was no evidence of stem bonding associated with the 100% cement constructs, though very little subsidence was observed. This suggests that these constructs functioned as a cohesive unit. This was corroborated by the similar cement-tube shear strengths values measured in pull-out and push-out tests without a distal plug. Data from the strain gauge rosettes also supported the "cohesive unit" notion: the tubes used in these constructs experienced primarily axial loading.  The impaction allografting constructs behaved similar to the 100% cement constructs. Again, there was no indication of stem-cement bonding, but subsidence was minimal. Furthermore, the interfacial shear strength measured with the pull-out test was nearly equal to that obtained from pure morsellized bone in a push-out test. This suggests that the IA construct functioned as a cohesive unit. This premise was supported radiographically with confirmation of morsellized bone-tube interface failure and morsellized bone consolidation and was further supported by load-strain data that identified axial compression as the primary loading type.  105  5.0 Conclusions and Recommendations From this study, the following conclusions may be made: •  There was not a distinct boundary between the cement mantle and the morsellized bone. Cement interdigitization was not consistent along the length of the stem and occasionally reached the tube wall.  •  The 100% cement (CE) construct was the stiffest through all loading regimes. The 100%o morsellized bone (MB) construct was initially the least stiff, but became nearly as stiff as the C E construct at the third loading regime.  •  The M B construct showed substantially more stem subsidence than the impaction allografting (IA) construct or the C E construct. Subsidence in the M B construct resulted from stem-morsellized bone interface slippage and morsellized bone consolidation.  •  Increased morsellized bone density was correlated with increased M B construct stiffness and decreased M B stem subsidence.  •  Without a distal plug, the IA and C E constructs slipped at the material-tube interface at 2»BW.  •  In pull-out tests, the IA and C E constructs failed at the tube-material interface. The M B constructs failed at the stem-morsellized bone interface.  •  There was no significant difference in any of the tests when using a 2000-grit polished stem and a stem coated in mould release.  •  The magnitude of M B strain was larger than that of IA or CE. Strain increased distally with the IA and C E constructs.  106  The common theme collected from this research is that the impaction allografting constructs behave similar to 100% cement constructs. For these constructs, the stems did not subside substantially and load transfer was predominantly due to shear. This resulted in greater axial strains at the outside surface of the tubes and suggests that the distal plug is critical to the rigidity of the overall construct. In contrast, the 100% morsellized bone constructs showed significant stem subsidence. Load was transferred radially and the subsequent morsellized bone compression manifested as large hoop stresses.  The problems associated with the impaction allografting procedure may be alleviated with optimisation of the composite material properties or with the development of a synthetic material to supplant MB. Perhaps structural testing of a synthetic composite material with the models developed for this study would yield acceptable structural properties without the need for impaction. If successful, subsequent research could focus on replacing the bone particles in the composite. Any further testing with the impaction allografting model outlined in this paper shouldfirstinvestigate the source of strain variability. Finally, testing relative motion between materials should be performed with the aid of RSA to supplement macroscopic radiographical findings. The accuracy of RSA may be as high as 0.15 - 0.6 mm (Karrholm et al 1997).  107  References 1. Oxford Reference: Concise Medical Dictionary. Martin, Elizabeth A. 4th ed. 1994. Oxford, Oxford University Press. 2. Baleani, M., Cristofolini, L, and Viceconti, M. Endurance testing of hip prostheses: a comparison between the load fixed in ISO 7206 standard and the physiological loads. Clin.Biomech.(Bristol., Avon.) 14[5], 339-345. 1999. 3. Barba, M. and Paprosky, W. G. Revision with Cementless Stem Technique. Sledge, C. B. Master Techniques in Orthopaedic Surgery: The Hip. [17b], 325-333. 1998. Philadelphia, Lippincott - Raven. 4. Bergmann, G., Graichen, F., and Rohlmann, A. Hip joint loading during walking and running, measured in two patients. J.Biomech. 26[8], 969-990. 1993. 5. Berzins, A., Sumner, D. R., Wasielewski, R. C , and Galante, J. O. Impacted particulate allograft for femoral revision total hip arthroplasty. In vitro mechanical stability and effects of cement pressurization. J.Arthroplasty 11 [5], 500-506. 1996. 6. Brewster, N. T., Gillespie, W. J., Howie, C. R., Madabhushi, S. P., Usmani, A. S., and Fairbairn, D. R. Mechanical considerations in impaction bone grafting. J.Bone Joint Surg.Br. 81[1], 118-124. 1999. 7. Brodt, M. D., Swan, C. C , and Brown, T. D. Mechanical behavior of human morselized cancellous bone in triaxial compression testing. J.Orthop.Res. 16[1], 43-49. 1998. 8. Cristofolini, L, Viceconti, M., Cappello, A., and Toni, A. Mechanical validation of whole bone composite femur models. J.Biomech. 29[4], 525-535. 1996. 9. Duncan, C. P., Masterson, E. L, and Masri, B. A. Impaction allografting with cement for the management of femoral bone loss. Orthop.Clin.North Am. 29[2], 297-305. 1998. 10. Eldridge, J . D., Smith, E. J., Hubble, M. J., Whitehouse, S. L, and Learmonth, I. D. Massive early subsidence following femoral impaction grafting. J.Arthroplasty 12[5], 535-540. 1997. 11. Elting, J. J., Mikhail, W. E., and Zicat, B. A. Preliminary report of impaction grafting for exchange femoral arthroplasty. Clin.Orthop. 319, 159-167. 1995. 12. Fowler, J. L, Gie, G. A., Lee, A. J., and Ling, R. S. Experience with the Exeter total hip replacement since 1970. Orthop Clin North Am 19[3], 477-489. 1988.  108  13. Franzen, H., Toksvig-Larsen, S., Lidgren, L, and Onnerfalt, R. Early migration of femoral components revised with impacted cancellous allografts and cement. A preliminary report of five patients. J.Bone Joint Surg.Br. 77[6], 862-864. 1995. 14. Frei, H. Research Proposal: Biomechanical changes of impacted allograft bone in vivo: a study in the rat. 4-24-2001. The University of British Columbia. 15. Frei, H., Mitchell, P., Masri, B. A., Duncan, C. P., and Oxland, T. R. Allograft impaction and cement penetration with impaction allografting: an in vitro model. 2003a. New Orleans, Orthopaedic Research Society. 16. Frei, H., Mitchell, P., Masri, B. A., Duncan, C. P., and Oxland, T. R. Mechanical characteristics of the bone graft-cement interface after impaction allografting. 2003b. 17. Gibson, L. J. and Ashby, M. F. Cellular solids: structure & properties. 1st ed. 1988. Oxford, Pergamon Press. 18. Gie, G. A., Fowler, J. L, Lee, A. J., and Ling, R. S. The long-term behaviour of a totally collarles, polished femoral component in cemented total hip arthroplasty. J Bone Joint Surg (Br) [72], 935. 1990. 19. Gie, G. A., Linder, L., Ling, R. S., Simon, J . P., Slooff, T. J., and Timperley, A. J . Impacted cancellous allografts and cement for revision total hip arthroplasty. J.Bone Joint Surg.Br. 75[1], 14-21. 1993. 20. Giesen, E. B., Lamerigts, N. M., Verdonschot, N., Buma, P., Schreurs, B. W., and Huiskes, R. Mechanical characteristics of impacted morsellised bone grafts used in revision of total hip arthroplasty. J.Bone Joint Surg.Br. 81 [6], 1052-1057. 1999. 21. Grimm, B., Miles, A. W., and Turner, I. G. Measurement of Impaction Quality and Correlation with Stability in Impaction Grafting. 2003. New Orleans, Orthopaedic Research Sociey. 22. Gross, S. and Abel, E. W. A finite element analysis of hollow stemmed hip prostheses as a means of reducing stress shielding of the femur. J.Biomech. 34[8], 995-1003. 2001. 23. Haddad, F. S. and Duncan, C. P. Impaction allografting of the proximal femur: fact or fad? Orthopedics 22[9], 855-858. 1999. 24. Harris, W. H. Is it advantageous to strengthen the cement-metal interface and use a collar for cemented femoral components of total hip replacements? Clin.Orthop. [285], 67-72. 1992. 25. Hayes, W. C. and Bouxsein, M. L. Biomechanics of Cortical and Trabecular Bone: Implications for Assessment of Fracture Risk. Mow, Van C. and Hayes, W. C. Basic Orthopaedic Biomechanics. 2nd Ed.[3], 69-112. 1997. Philadelphia, Lippincott - Raven.  109  26.  Hedia, H. S., Barton, D. C , Fisher, J., and Elmidany, T. T. A method for shape optimization of a hip prosthesis to maximize the fatigue life of the cement. Med Eng Phys 18[8], 647-654. 1996.  27.  Hibbeler, R. C. Mechanics of Materials. 2nd Ed. 1994. New York, Macmillan College Publishing Company.  28.  Huiskes, R. Stress analyses of implanted orthopaedic joint prostheses for optimal design and fixation. Acta Orthop Belg 46[6], 711-727. 1980.  29.  Huiskes, R. and Boeklagen, R. Mathematical shape optimization of hip prosthesis design. J.Biomech. 22[8-9], 793-804. 1989.  30.  Huiskes, Rik and Verdonschot, N. Biomechanics of Artificial Joints: The Hip. Mow, Van C. and Hayes, W. C. Basic Orthopaedic Biomechanics. 2nd ed[11], 395-460. 1997. Philadelphia, Lippincott-Raven.  31.  Karrholm, J., Herberts, P., Hultmark, P., Malchau, H., Nivbrant, B., and Thanner, J. Radiostereometry of hip prostheses. Review of methodology and clinical results. Clin.Orthop. [344], 94-110. 1997.  32.  Karrholm, J., Hultmark, P., Carlsson, L, and Malchau, H. Subsidence of a nonpolished stem in revisions of the hip using impaction allograft. Evaluation with radiostereometry and dual-energy X-ray absorptiometry. J.Bone Joint Surg.Br. 81[1], 135-142. 1999.  33.  Karrholm, J., Nivbrant, B., Thanner, J., Anderberg, C , Borlin, N., Herberts, P., and Malchau, H. Radiostereometric evaluation of hip implant design and surface finish. 2000. Orlando, American Academy of Orthopaedic Surgeons.  34.  Keene, G. S., Graves, S. E., Howie, D. W., Mintz, A. D., and McGee, M. A. Impaction allografting with cement or cement alone for revision arthroplasty of the femur. J Bone Joint Surg (Br) [80 (suppl 2)], 135. 1998.  35.  Knight, J . L. and Helming, C. 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[2], 159-165. 2000.  36.  Leopold, S. S., Berger, R. A., Rosenberg, A. G., Jacobs, J . J., Quigley, L. R., and Galante, J . O. 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 [8], 1080-1092. 1999.  37.  Leopold, S. S., Jacobs, J . J., and Rosenberg, A. G. Cancellous allograft in revision total hip arthroplasty. A clinical review. Clin.Orthop. [371], 86-97. 2000.  38.  Leopold, S. S. and Rosenberg, A. G. Current status of impaction allografting for revision of a femoral component. Instr.Course Lect. 49, 111-118. 2000.  110  39. Linder, L. Cancellous impaction grafting in the human femur: histological and radiographic observations in 6 autopsy femurs and 8 biopsies. Acta Orthop.Scand. 71 [6], 543-552. 2000. 40. Ling, R. S. Cemented revision for femoral failure. Orthopedics 19[9], 763-764. 1996. 41. Ling, R. S. Femoral component revision using impacted morsellised cancellous graft. J.Bone Joint Surg.Br. 79[5], 874-875. 1997. 42. Ling, R. S., Timperley, A. J., and Linder, L. Histology of cancellous impaction grafting in the femur. A case report. J.Bone Joint Surg.Br. 75[5], 693-696. 1993. 43. Malkani, A. L., Voor, M. J., Fee, K. A., and Bates, C. S. Femoral component revision using impacted morsellised cancellous graft. A biomechanical study of implant stability. J.Bone Joint Surg.Br. 78[6], 973-978. 1996. 44. Mann, K. A., Werner, F. W., and Ayers, D. C. Mechanical strength of the cementbone interface is greater in shear than in tension. J.Biomech. 32[11], 1251-1254. 1999. 45. Masri, B. A. Personal Communication. 6-18-2001. 46. Masterson, E. L. and Duncan, C. P. Subsidence and the cement mantle in femoral impaction allografting. Orthopedics 20[9], 821-822. 1997. 47. Masterson, E. L., Masri, B. A., Duncan, C. P., Rosenberg, A., Cabanela, M., and Gross, M. The cement mantle in femoral impaction allografting. A comparison of three systems from four centres. J.Bone Joint Surg.Br. 79[6], 908-913. 1997. 48. McLaren, A. C , Brown, S. G., and Lucero, E. M. Early failure of impaction bone grafting for revision total hip replacement. 1998. San Francisco, Annual Meeting of the American Academy of Orthopaedic Surgeons. 49. Meding, J . B., Ritter, M. A., Keating, E. M., and Faris, P. M. Impaction bonegrafting 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[12], 1834-1841. 1997. 50. Mikhail, W. E. and Weidenhielm, L. R. The hip: master techniques in orthopaedic surgery. Sledge, C. B. 1998. Philadelphia, Lippincott-Raven. Master techniques in orthopaedic surgery. Thompson, R. C. 51. Mitchell, P. Personal Communication. 10-20-2002. 52. Mohler, C. G. and Cowin, S. C. Revision with Cemented Femoral Component. Sledge, C. B. Master Techniques in Orthopaedic Surgery: The Hip. [17a], 309-323. 1998. Philadelphia, Lippincott - Raven.  111  53. Mow, V. C , Kuei, S. C , Lai, W. M., and Armstrong, C. G. Tl - Biphasic creep and stress relaxation of articular cartilage in compression? Theory and experiments. J Biomech Eng 102[1], 73-84. 1980. 54. Nivbrant, B., Karrholm, J., Lundblad, M., and Soderlund, P. Migration of cemented and cementless revision stems using impacted allograft: 6-year follow-up with radiostereometry . Acta Orthop.Scand.SuppI 287[49]. 1999. 55. Ornstein, E., Franzen, H., Johnsson, R., and Sundberg, M. 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 [4], 360-364. 2000. 56. Ornstein, E., Johnsson, R., Franzen, H., Sandqvist, P., and Sundberg, M. Prosthetic migration during 1.5 years after hip revision with impacted morselized allograft evaluated by RSA. Trans Orthop Res Soc , 843. 1997. 57. Paprosky, W. G., Weeden, S. H., and Bowling, J. W. Component removal in revision total hip arthroplasty. Clin.Orthop. [393], 181-193. 2001. 58. Piccaluga, F., Gonzalez, Delia, V, Encinas Fernandez, J. O, and Pusso, R. 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[4], 544-549. 2002. 59. Pilliar, R. M., Lee, J. M., and Maniatopoulos, C. Observations on the effect of movement on bone ingrowth into porous-surfaced implants. Clin.Orthop. [208], 108-113. 1986. 60. Rice, J. C , Cowin, S. C , and Bowman, J. A. On the dependence of the elasticity and strength of cancellous bone on apparent density. J.Biomech. 21, 155168. 1988. 61. Robinson, D., Hendel, D., and Halperin, N. Changes in femur dimensions in asymptomatic non-cemented hip arthroplasties. 20 cases followed for 5-8 years. Acta Orthop.Scand. 65[4], 415-417. 1994. 62. Roques, A., New, A., Taylor, A., Baker, D., and Browne, M. Measurement of residual stresses due to volumetric shrinkage in bone cement. 2002. Calgary, World Congress of Biomechanics. 63. Rosenberg, A. G. Revision Total Hip Arthroplasty: The Femoral Stem. Sledge, C. B. Master Techniques in Orthopaedic Surgery: The Hip. [17], 305-308. 1998. Philadelphia, Lippincott- Raven. 64. Slooff, T. J., Buma, P., Schreurs, B. W., Schimmel, J. W., Huiskes, R., and Gardeniers, J. Acetabular and femoral reconstruction with impacted graft and cement. Clin.Orthop [324], 108-115. 1996.  112  66.  Speirs, A. D. Thesis: Calcium phosphate cement composites in revision hip replacement. 2001b. The University of British Columbia.  67.  Stulberg, S. D. Impaction grafting: the science and clinical reality. Orthopedics 23[9], 945-947. 2000.  68. Tagil, M. and Aspenberg, P. Fibrous tissue armoring increases the mechanical strength of an impacted bone graft. Acta Orthop.Scand. 72[1], 78-82. 2001. 69. Tanabe, Y., Wakui, T., Kobayashi, A., and et al. Determination of mechanical properties of impacted human morsellized cancellous allografts for revision joint arthroplasty. J M A T E R SCI-MATER M 10[12], 755-760. 1999. 70. Thibodeau, G. A. and Patton, K. T. Anatomy & Physiology. 2nd. 1993. St. Louis, Mosby. 71.  van Biezen, F. C , ten Have, B. L., and Verhaar, J . A. Impaction bone-grafting of severely defective femora in revision total hip surgery: 21 hips followed for 41-85 months. Acta Orthop.Scand. 71 [2], 135-142. 2000.  72. Verdonschot, N., Dalstra, M., and Huiskes, R. The relevance of implant telemetry for mechanical analyses of total hip arthroplasties. Bergmann, G., Graichen, F., and Rohlmann, A. 249-258.1991. Berlin, Proceedings Book Workshop "Implantable Telemetry in Orthropaedics". 73. Verdonschot, N. and Huiskes, R. Mechanical effects of stem cement interface characteristics in total hip replacement. Clin.Orthop. [329], 326-336. 1996. 74. Voor, M. J., Nawab, A., Malkani, A. L., and Ullrich, C. R. Mechanical properties of compacted morselized cancellous bone graft using one-dimensional consolidation testing. J.Biomech. 33[12], 1683-1688. 2000. 75. Weidenhielm, L. R., Mikhail, W. E., Wretenberg, P., Fow, J., Simpson, J., and Bauer, T. W. Analysis of the retrieved hip after revision with impaction grafting. Acta Orthop.Scand. 72[6], 609-614. 2001. 76. Yoon, Y. S., Jang, G. H., and Kim, Y. Y. Shape optimal design of the stem of a cemented hip prosthesis to minimize stress concentration in the cement layer. J.Biomech. 22[11-12], 1279-1284. 1989. 77. Yoon, Y. S., Oxland, T. R., Hodgson, A. J., Duncan, C. P., Masri, B. A., and Lee, J. J . Report: Effects of stem-cement debonding and taper angle on axial failure loads in femoral implant constructs for hip replacement surgery. 2001b. Korea Advanced Institute of Science and Technology.  113  APPENDIX A: Testing Machine and Strain Gauge Calibration  114  Load Frame Calibration Before applying a standard warm-up routine, the load and position limits were set, the PID Controller was set to 10 dB, and the load cell was balanced. To verify the load cell, a 75 N dead weight was applied to the frame and compared to the recorded value. To check the LVDT, the crosshead position was measured with Vernier calipers and compared to the measured value.  Strain Gauge Calibration The aluminum tubes were loaded and unloaded in compression to compare the strain gauge readings to the predicted strains from simple mechanics of materials calculations. A custom data acquisition software application was developed using Lab VIEW 6.1 to read, process, and record data obtained through two National Instruments 8-channel strain gauge input modules. Each tube was loaded in a restraining jig and centred in the load frame. A 3/8-inch ceramic ball and lubricated aluminum platen placed on top of the tube ensured that a point load was applied to an unconstrained tube. At the "zero-load" condition, all channels were nulled and the gain was adjusted via shunt calibration with 100,000-Ohm resistors. A linearly increasing compressive load was applied to each specimen such that 3,000 N was reached in 5 seconds. This load was held for 5 seconds and then returned to 0 N over 5 seconds. Force, displacement, and strain were recorded at 5 Hz. This procedure was repeated before each round of testing. Another test was performed on each tube to assess the drift in the system. The procedure for these tests was identical to that indicated above, except that the compressive load only reached 2,000 N and the plateau was held for 20 minutes instead of 5 seconds.  115  APPENDIX B: Strain Gauge Data  116  The following are plots of hoop and axial strain (left column) and an expanded view of the beginning of the third loading regime (right column) for each specimen. Specimen 1 (r1s1): IA, no centralizer, distal plug, regular stem (Rosettes A, B, C, D) 200 -200  150  £2 (hoop) c (axial)  —  ^ —  100  -400 -600  50  -soo  0  -1000  -50  -1200 -1400  -100  -1600  -150  -1800  -200  -2000 Time |s) -E2 (hoop) -  -E (axial)  Force (N)  200 -200  150  -400  100  1 B  -600  1  50 0  Stra  e  • c2 (hoop)  I  • E (axial)  I  -50  B  50  -800 J-  0  -1000 ~Z -1200 °  -50  -1400  •100  -100  -150  •150  -200  -1600 -1800 -2000 10  -200 2000  3000  20  30  40  Time |s)  4000  Tims (s)  —•— £2 (hoop)  •  E (axial)  Force (N)  200 -200  150  -400  100  -600 -800  -1200 °  | E (axial)  -1400  -100  -1600  -150  -1800  -200  -2000 10  1000  2000  3000  4000  5000  £  -1000  i E2 (hoop)  20  30  40  Time |s)  6000  Time (8)  - E 2 (hoop)  117  •  E (axial)  Force (N)  50  Specimen 2 (r1s2): IA, centralizer, distal plug, regular stem  -200 -400 -600 -600  c2 (hoop) £ (axial)  £  -1000  «j  -1200  £  -1400 -1600 -1800 -2000  0  1000  2000  3000  4000  5000  Time (s)  6000  Tim* (I)  - E2 (hoop)  200 200  2  50  1  0  6  -50  .«  50  » z2 (hoop) • E (axial)  0  1  -400  100  100  so  -200  150  150  c  -600 -800  -1000 -1200  £  -1400  -50 -100 -100  £  jj  -1600  -150  -1800  -150  10  -200  20  3 0  40  59  Time (s)  3000 Time (s)  | —•—t2 (hoop) —•— t (axial)  Force (N)  -200 -400 -600 -800  ^ ^ i ^ ^ ^ ^ ^ ^ ^ S ^ O x ^ S ^ ^ - J ^ f j -1000  • £2 (hoop) • £ (axial)  -1000 -1200  o  -1400 -1600 -1800 -2000 0  10  20  30  4 0  50  Time (e) -E2(hoop)  200  1 i  Force (N)  200  150  150  100  100  S  50  • E2 (hoop) • E (axial)  0  £  1 tn  £ (axial)  •50  Q  I °  I 8  -100  50  -50  0 -200 -400 -600  A^A^A  A A A A A A T  —it*———is*—is*—'Is*—'ii*'—Is*——  -1800 -2000  -200 3000  Time (e)  4000  Time (a)  - E2 (hoop) —•— E (axial)  118  -1200  -1600  •200  2000  -1000  -1400  -100 •150  -150  -800  Force (N)  £  £j o  Specimen 3 (r1s3): IA, centralizer, distal plug, regular stem 200 150 100 ?  50  I  0  o  « E2 (hoop)  e -so  5)  -100 -150 -200 2000  3000  -200 -400 -600 -800 -1000 -1200 V^—V^—V~V—V~V^—V^^V—VI -1400 -1600 -1800 -2000  4000  Time |s)  - E 2 (hoop) --»—£(axia  Force (N)  200  -200 -400 -600 -800 j -1000 • -1200 £ -1400 -1600 -1800 -2000  150 100  s  50  I c  0  1  c2 (hoop) c (axial)  -50 -100 -150 -200 2000  3000  Time (s)  4000  Time (s)  - E2 (hoop) — -  200  s  o I  Force (N)  0 -200 -400 -600 -800 £• -1000 -1200 £ -1400 -1600 -1800 -2000  200  150  150  100  100 50  50 E2 (hoop) E (axial)  0  g I -50 55  E (axial)  0 -50 -100  -100  -150  -150  -200 20  -200 2000  30 Time (s)  3000 Time |s)  Force (N)  119  Specimen 4 (r1s4): IA, centralizer, distal plug, regular stem  E2(hoop) E (axial)  2000  3000  ^^V-N.  A  .A ,A ,A ,A A  Time (a)  4000  Time (•)  - E2 (hoop)  120  E (axial)  Force (N)  -200 -400 -600 -800 £ -1000 ^ -1200 -1400 -1600 -1800 -2000  S.  Specimen 5 (r1s5): IA, centralizer, distal plug, regular stem  200 200  -200  150  -400  150  100  100  50  50  0  -1000  -50  -1200  • £2(hoop) • £ (axial)  0 -50  -100  -100  -150  -150  -200  -600 -800  -1400 -1600 -1800 -2000 20  -200 0  1000  2000  3000  4000  5000  30 Time (a)  6000  Time (a)  -£2(hoop)  Force (N)  200 150  -200 -400 -600 -800 -1000 -1200 -1400 -1600 -1800 -2000  so  50  1C  0  Stra  100  -50  • £2 (hoop) • £ (axial)  •  **'  »•  -100 -150 10  -200 3000  20  30  40  Time (a)  4000  Time (a)  - £2 (hoop) • - £ (axial)  121  Force (N)  Specimen 6 (r2s1): C E , centralizer, distal plug, regular stem 200  s  Straii  400  100  100  .« E  200  150  150  600  50 50  o  * E 2 (hoop) 0  • E (axial)  -50  -50  -100  -100  -150  -150  -200  800  •  y y.-v-.y. y v. y.-.\i ;  j  1000 1200 1400  z  I 0  u.  1600 -1800 -2000 20  30  -200 3000  Time |s)  4000  Time (8)  - E 2 (hoop) — • - E (axial)  Force (N)  200 200  150  400 100 600 |  • E2 (hoop) • E (axial)  50  I  0  1  -50  c  800  y y y y v y y y y M  «  1600 1800  -150 -200  3000  2000 20  30 Time (s)  4000  Time |s)  122  1200 1400  -100  2000  1000  z ® If o  Specimen 7 (r2s2): C E , no centralizer, distal plug, regular stem Note: not full of cement 0  200  -200  150 :  100 e u I  -600  - A - A - A - A - A - A - A - A - A - £  50 » £2 (hoop)  0  • E (axial)  C -50  1  tn  -400 -800  £  -1000  •  -1200  |  -1400 -1600  -100  -1800 .  -150 0  -200  10  ...  20  30  i  40  5C  Time (s)  3000 Time (s)  - c 2 (hoop)  •  E (axial)  Force (N) |  200 200  g  50  I  0  B |  -50  tt  50  1  ffflnBI  ;  100  100  J. c  -200  150  150  E2(hoop)  0  E (axial)  -400 -600  -150  -150  -200  -200  -1400  -1800 -2000 20  «  30 Time (a)  3000 Time (a)  I  *  -1600  « -100  •100  2 o 1 c  £  -1000  -1200 £  tn  -50  -800  E (axial)  -E2(hoop)  Force (N)  200  200  0  150  150  -200  100  100  -400 -600 -800  50 0  E (axial)  -1200 £  -50  •50  -1400  -100  •100  -150  -150  •200  -1600 -1800 -2000 0  -200 0  1000  2000  3000  4000  5000  10  20  30  40  50  Time (a)  6000  Time (a)  - E 2 (hoop) — • - E (axial)  200  Force (N)  ^-:~--- --- r  -400  100  -600  50 • E 2 (hoop) • E (axial)  0 -50  £  -1000  jj  -1400 -1600  •150  -1800  -200  -2000  Time (a)  4000 -E2(hoop)  123  -800  -1200 o  - V - V -  •100  Time (a)  r  -200  150  3000  £  l * ^ * * * * ' ' * * * * * * ! * * * * ' * * * * * ^ * * * ! ^ * * * ^ ^ * ^ ! * ^ -1000 jf  £2 (hoop)  - E (axial)  Specimen 8 (r2s3): G R , no centralizer, distal plug, regular stem 200  200  150  150  -200 -400  100  100  -600 -800  50 • E 2 (hoop)  0  J  -50  • £ (axial)  -150  -150  -200  >_J  J  _J  k  ^ >^ >_i  U  1  J T t t t t f ^  -100 -100  J  tl  g  -1000  jj  -1200  |  -1400 -1600 -1800 -2000  20  -200  30 Time {•)  3000 Time (>)  £ (axial)  -£2(hoop)  Force (N)  0 -200 -400 -600  £2 (hoop)  II  E (axial)  /  A-v* -N AMA A M/s A M M  » , i ( l  AAA  A  A* i '  -800  -  -1000  <s  4- 1 2 0 0  £  A^ A  y v v v v v v v v y  -1400 -1600 -1800 -2000  3000  Time (e)  4000  Time (I)  - E 2 (hoop) — » -  £ (axial)  F o r c e (N)  g  200  -200  150  -400  100 | > £2(hoop)  £  0  E  -50  • £ (axial)  -600  50  8  -800  •/v^^v^,V•,^,^,^  -100  i  „ 10  3000  -1200  <f  £  0  -1400 -1600  -150 -200  -1000  pl—y  v'  20  1* 30  -1800 >  » ' — » * — y 40  5C  Time (s)  4000  Time (t)  | — • — £2 ( h o o p ) — • — £ (axial)  F o r c e (N)  200 -200  150  -400  100 ? u  • £2 (hoop) • £ (axial)  -600  50  -800  I °U ^ A ^ A ^ A ^ A ^ A ^ A ^ |  -50  "  .100  °  -1600  •150  -1800  -200  -2000 30 Time (a)  -E2(hoop) -  124  *  -1200 -1400  20  Force (N)  J  -1000  Specimen 9 (r2s4): GR, no centralizer, distal plug, regular stem  125  Specimen 10 (r2s5): G R , no centralizer, distal plug, regular stem  200 -200 -400 -600 -800 £ -1000 J -1200 ° -1400 -1600 -1800 -2000  150 100 50  |_ c I  55  • i2 (hoop) > E (axial)  0 -50 -100 -150  20  •200 2000  3000  30 Time {•)  4000  Time (s)  - E2 (hoop)  126  E (axial)  Force (N)  Specimen 11 (r3s1): IA, no centralizer, no distal plug, regular stem Note: slipped at 2 B W 0 -200 -400 -600 -800 • -1000 1200 £ 1400 -1600 -1800 -2000  200 150  150  -  100  E2(hoop) E (axial)  e a |  E  1 8  50 0 -50 -100 -150 •200  20  30 Time (s)  -E2(hoop)-  127  - E (axial)  Force (N)  Specimen 12 (r3s2): G R , no centralizer, no distal plug, regular stem 200 -200 -400 -600  150 100 , 50  -800  :  -1000 -1200 i -1400 -1600 -1800 -2000  > £2 (hoop) • £ (axial)  - E2 (hoop) •  E (axial)  Force (N)  200 -200  150  -400  100  e  -600  so  -800  u • £2 (hoop) • E (axial)  I  0 -100 -150  ^WWWW^A  -200 3000  £  g  -1200 £ -1400 -1600 -1800 -2000  20 2000  -1000  30 Time (a)  4000  Time (s)  - £2 (hoop) » £ (axial)  Force (N)  200 150 100 50  MAAAAAAA/I  0  * E2 (hoop) • £ (axial)  -50 -100 -150 -200 10  20  200 400 600 800 z 1000 ~* 1200 £ 1400 1600 1800 2000  30 Time (a)  — » — E2 (hoop) —•— E (axial)  200 150  I E2 (hoop) E (axial)  I B 1 a  .  100 50 0 -50  m  -100 -150 -200 2000  3000  Force (N)  20  a  30 Time (a)  4000  Time (a)  £2 (hoop) -  128  Force (N)  0 -200 -400 -600 -800 £ -1000 ^ -1200 £ -1400 -1600 -1800 -2000  Specimen 13 (r3s3): IA, centralizer, distal plug, Si stem  2  u I  200  200  0  150  150  -200  100  100  -400 -600  50  50 • E 2 (hoop) 0  • E (axial)  -800  0  -1200 |  -50  -1400  -100  -1600  -100  -150  -150  -200 -  -1800 -2000  -200 2000  3000  4000  Time (s)  f  200  150  150  100  100  - c (axial)  F o r c e (N)  -200 -400 -600  50  50  o  1. 1a  -E2(hoop)  200  • £2 (hoop)  0  • E (axel)  c  0 -50  -50  -800  V  •v' '."  'v?  \<  Q  -1000 -1200 -1400  -100  -1600  -100  -150  -150  200  -1800 -2000 10  -200 2000  20  30  40  Time (s)  3000  Time |»)  -e2(hoop) -  129  ;  -1000  £ (axial)  F o r c e (N)  50  Specimen 14 (r3s4): C E , no centralizer, no distal plug, regular stem Note: slipped at 2BW, stopped machine at 3 B W 200  200  150  150  |  I  50  £2 (hoop) £ (axial)  0  c  Stra  -200 -400  100  100  2  0  I c 1  -600  50  -800 0  -1000 -1200  -50  &  -50  -1400 -100  -100  -150  -150  -200  -1600 -1800 -2000  -200 0  1000  2000  3000  4000  5000  6000  Tlma (a)  - £2 (hoop) —•— £ (axial)  200 150  s  E. c  c  55  200  0  150  -200 -400  100  |  100  .2  Force (N)  50 0  _—I  £2 (hoop) £ (axial)  f  50  -600  -A— A — A — A—A-  °  -50  -100  -100  -150  -150  -200  -800  £  -1000  ^  -1200  °  -1400 -1600 -1800 -2000 20  -200  30  Time (a)  3000  Time (a)  - £2 (hoop) - »-£ (axial)  200 150 100  • £2 (hoop) • £ (axial)  f Ic 5  5)  200  - V V V V V W \ )  50 0 -50  400 600 800  £  1000  ^  1200  °  1400  -100  1600 -150  1800  -200  2000  2000  3000  Time (a)  -E2(hoop) -  130  Force (N) I  Specimen 15 (r3s5): C E , no centralizer, distal plug, regular stem  200  200  0  150  150  -200  100  100  50  • £2 (hoop) * £ (axial) -50  -400 -600  50  (—  -800  0  -1000  -1200 u-  -50  -1400  •100  -1600  -100  -150  -1800  -150  -200  -2000  -200 2000  3000  Time (s)  -E2(hoop) -  131  z 8 0  Force (N)  Specimen 16 (r4s1): C E , no centralizer, no distal plug, regular stem Note: stopped at 2 B W  132  Specimen 17 (r4s2): G R , no centralizer, no distal plug, regular stem 200 -200  150  -400 !00 ?  50  • E2 (hoop)  £  • t (axial)  J  2  55  -600 -800  £  0  -1000  ^  -50  -1200  |  -1400 -1600  -100  -1800  -150  -2000 -200 1000  2000  3000 Time  4000  5000  20  30 T i m e (»)  6000  (I)  - £ 2 (hoop)  £ (axial)  200  F o r c e (N)  0 -200  150  -400  100  -600 50  0 • £2 (hoop) -50  • £ (axel)  -800  -  -looo  £  -1200  £  -1400 -100 -1600 -150  -1800 -2000  3000 Time  Time  4000  (•)  133  (s)  Specimen 18 (r4s3): IA, no centralizer, distal plug, regular stem  s  & £ c  1  tn  200  200  150  150  100  100  200 400 600 800 z 1000 0 1200 u. 1400 1600 1800 2000  50  50 • E2(hoop) • E (axial)  0 -50  0 -so -100  -100  •150  -150  -200  -200 3000  Time (a)  4000  Time (a)  - £2 (hoop)  Force (N)  200  - 0  150  - -200 -400  100  -600  g" 50  • E2 (hoop) • E (axial)  u i o c C -50 at  •w  \.V  \J  \J \J \J  \J  1000  2000  3000  --1800  (  10  20  30  40  Time (a) | —•— C2 (hoop)  134  |  -1200 £  - -1600  4000  Time (s)  ~  -1000  - -1400  -100 -150 -200  \,  -800  £ (axial)  Force (N) |  50  Specimen 19 (r4s4): IA, no centralizer, no distal plug, regular stem Note: slipped at 2 B W  200 200  -600  g  50  u  1£ I  -400  100  100  s  -200  150  150  > c2(hoop)  0  ' E (axial)  I c J  50  -800  0  -1000  -50  -1200  in  -150  iJ  -1400  -50 -100  j;  1  -100  -1600  -150  -1800 -2000  -200 10  -200  20  30  40  Time (t)  3000 Time (•)  —•— E 2 (hoop) -  E (axial)  Force (N)  200 200  150  400  100  600 50  • E 2 (hoop) • E (axial) £  -50  800  0 -50  w—v^—^—w—w—M  -100  •  1200  o  1400 1600  -150  1800  -200  2000  3000  5000  2000  Time (•)  6000  Time {•)  - E 2 (hoop) —•— E (axial)  135  Force (N)  J  1000  Specimen  20 (r4s5): G R , no centralizer, distal plug, regular stem  200  200  150  150  •  ,  100 B .ii  B c  I  55  50 • E2 (hoop) * £ (axial)  0 -50 -100  ^  •150  ]  -200 0  1000  2000  3000  4000  5000  10  . 30  2C  ..  . 40  Time (s  6000  Time (a)  - E2 (hoop) - •— E (axial)  136  Force (N)  50  -200 -400 -600 -800 £ -1000 <f -1200 o -1400 -1600 -1800 -2000  Specimen 21 (r5s1): IA, centralizer, no distal plug, Si stem Note: slipped at 1BW 200  0 -200 -400 -600 -800 • -1000 -1200 | -1400 -1600 -1800 -2000  150  -oo  I • E2 (hoop) • E (axial)  |  50 0  G -50  -100 -150  mm  •200 0  1000  2000  3000  4000  5000  Tims |s)  6000  Time |s)  - E (axial)  150  Force (N)  0 -200 -400 -600 -800 : -1000 -1200 i -1400 -1600 -1800 I -2000  — I  100 60 £2 (hoop) E (axial) -150 -200 10 3000  20  30  40  Time (s)  4000  Time (s)  —•—c2 (hoop)  »— E (axial)  Force (N) I  200 150 100 E  | I  c B  a  50 * tl (hoop) • E (axial)  0 -50 -100 -150 -200 3000  Time (s)  4000  Time (s)  -z2 (hoop)-  137  - e (axial)  Force (N)  50  Specimen 22 (r5s2): IA, centralizer, no distal plug, regular stem Note: slipped at 1BW & 2 B W 200  200  150  150  100 £  -600  5  • E2 (hoop)  o  -800  J  -50  • c (axial)  C 50  -1600 -1800  -200  -2000 30  20  -100  -200  -1200  -1400  -150  I -  2000  3000 Time  Time  4000 (a)  - E2 ( h o o p ) -  138  £  S ^ S w l .,«„ |  0  -100  -150  -200 -400  50  50  I  '  100  50  (a) F o r c e (N)  £  Specimen 23 (r5s3): C E , centralizer, no distal plug, regular stem Note: slipped at 2BW, stopped at 2 B W Specimen 1: Principal Strains  Principal Strain Direction  l» *p|  0  1000  2000  3000  4000  5000  6000  1000  2000  Specimen 1: Principal Strains  1000  2000  3000  4000  3000  4000  5000  6000  5000  6000  5000  6000  5000  6000  Principal Strain Direction  5000  6000  1000  2000  Specimen 1: Principal Strains  3000  4000  Principal Strain Direction  150 100  •30  1000  2000  3000  4000  5000  6000  1000  2000  Specimen 1: Principal Strains  1000  2000  3000  4000  3000  4000  Principal Strain Direction  5000  6000  1000  139  2000  3000  4000  Specimen 24 (r5s4): G R , no centralizer, no distal plug, regular stem 200  e o I c a  150  150  100  100  0 -200  piflnfl  -  :  —  50  50 0  -400  -600 -800  s  • £2 (hoop)  -1000  •  • E (axial)  -1200  £  -50  -1400 -1600  -100  -1800 -150  -2000  -200  Time (s)  3000  Time (s)  - £2 (hoop) —•— E (axial)  Force (N)  0 -200 -400 -600  ^VXAAAAAAA/"  » £2(hoop) • E (axial)  -800  £  -1000 -1200  £  -1400 -1600 -1800  0  1000  2000  3000  4000  5000  Time (a)  6000  Time (a)  - E 2 (hoop) • — E (axial)  200  0  150  -200 -400  100 50  • £2 (hoop) • £ (axel)  Force (N) I  *C  i  -y  ~*  V  -600  **  -800  z  0  -1000  ^  -50  -1200  °  -1400  -100  -1600 -150  -1800  -200  2000  3000  -2000  4000  Time (a)  - E2 (hoop) —» - £ (axial)  200 400 600  • E2 (hoop)  ./V \ / \ / \  vA/V  • E (axial)  800  Q  1200  £  1400 1600 1800 2000 20 3000  30  Time (a)  4000  Time (a)  - £ 2 (hoop)  140  Force (N)  z  1000  Specimen 25 (r5s5): IA, centralizer, distal plug, Si stem  e  200  150  150  100  100  c 1  -200 -400 -600  50  50  .2  §  200  • E 2 (hoop) • E (axel)  0 -50  -800  0  •1000 -1200 i  •50  -1400  -100  -1600  -100  -150  -150  -1800  -200 10  -200 2000  3000  20  30  40  50  Time (s)  4000  Time (s)  - £ (axial)  Force (N)  0 -200  inn  -400  o <  150  -600 -800  o  o  • £2 (hoop) • E (axial) —  I>  Strain (micro)  200  H H HHlAlBi  :  £  -1000  •  -1200  °  -1400 -1600 -1800  -150 -200 1000  2000  3000  4000  5000  Time (s)  6000  Time (a)  - E 2 (hoop) - - E (axial)  Force (N)  0  200 150  -200  100  -400  -600  50  -800  • E2 (hoop) ' E (axel)  0 -50  -1200 , -1400  -100  -1600  -150  -1800 -200  0  1000  2000  3000  4000  5000  -2000  Time (s)  6000  Time |s)  - E 2 (hoop) - £ (axial)  141  j  -1000  Force (N)  

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.831.1-0079037/manifest

Comment

Related Items