UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Patellofemoral osteoarthritis : characterizing knee alignment and morphology Macri, Erin Michelle 2017

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

Item Metadata

Download

Media
24-ubc_2017_november_macri_erin.pdf [ 5.43MB ]
Metadata
JSON: 24-1.0356572.json
JSON-LD: 24-1.0356572-ld.json
RDF/XML (Pretty): 24-1.0356572-rdf.xml
RDF/JSON: 24-1.0356572-rdf.json
Turtle: 24-1.0356572-turtle.txt
N-Triples: 24-1.0356572-rdf-ntriples.txt
Original Record: 24-1.0356572-source.json
Full Text
24-1.0356572-fulltext.txt
Citation
24-1.0356572.ris

Full Text

PATELLOFEMORAL OSTEOARTHRITIS:  CHARACTERIZING KNEE ALIGNMENT AND MORPHOLOGY by  Erin Michelle Macri  B.Sc., Simon Fraser University, 1997 M.P.T., The University of British Columbia, 2006 M.Sc., The University of British Columbia, 2013  A DISSERTATION SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Experimental Medicine)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  October 2017  © Erin Michelle Macri, 2017  ii  Abstract Patellofemoral osteoarthritis (OA) is an under-recognized medical condition that may progress to generalized knee OA. It is highly prevalent and is associated with knee pain and disability. Patellofemoral alignment, tibiofemoral alignment or trochlear morphology may influence the prevalence, onset or worsening of patellofemoral OA. Therefore, my main objective in this dissertation was to use magnetic resonance (MR) imaging to quantify the relationships between knee alignment or morphology and prevalent patellofemoral OA.   This dissertation is comprised of four studies. First, I systematically reviewed the associations among knee alignment or morphology with patellofemoral OA presence, severity, onset or worsening. Lateral patellar tilt, lateral displacement, patellar height, frontal plane knee alignment, and a shallow sulcus were all associated with OA prevalence and severity.   Second, I examined the relationship between key alignment and morphology measures and early patellofemoral OA, one year following anterior cruciate ligament reconstruction. Lateral displacement and a shallow sulcus were associated with early patellofemoral OA.  Third, I evaluated alignment and morphology in the Framingham community cohort, a population-based cohort. I established reference values in knees without patellofemoral OA or pain. I then evaluated dose-response patterns of the full sample. The odds of prevalent patellofemoral OA rose monotonically with greater lateral patellar displacement and a shallower sulcus. Odds rose for patellar tilt in both directions, i.e. increased medial or lateral tilt.    iii  Fourth, using an upright, open-bore MR scanner, I developed reproducible methods to evaluate three dimensional knee alignment in standing. I then examined differences in alignment from supine to standing positions, and differences in alignment in supine and standing in individuals with patellofemoral OA and matched controls. Key measures of malalignment were greater in standing then supine. In knees with patellofemoral OA, the patella was more laterally tilted, laterally displaced, and proximally displaced compared to controls, and the tibia was more externally rotated.   Taken together, these studies suggest that knee alignment and trochlear morphology may influence the risk of patellofemoral OA.     iv  Lay Summary Arthritis of the knee caps is an under-recognized yet common medical condition that may progress to generalized knee OA. It is associated with knee pain and disability. Knee alignment and/or bony shape at the knee joint may be important features related to patellofemoral OA. Therefore, the purpose of my thesis was to use magnetic resonance imaging (MRI) to describe the relationships between knee alignment or bony shape and OA of the knee caps.   I completed four studies within this thesis that contribute to this research field. Through these studies, I have characterized key alignment and bony shape measures that are related to OA of the knee caps. I have shown that these measures may be associated with very early OA through to more severe OA. I have also shown that measuring alignment in standing, compared to lying down, may be more clinically important in detecting and quantifying malaligned knees.    v  Preface Chapter 2. A version of this material has been published as Macri EM, Stefanik JJ, Khan KM, Crossley KM. (2016). Is tibiofemoral or patellofemoral alignment or trochlear morphology associated with patellofemoral osteoarthritis? A systematic review. Arthritis Care and Research. 68(10): 1453 – 70. I developed and registered the study protocol in collaboration with my three coauthors (Josh Stefanik, Karim Khan, Kay Crossley). I developed the search strategy in consultation with a research librarian at UBC (Charlotte Beck). I screened titles, extracted data, and assessed quality of included studies independently. Josh Stefanik also screened titles, extracted data and assessed study quality independently. We finalized study inclusion and ensured accurate data extraction and quality assessment by comparing results after conducting the independent work. I interpreted the results and wrote the manuscript with intellectual input from all coauthors. Since this was a systematic review, ethics approval was not required.  Chapter 3. A version of this material has been published as Macri EM, Culvenor AG, Morris HG, Whitehead TS, Russell TG, Khan KM, & Crossley KM. (2017). Lateral displacement, sulcus angle and trochlear angle are associated with early patellofemoral osteoarthritis following anterior cruciate ligament reconstruction. Knee Surgery, Sports Traumatology, Arthroscopy. Published Online First: doi: 10.1007/s00167-017-4571-1. This was an ancillary study of an existing cohort based in Melbourne, Australia and managed by Adam Culvenor within the research laboratory of Kay Crossley (The University of Melbourne and The University of Queensland, Australia). Radiographs were read by Kay Crossley and Adam Culvenor and results provided to me. MR images were provided to me and I independently evaluated patellofemoral alignment and morphology. Trevor Russell developed a software   vi  program for me to use for assessing alignment. Tim Whitehead and Hayden Morris performed surgery on study participants. I conducted statistical analyses, interpreted results, and wrote the manuscript with intellectual input from all coauthors. Ethical approval for this study was granted by The University of Melbourne (ID: 1136167) and The University of Queensland (ID: 2012000567 and ID: 2013001448) Human Research Ethics Committees.   Chapter 4. A version of this material has been published as Macri EM, Felson DT, Zhang Y, Guermazi A, Roemer FW, Crossley KM, Khan KM, Stefanik JJ. (2017) Patellofemoral morphology and alignment: reference values and dose-response patterns for the relation to MRI features of patellofemoral osteoarthritis. Osteoarthritis and Cartilage. Published Online First: doi: 10.1016/j.joca.2017.06.005. This project took place in Boston MA at Boston University and Northeastern University under the supervision of Josh Stefanik and David Felson. The Framingham Community Cohort is a longitudinal cohort managed by David Felson. Ethical approval for this study was provided by the institutional review board of Boston University Medical Center (ID # H-22674). I developed the study design and methodology with intellectual input from Josh Stefanik, David Felson, Karim Khan, Kay Crossley, and members of the RESCU team (Research Evaluation and Support Core Unit). MR images were provided to me to independently evaluate knee alignment and morphology. Protocols for standardizing alignment and morphology measures were developed in collaboration with Josh Stefanik. MRI structural outcomes were read and scored by Ali Guermazi and Frank Roemer. Statistical methods were developed in collaboration with Yuqing Zhang.  I conducted alignment and morphology measures, statistical analysis, interpretation and manuscript writing with intellectual input from all coauthors.   vii   Chapters 5 & 6. A version of this material has been accepted as Macri EM, Crossley KM, d’Entremont AG, Hart HF, Forster BB, Wilson DR, Ratzlaff CR, Walsh AM, Khan KM. (2017) Patellofemoral and tibiofemoral alignment in a fully weight bearing upright magnetic resonance scanner: repeatability and feasibility. Journal of Magnetic Resonance. Imaging Published Online First: doi:10.1002/jmri.25823. This study took place at the University of British Columbia. Ethical approval for this study was provided by the UBC Clinical Research Ethics Board (certificate number H13 – 01993) as well as Vancouver Coastal Health Research Institute Operational Review (number V13-01993). I developed the study protocol in collaboration with Kay Crossley, Harvi Hart, and Karim Khan, and in consultation with Agnes d’Entremont, Bruce Forster, David Wilson, Andrew Yung, Honglin Zhang, Trudy Harris, Jen Patterson and Chuck Ratzlaff. I conducted participant screening and recruitment, collected all clinical and demographic data, and led image acquisition with MRI technicians (Jen Patterson, Trudy Harris, Laura Barlow, James Zhang, Alex Mazur, Jeanette Procter). Harvi Hart assisted with a portion of image acquisition and clinical data collection for quality assurance. Charlie Goldsmith provided statistical support. I conducted statistical analysis, interpreted results, and wrote the manuscript with intellectual input from all coauthors.   On July 11, 2017, I verified that all internet links provided within this thesis are active.    viii  Table of Contents Abstract .......................................................................................................................................... ii Lay Summary ............................................................................................................................... iv Preface .............................................................................................................................................v Table of Contents ....................................................................................................................... viii List of Tables .............................................................................................................................. xix List of Figures ............................................................................................................................ xxii List of Abbreviations ............................................................................................................... xxiv Acknowledgements ....................................................................................................................xxx Dedication ................................................................................................................................ xxxii Chapter 1 Introduction....................................................................................................................1 1.1 The burden of knee OA................................................................................................... 2 1.1.1 Prevalence ................................................................................................................... 2 1.1.2 Physical impact ........................................................................................................... 3 1.1.3 Economic impact ........................................................................................................ 5 1.2 The contribution of the patellofemoral joint to knee OA is under-recognized ............... 7 1.2.1 Compartmental patterns of knee OA .......................................................................... 8 1.2.2 Patellofemoral OA is under-diagnosed ..................................................................... 10 1.2.3 Patellofemoral OA is clinically important ................................................................ 11 1.2.4 Patellofemoral OA is a research and public health priority ...................................... 11 1.3 Patellofemoral joint anatomy and function ................................................................... 12 1.3.1 Patellofemoral alignment .......................................................................................... 14 1.4 What is OA? .................................................................................................................. 17   ix  1.4.1 Cartilage .................................................................................................................... 17 1.4.2 Bone .......................................................................................................................... 19 1.4.3 Other joint features ................................................................................................... 20 1.5 Diagnosing knee OA ..................................................................................................... 22 1.5.1 Knee OA ................................................................................................................... 22 1.5.2 Patellofemoral OA .................................................................................................... 26 1.5.3 The limitations of radiographic methods .................................................................. 28 1.5.4 Magnetic resonance imaging (MRI) as a diagnostic tool ......................................... 30 1.5.4.1 Semiquantitative scoring ................................................................................... 31 1.6 Risk factors ................................................................................................................... 34 1.7 Patellofemoral OA as a distinct phenotype ................................................................... 39 1.7.1 What is a phenotype? ................................................................................................ 39 1.7.2 Patellofemoral OA is unique from tibiofemoral OA ................................................ 40 1.7.3 Patellofemoral OA may be linked to patellofemoral pain ........................................ 41 1.8 Patellofemoral joint mechanics and OA ....................................................................... 42 1.9 Using MRI to measure patellofemoral alignment and morphology ............................. 44 1.10 Theme of dissertation and study objectives .................................................................. 45 1.10.1 Study #1: Systematic review ................................................................................. 46 1.10.2 Study #2: Alignment and morphology one year following ACL reconstruction .. 46 1.10.3 Study #3: Framingham community cohort: reference values and dose-response patterns 46 1.10.4 Study #4: The “MRI open-bore for osteoarthritis Vancouver” (MOOV) study ... 47   x  Chapter 2 Is tibiofemoral or patellofemoral alignment or trochlear morphology associated with patellofemoral osteoarthritis? A systematic review. ......................................................................48 2.1 Background ................................................................................................................... 48 2.1.1 Study aims ................................................................................................................. 49 2.2 Methods......................................................................................................................... 49 2.2.1 Study design and protocol registration ..................................................................... 49 2.2.2 Ethics......................................................................................................................... 49 2.2.3 Eligibility .................................................................................................................. 49 2.2.3.1 Participants ........................................................................................................ 49 2.2.3.2 Variables of interest .......................................................................................... 50 2.2.3.3 Outcome measures ............................................................................................ 50 2.2.4 Search strategy .......................................................................................................... 50 2.2.5 Data extraction .......................................................................................................... 51 2.2.6 Quality assessment .................................................................................................... 52 2.2.7 Statistical analyses .................................................................................................... 53 2.3 Results ........................................................................................................................... 54 2.3.1 Search strategy .......................................................................................................... 54 2.3.2 Study characteristics ................................................................................................. 54 2.3.3 Quality assessment .................................................................................................... 58 2.3.4 Axial plane patellofemoral alignment: limited evidence .......................................... 60 2.3.5 Sagittal plane patellofemoral alignment: limited evidence ....................................... 64 2.3.6 Frontal plane tibiofemoral alignment: strong evidence ............................................ 67 2.3.7 Trochlear morphology: strong evidence ................................................................... 70   xi  2.4 Discussion ..................................................................................................................... 75 2.4.1 Methodological considerations ................................................................................. 76 2.4.2 Limitations ................................................................................................................ 78 2.4.3 Clinical implications ................................................................................................. 79 Chapter 3 Lateral displacement, sulcus angle and trochlear angle are associated with early patellofemoral OA following ACL reconstruction ........................................................................80 3.1 Background ................................................................................................................... 80 3.1.1 Study aims ................................................................................................................. 81 3.2 Methods......................................................................................................................... 82 3.2.1 Study design .............................................................................................................. 82 3.2.2 Ethics......................................................................................................................... 82 3.2.3 Participants ................................................................................................................ 82 3.2.4 Surgery – ACL reconstruction .................................................................................. 82 3.2.5 Radiography: defining patellofemoral OA ............................................................... 83 3.2.6 MRI: patellofemoral alignment and trochlear morphology ...................................... 84 3.2.7 Statistical analyses .................................................................................................... 86 3.3 Results ........................................................................................................................... 87 3.3.1 Participants ................................................................................................................ 87 3.3.2 Alignment and morphology ...................................................................................... 89 3.3.3 Sensitivity analyses ................................................................................................... 91 3.4 Discussion ..................................................................................................................... 92 3.4.1 Alignment and patellofemoral OA: integration ........................................................ 92 3.4.2 Morphology and patellofemoral OA: integration ..................................................... 94   xii  3.4.3 Limitations ................................................................................................................ 94 3.4.4 Clinical implications ................................................................................................. 95 Chapter 4 The Framingham Community Cohort: a population-based study of patellofemoral alignment and morphology ............................................................................................................97 4.1 Background ................................................................................................................... 97 4.1.1 Study aims ................................................................................................................. 98 4.2 Methods......................................................................................................................... 98 4.2.1 Study Design ............................................................................................................. 98 4.2.2 Ethics......................................................................................................................... 98 4.2.3 Participants ................................................................................................................ 98 4.2.4 Image acquisition ...................................................................................................... 99 4.2.5 Patellar alignment and trochlear morphology ......................................................... 100 4.2.6 MRI-defined patellofemoral OA grading ............................................................... 102 4.2.7 Knee pain outcomes ................................................................................................ 104 4.2.8 Statistical analyses .................................................................................................. 104 4.2.8.1 Reliability ........................................................................................................ 104 4.2.8.2 Reference intervals.......................................................................................... 105 4.2.8.3 Dose-response patterns ................................................................................... 105 4.3 Results ......................................................................................................................... 107 4.3.1 Participants .............................................................................................................. 107 4.3.2 Reference values ..................................................................................................... 108 4.3.3 Dose-response patterns ........................................................................................... 109 4.4 Discussion ................................................................................................................... 116   xiii  4.4.1 Limitations .............................................................................................................. 119 4.4.2 Clinical implications ............................................................................................... 120 Chapter 5 The ‘MRI Open-bore Osteoarthritis Vancouver’ (MOOV) study part I: 3D knee alignment in a vertically oriented open-bore scanner differs in standing compared to supine....122 5.1 Background ................................................................................................................. 122 5.1.1 Study aims ............................................................................................................... 124 5.2 Methods....................................................................................................................... 124 5.2.1 Study design ............................................................................................................ 124 5.2.2 Ethics....................................................................................................................... 125 5.2.3 Recruitment ............................................................................................................. 125 5.2.4 Participants .............................................................................................................. 127 5.2.5 Demographic data collection .................................................................................. 129 5.2.6 Imaging ................................................................................................................... 130 5.2.6.1 Conventional scanner: high-resolution images ............................................... 130 5.2.6.2 Upright scanner: low resolution images ......................................................... 132 5.2.6.3 Image processing ............................................................................................ 136 5.2.7 Statistical analyses .................................................................................................. 138 5.2.7.1 Repeatability ................................................................................................... 138 5.2.7.2 Descriptive statistics ....................................................................................... 139 5.2.7.3 Inferential statistics: unadjusted within-group differences ............................. 139 5.2.7.4 Inferential statistics: mixed effects models ..................................................... 139 5.2.7.5 Two-dimensional alignment: malalignment subgrouping .............................. 142 5.3 Results ......................................................................................................................... 142   xiv  5.3.1 Participants .............................................................................................................. 142 5.3.2 Repeatability ........................................................................................................... 144 5.3.3 Patellofemoral alignment ........................................................................................ 146 5.3.3.1 Unadjusted within-group differences .............................................................. 146 5.3.3.2 Mixed effects models ...................................................................................... 147 5.3.3.3 Two-dimensional alignment: malalignment subgrouping .............................. 151 5.3.4 Tibiofemoral alignment .......................................................................................... 152 5.4 Discussion ................................................................................................................... 153 5.4.1 Limitations .............................................................................................................. 155 5.4.2 Clinical implications ............................................................................................... 157 Chapter 6 The MOOV study part II: 3D knee alignment differs between patellofemoral OA and matched controls. .........................................................................................................................159 6.1 Background ................................................................................................................. 159 6.1.1 Study aims ............................................................................................................... 161 6.2 Methods....................................................................................................................... 162 6.2.1 Statistical analyses .................................................................................................. 162 6.2.1.1 Inferential statistics: unadjusted difference method ....................................... 162 6.2.1.2 Inferential statistics: mixed effects models ..................................................... 162 6.2.1.3 Two-dimensional alignment ........................................................................... 162 6.2.1.4 Two-legged vs. one-legged alignment ............................................................ 163 6.3 Results ......................................................................................................................... 163 6.3.1 Patellofemoral alignment ........................................................................................ 163 6.3.1.1 Unadjusted between-group differences ........................................................... 163   xv  6.3.1.2 Mixed effects models ...................................................................................... 163 6.3.1.3 Two-dimensional alignment between-group differences ................................ 170 6.3.1.4 Two legged vs. one legged squat .................................................................... 171 6.3.2 Tibiofemoral alignment .......................................................................................... 172 6.4 Discussion ................................................................................................................... 173 6.4.1 Limitations .............................................................................................................. 175 6.4.2 Clinical implications ............................................................................................... 176 Chapter 7 Conclusion ................................................................................................................179 7.1 Study #1: Systematic review ....................................................................................... 179 7.1.1 Key findings and contributions ............................................................................... 179 7.2 Study #2: Alignment and morphology one year following ACL reconstruction ........ 181 7.2.1 Key findings and contributions ............................................................................... 181 7.3 Study #3: Framingham community cohort: reference values and dose-response patterns 182 7.3.1 Key findings and contributions ............................................................................... 182 7.4 Study #4: The “MRI open-bore for osteoarthritis Vancouver” (MOOV) study ......... 184 7.4.1 Key findings and contributions ............................................................................... 184 7.5 Clinical implications ................................................................................................... 186 7.5.1 Possible clinical contributors to malalignment ....................................................... 187 7.5.2 Clinical measurement of patellofemoral alignment ................................................ 189 7.5.3 Treating malalignment in patellofemoral OA – state of the art .............................. 191 7.5.3.1 Conservative approach is generally preferred over surgical ........................... 191   xvi  7.5.3.2 Conservative patellofemoral OA treatments with or without alignment outcomes 192 7.5.3.3 Treatments targeting patellar alignment ......................................................... 196 7.5.3.4 Treatments for patellofemoral pain without alignment outcomes .................. 196 7.5.3.5 Who will benefit most from treating alignment? ............................................ 197 7.5.4 The bigger picture – public health .......................................................................... 198 7.6 Future directions ......................................................................................................... 200 References ...................................................................................................................................202 Appendices ..................................................................................................................................261 Appendix A Search strategies for systematic review protocol ............................................... 261 A.1 Medline search strategy .......................................................................................... 261 A.2 CINAHL search strategy......................................................................................... 263 A.3 EMBASE search strategy ....................................................................................... 265 A.4 PEDro search strategy ............................................................................................. 267 A.5 CENTRAL search strategy ..................................................................................... 267 A.6 SPORTDiscus search strategy ................................................................................ 269 A.7 Web of Knowledge search strategy ........................................................................ 272 A.8 Google Scholar search strategy (Google Chrome) ................................................. 274 A.9 Conference proceedings search strategy ................................................................. 274 Appendix B Quality assessment tool for systematic review ................................................... 276 Appendix C Alignment and morphology measures included in systematic review studies ... 278 C.1 Axial plane alignment ............................................................................................. 278 C.2   Sagittal plane alignment ........................................................................................ 279   xvii  C.3 Frontal plane alignment .......................................................................................... 280 C.4   Trochlear morphology .......................................................................................... 280 Appendix D Patient reported outcome measures used in MOOV study ................................ 282 D.1   International Physical Activity Questionnaire – Short (IPAQ-S) ......................... 282 D.2 EQ-5D-5L ............................................................................................................... 285 D.3 Knee injury and Osteoarthritis Outcome Score (KOOS) ........................................ 287 D.4 KOOS Patellofemoral Subscale (KOOS-PF).......................................................... 292 D.5 Anterior Knee Pain Scale (AKPS) .......................................................................... 294 D.6   Tampa Scale of Kinesiophobia ............................................................................. 295 D.7 Knee Self-Efficacy Scale (K-SES) ......................................................................... 296 Appendix E MOOV study: tibiofemoral 3D alignment .......................................................... 298 E.1 Table: tibiofemoral alignment in full extension, standing compared to supine ...... 298 E.2 Tibial adduction results, mixed effects: model coefficients (β) with 95% confidence intervals (CI) and p-values.. ................................................................................................ 299 E.3 Tibial internal rotation results, mixed effects: model coefficients (β) with 95% confidence intervals (CI) and p-values.. ............................................................................. 300 E.4 Tibial proximal translation results, mixed effects: model coefficients (β) with 95% confidence intervals (CI) and p-values. .............................................................................. 301 E.5 Tibial lateral translation results, mixed effects: model coefficients (β) with 95% confidence intervals (CI) and p-values. .............................................................................. 302 E.6 Tibial anterior translation results, mixed effects: model coefficients (β) with 95% confidence intervals (CI) and p-values. .............................................................................. 303 Appendix F MOOV study: one-legged stance in those with and without malalignment ....... 304   xviii  F.1 Within group difference (one-legged alignment minus two-legged alignment, 30° knee flexion), mixed effects model ..................................................................................... 304 F.2 Between group difference in one-legged stance, with vs. without malalignment, mixed effects model ............................................................................................................ 304    xix  List of Tables Table 1.1 Estimates of compartmental prevalence of radiographic knee OA .............................. 10 Table 1.2 Kellgren and Lawrence grading system for knee OA ................................................... 23 Table 1.3 Osteoarthritis Research Society International (OARSI) grading system for knee OA. 23 Table 1.4 American College of Rheumatology classification criteria for knee OA. .................... 25 Table 1.5. Proposed OARSI definitions for MRI-defined tibiofemoral and patellofemoral OA. 33 Table 1.6 Risk factors for OA at the tibiofemoral joint. ............................................................... 35 Table 1.7 Risk factors for OA at the patellofemoral joint ............................................................ 36 Table 2.1 Best evidence synthesis guidelines. .............................................................................. 54 Table 2.2 Study characteristics ..................................................................................................... 56 Table 2.3 Quality assessment scores, in order of publication year. .............................................. 59 Table 2.4 Axial plane alignment ................................................................................................... 60 Table 2.5 Sagittal plane alignment................................................................................................ 64 Table 2.6 Frontal plane tibiofemoral alignment. .......................................................................... 68 Table 2.7 Trochlear morphology. ................................................................................................. 70 Table 3.1 Reliability for alignment and morphology measures .................................................... 89 Table 3.2 Demographics and summary of alignment and morphology measures ........................ 90 Table 3.3  Logistic regression models for prevalent patellofemoral OA. ..................................... 91 Table 4.1 MRI sequence parameters. .......................................................................................... 100 Table 4.2 Participant characteristics. .......................................................................................... 107 Table 4.3 Inter-rater reliability .................................................................................................... 108 Table 4.4 Intra-rater reliability .................................................................................................... 108   xx  Table 4.5 Reference values among those with no patellofemoral joint full-thickness cartilage damage or knee pain ................................................................................................................... 109 Table 4.6 Reference values for full thickness patellofemoral cartilage damage regardless of symptoms .................................................................................................................................... 110 Table 4.7 Thresholds above or below which predicted odds ratios are significantly higher than unity for having patellofemoral OA features and/or pain. .......................................................... 114 Table 4.8 Thresholds where predicted odds ratios (OR) achieve clinical relevance for having full thickness cartilage damage. ........................................................................................................ 115 Table 5.1 MRI sequence parameters. .......................................................................................... 132 Table 5.2 Participant characteristics ........................................................................................... 144 Table 5.3 Repeatability reported separately by group, and combined ........................................ 145 Table 5.4 Mean values of patellofemoral alignment at full extension, supine and standing ...... 145 Table 5.5 Patellar flexion results, mixed effects ......................................................................... 148 Table 5.6 Patellar medial spin results, mixed effects.................................................................. 149 Table 5.7 Patellar medial tilt results, mixed effects .................................................................... 149 Table 5.8 Patellar proximal translation results, mixed effects .................................................... 150 Table 5.9 Patellar lateral translation results, mixed effects ........................................................ 150 Table 5.10 Patellar anterior translation results, mixed effects .................................................... 151 Table 5.11 Within-group difference in patellar alignment by position ...................................... 152 Table 6.1 Between-group difference in alignment of matched pairs .......................................... 165 Table 6.2 Patellar flexion, mixed effects .................................................................................... 166 Table 6.3 Patellar medial spin, mixed effects ............................................................................. 166 Table 6.4 Patellar medial tilt, mixed effects ............................................................................... 166   xxi  Table 6.5 Patellar proximal translation, mixed effects ............................................................... 168 Table 6.6 Patellar lateral translation, mixed effects .................................................................... 168 Table 6.7 Patellar anterior translation, mixed effects ................................................................. 168 Table 6.8 Two-dimensional alignment results, by group (in supine). ........................................ 171 Table 6.9 Within-group difference in alignment (one-legged minus two-legged), standing at 30° flexion ......................................................................................................................................... 171 Table 6.10 Between group differences for patellofemoral OA vs. controls, one legged stance at 30° knee flexion .......................................................................................................................... 172 Table 7.1 Overview of studies investigating interventions for patellofemoral OA .................... 192     xxii  List of Figures Figure 1.1 Forest plot: meta-analysis of knee OA prevalence ........................................................ 4 Figure 1.2 MR image of an asymptomatic knee ............................................................................. 8 Figure 1.3 Image of patella ........................................................................................................... 13 Figure 1.4 Patella rotations and translations ................................................................................. 13 Figure 1.5. Kinematics (patellar tracking) of the patellofemoral joint ......................................... 15 Figure 1.6 Aggrecans, by electron micrograph ............................................................................. 18 Figure 1.7 Major components of diagnosis of knee OA per EULAR recommendations. ............ 25 Figure 1.8 The natural history of OA............................................................................................ 30 Figure 1.9 A conceptual model of OA .......................................................................................... 38 Figure 1.10 Patellofemoral joint contact pressure profiles ........................................................... 44 Figure 2.1 Flow chart for study inclusion. .................................................................................... 55 Figure 3.1 Sagittal plane alignment .............................................................................................. 85 Figure 3.2 Axial plane alignment.................................................................................................. 85 Figure 3.3. Trochlear morphology ................................................................................................ 86 Figure 3.4 Flow chart of participant recruitment. ......................................................................... 88 Figure 4.1 Alignment and morphology measures ....................................................................... 101 Figure 4.2 Using the WORMS scoring method, the knee is subdivided into regions. ............... 103 Figure 4.3 Using the WORMS scoring method, cartilage signal and morphology is divided into eight categories. .......................................................................................................................... 103 Figure 4.4 Dose-response patterns for alignment. ...................................................................... 111 Figure 4.5 Dose-response patterns for morphology measures. ................................................... 112 Figure 4.6 Symptomatic patellofemoral OA dose-response patterns ......................................... 116   xxiii  Figure 5.1 Illustration of side view of an open-bore MRI scanner and backrest ........................ 123 Figure 5.2 Flow chart of participant screening procedures ......................................................... 126 Figure 5.3 a. Dual SENSE Flex-M coil set-up for high-resolution scanner ............................... 131 Figure 5.4 Upright scanner positioning in supine. ...................................................................... 133 Figure 5.5 Donning knee coil in preparation for standing scans. ............................................... 134 Figure 5.6 Upright scanner positioning in standing. ................................................................... 135 Figure 5.7 Image processing methods for one participant. ......................................................... 136 Figure 5.8 Patella alignment relative to femur............................................................................ 137 Figure 5.9 Tibia alignment relative to femur .............................................................................. 137 Figure 5.10 Flow chart for participant screening. ....................................................................... 143 Figure 5.11 Patellar flexion, standing (solid) and supine (dash) ................................................ 148 Figure 5.12 Patellar medial spin, standing (solid) and supine (dash) ......................................... 149 Figure 5.13 Patellar medial tilt, standing (solid) and supine (dash) ........................................... 149 Figure 5.14 Patellar proximal translation, standing (solid) and supine (dash) ........................... 150 Figure 5.15 Patellar lateral translation, standing (solid) and supine (dash) ................................ 150 Figure 5.16 Patellar anterior translation, standing (solid) and supine (dash) ............................. 151 Figure 6.1 Between-group alignment in standing and supine, three rotations............................ 167 Figure 6.2 Between-group alignment in standing and supine, three translations ....................... 169 Figure 7.1 Clinical method for assessing patella position .......................................................... 189 Figure 7.2 Single-leg squat test, four types of performance ....................................................... 190 Figure 7.3 Taping to unload the infrapatellar fat pad ................................................................. 194 Figure 7.4 BioSkin Q Brace ........................................................................................................ 195 Figure 7.5 Obesity is associated with multiple comorbidities .................................................... 199   xxiv  List of Abbreviations β  beta (coefficient size) κ   kappa 2D  two-dimensional 3D   three-dimensional  A  anterior ACL  anterior cruciate ligament ACL  anterior condylar line ACLR  anterior cruciate ligament reconstruction ADLs  activities of daily living AKPS  anterior knee pain scale AL  anterior condylar line ANOVA  analysis of variance BL  baseline BME  bone marrow edema BMI  body mass index BML  bone marrow lesion BO  bisect offset BPTB  bone-patellar-tendon-bone C  central CI  confidence interval CON  control(s) dGEMRIC  delayed gadolinium enhanced magnetic resonance imaging of cartilage   xxv  HT  hamstring tendon EQ 5D 5L EuroQOL (quality of life) scale, five dimensional 5 Likert EULAR The European League Against Rheumatism d  delta/change (Cohen’s d) D  Delta/change (Cook’s D) dfbeta  change in beta FLASH fast low angle shot MRI sequence GE  gradient echo HKA  hip-knee-ankle angle HRQOL health-related quality of life ICC   intra-class correlation coefficients ICRS  International Cartilage Repair Society IPAQ-S International Physical Activity Questionnaire-Short iPFOA  isolated patellofemoral osteoarthritis ISR  Insall-Salvati Ratio JSN  joint space narrowing KL  Kellgren & Lawrence KOOS  Knee injury and Osteoarthritis Outcome Score KOOS-PF Knee injury and Osteoarthritis Outcome Score - patellofemoral subscale K-SES  knee self-efficacy scale L  lateral LARS  ligament augmentation and reconstruction system  LCI  lower bounds of a confidence interval   xxvi  LD  lateral displacement LF  lateral facet LTI  lateral trochlear inclination lPFOA  lateral compartment patellofemoral osteoarthritis LPFTA lateral patellofemoral tilt angle (Laurin method) LPTA  lateral patellar tilt angle M  medial MF  medial facet MOAKS MRI osteoarthritis knee score mPFOA medial compartment patellofemoral osteoarthritis  mISR  modified Insall-Salvati Ratio MR  magnetic resonance MRI  magnetic resonance imaging MTI  medial trochlear inclination NA  no association NS  non-significant OA  osteoarthritis OARSI Osteoarthritis Research Society International OR   odds ratio OST  osteophyte(s) P  posterior PA  postero-anterior PANT  patellar anterior translation   xxvii  PC  patellar cartilage PCL  posterior condylar line PDw  proton density weighted MRI sequence PF  patellar facet PFJ  patellofemoral joint PFLEX patellar flexion PFOA  patellofemoral osteoarthritis PL  patellar length PLAT  patellar lateral translation PPROX patellar proximal translation PSPIN  patellar spin PTA  patellar tilt angle PTI  patellotrochlear index PTILT  patellar tilt RCT  randomized controlled trial S  tibial spine SA  sulcus angle SBA  subchondral bone attrition SD  standard deviation SDC95  smallest detectable change at 95% confidence SE  spin echo MRI sequence SEM  standard error or measurement SF-36  Medical Outcomes Score Short Form-36   xxviii  St  Standing Su  Supine T  tesla TADD  tibial adduction TANT  tibial anterior translation TFLEX tibial flexion TIR  tibial internal rotation TLAT  tibial lateral translation TPROX tibial proximal translation T1w  T1 weighted MRI sequence TA  trochlear angle TC  trochlear cartilage TD  trochlear depth TFA  tibiofemoral angle TFJOA tibiofemoral joint osteoarthritis TFR  trochlear facet ratio TL  tendon length TSE  turbo spin echo MRI sequence UBC   University of British Columbia UCI  upper bounds of a confidence interval U.K.  United Kingdom U.S.  United States USD  United States dollars   xxix  VAS  visual analogue scale VCHRI Vancouver Coastal Health Research Institute VISTA  Volume ISotropic Turbo spin echo Acquisition WOMAC  Western Ontario McMaster Universities Osteoarthritis Scale WORMS Whole organ magnetic resonance imaging score    xxx  Acknowledgements I start by deeply acknowledging that most of my work, and indeed most of my life, has taken place on the traditional, ancestral, unceded territories of the Musqueam, Tsleil-Waututh, and Squamish First Nations. Thank you to the many indigenous leaders who have opened their hearts and offered me guidance on how to live in a good way. Huy chexw.  I would like to thank Professor Karim Khan for seeing possibility in me from our very first skype while I was traveling in 2010. Thank you for your honesty, insight, humor, and passion. Thank you, Professor Kay Crossley, for inspiring me to want to dedicate such a large part of my life to the elegant knee cap. I am indebted for your commitment from across the globe, encouraging me to grow in ways I could not have imagined. Thank you, Dr. Bruce Forster, for offering an open door and a new perspective. Thank you, Dr. Charles Ratzlaff, for your critical thinking and strategic mind. Thank you to all four of you, for forming my committee, anchoring me in a solid and supportive learning environment. Thank you for your time, hard hitting feedback, and encouragement.  In addition to my committee, there are countless others without whom I could not have succeeded. They are coauthors, mentors, professional acquaintances and friends. Thank you, Josh Stefanik, for the frequent and ongoing patella-related geek out sessions that have led to a long wish list of projects, an international collaboration, and two publications (so far) that formed key projects for this dissertation. Thank you, Adam Culvenor, for entrusting me with your KOALA participants, for leading by example, and for your ongoing advice and feedback. Thank you, Harvi Hart and Nat Collins for sharing ideas, projects and sweat. Thank you, Agnes   xxxi  d’Entremont, Ewa Roos, David Felson, David Wilson, Charlie Goldsmith, Andrew Yung, Amy Phillips, Honglin Zhang, Trudy Harris, Jen Patterson, Christina Thiele, Michael Hunt, Mike Brenneman, Karen Zalamea, Paul Drexler, and members of the RESCU team in Boston. I know there are many others who have supported me and deserve mentioning here – your names will most certainly come to me right after I have sent this off for publication. You are all generous and kind. Your contributions are immense and will always be remembered.  Financial support made this dissertation possible. Thank you to the Canadian Institutes of Health Research and the Vanier Canada Graduate Scholarship that made full time doctoral study possible. Thank you to the Osteoarthritis Research Society International for the OARSI travel award that enabled me to spend time with Josh Stefanik, David Felson and the RESCU team in Boston. Thank you to the UBC and VGH Hospital Foundation for funding my research. Thank you to the Centre for Hip Health and Mobility for in kind contributions. Thank you to the Australian International Endeavour Award Research Fellowship that allowed me to work with the Kay Crossley’s team, prior to starting my PhD, and ultimately inspired me to pursue this research field.   Jaia, thank you for supporting my ongoing love affair with learning. Thank you for the music, the laughter, the antics. Thank you for the opportunity to aim for perpetual grace, all the while tripping over my own feet.   Colleen and Mike Macri, mom and dad, thank you. I dreamed of one day giving our family a degree, instead I gave them four. What a privilege it is to be your daughter.          xxxii  Dedication    I dedicate this dissertation to the heroes, activists, and lovers who live courageously every day, so that we may all dream big, know peace, express our deepest truths,  and love without limits.    xxxiii         “Be kind to your knees. You'll miss them when they're gone.” ~ Mary Schmich, erroneously attributed to Kurt Vonnegut's commencement address at MIT, 1997   1  Chapter 1 Introduction Osteoarthritis (OA) is a disabling musculoskeletal condition of total joint failure that results from loss of homeostasis. It affects over 30 million adults in the U.S. [1-3] and over 4.4 million in Canada [4]. Worldwide, OA is among the top three most prevalent musculoskeletal disorders, and ranks 11th in years lost to disability amidst 291 diseases or injuries [5]. On account of an aging population and rising obesity rates, prevalence is projected to increase substantially [6 7], potentially making OA the fourth leading cause of disability by the year 2020 [8]. Total annual direct and indirect costs associated with managing OA are estimated at $140 to $160 billion in the U.S. [9-11] and $27.5 billion in Canada [4]. The joint most commonly affected by OA – and associated with the most OA-related disability – is the knee.   In this chapter, I provide an overview of the importance of patellofemoral OA within the context of knee (primarily tibiofemoral) OA, which has received substantially more clinical and research focus. I discuss the burden of patellofemoral and knee OA, describe the OA process, discuss diagnosis, and provide an overview of what is known regarding risk factors for patellofemoral and knee OA. It is ultimately the fact that so little is known about risk factors for patellofemoral OA that justifies the main aim of my thesis, which was to quantify the cross-sectional associations between measures of patellofemoral alignment, tibiofemoral alignment, or trochlear bony morphology with patellofemoral OA.    2  1.1 The burden of knee OA According to self-reported doctor’s diagnosis, over 15 million adults aged ≥25 years live with knee OA in the U.S. [12]. When considering both prevalence and disability, knee OA accounts for 83% of the worldwide burden of OA [5].   1.1.1 Prevalence In a recent meta-analysis (24 radiographic studies of structural features alone, 13 symptomatic knee OA studies which include pain, and 7 self-report studies), the overall prevalence of knee OA was estimated to be 23.9% [13] (see Figure 1.1). However, a key challenge in quantifying prevalence lies in the many ways to diagnose knee OA, such as self-report, structural features on imaging, reported symptoms, and/or other clinical features (for details regarding diagnosis, see Section 1.5) [14]. Estimates also vary due to sample characteristics such as sex, age, body mass index (BMI), socioeconomic factors, occupation, or ethnicity.   Overall, there is a secular trend towards increasing knee OA prevalence. This is in part due to a worldwide aging population and increasing rates of obesity, but may also be related to greater patient and physician awareness of knee OA (i.e. increasing likelihood of diagnosis) [15], and increasing reports of knee pain not explained by age or body mass index (BMI) [16].   To illustrate how much these features can affect outcomes, consider that the lifetime risk of developing symptomatic knee OA ranges from 13.8% of adults aged ≥25 years based on self-reported doctor’s diagnosis (19.7% if obese) [15], up to 44.7% of adults aged ≥45 years based on   3  symptomatic radiographic OA (60.5% if obese) [17]. Sample demographics and diagnostic methods must be carefully considered when interpreting the results of knee OA studies.  1.1.2 Physical impact Knee OA is responsible for more physical disability than any other disease among older community-dwelling adults in the U.S. [18 19]. Worldwide, this condition was responsible for 14.2 million years lived with disability in 2010 [5]. Knee OA is associated with pain, muscle weakness or loss of motor control, reduced proprioception or balance, and limited range of motion [20-22]. Because the knee is a weightbearing joint, knee pain results in substantial loss of function and mobility, and reduction in independence and participation in personal care, occupations, and social engagement. For example, men with symptomatic knee OA report 3.2 times higher odds of having difficulty walking compared to those without symptomatic knee OA, and 2.5 times higher odds of having difficulty with transfers (for example, getting in and out of a car or bath). The odds for women with these tasks are 5.3 and 4.4 times higher, respectively [23]. Consequently, people with knee OA report worse health-related quality of life (HRQOL). Using the generic Short Form 36 (SF-36) questionnaire [24], those with knee OA scored worse than age- and sex-matched healthy controls on six out of the eight SF-36 subscales which represent both physical and mental aspects of HRQOL [25]. Those with symptomatic knee OA report worse HRQOL than those with only radiographic OA or only knee pain [26].       4   Figure 1.1 Forest plot: meta-analysis of knee OA prevalence, heterogeneity and 95% CIs by OA definition (self-report, radiographic, symptomatic), in women. Reprinted with permission1 .                                                  1 Reprinted from Osteoarthritis and Cartilage, Volume 19, Pereira D. et al., “The effect of osteoarthritis definition on prevalence and incidence estimates: a systematic review”, pp 1270-1285, 2011, with permission from Elsevier.   5   While knee OA is typically considered a condition affecting older adults, the burden of knee OA is not limited to the aged. Recent epidemiology studies report knee OA (defined as self-reported knee pain or stiffness along with a doctor diagnosis of arthritis that is not rheumatoid arthritis, fibromyalgia, lupus, or gout) in age groups as young as 25 years [15]. In addition, OA can develop in young people as early as one year following surgical reconstruction of traumatic knee injuries [27]. The result is a loss of function and quality of life much earlier in life when disability can substantially impact on work and family obligations [28]. Indeed, in individuals aged 30 – 55 years with knee OA, mean HRQOL is 32% worse than population estimates [29].   1.1.3 Economic impact The financial implications of a condition with both high prevalence and considerable physical disability are substantial. The costs of knee OA can be described in terms of direct costs (those associated with health care) and indirect costs (those associated with lost occupational productivity of the individual or their care givers). While these costs are attributable to individuals living with knee OA, the burden is borne at both a personal and societal level.  There are currently no known disease-modifying treatment approaches for knee OA. Thus, management of the condition is generally considered palliative until pain and disability becomes severe enough to warrant total joint arthroplasty (knee replacement). This surgical procedure is costly and requires admission to hospital, resulting in OA being the second most costly hospital expense in the U.S. next to septicemia [30]. In 2011, OA was the 5th highest reason overall for a hospital stay in the U.S. [31], yet it was the most common reason for hospitalization in people aged 65-84 years (142 / 10,000 population, up 54% over previous 15 years), and even in younger   6  people aged 45-64 years (49 / 10,000 population, up 160% in 15 years) [31]. The estimated cost (converted to 2017 USD2) of a total knee arthroplasty plus associated rehabilitation is over $21,000 USD [32 33]. Average annual direct costs per patient in developed countries are estimated at $10,700 USD per year [34]. Extending this to a lifetime, mean estimated direct costs are about $133,000 USD [32].   The indirect costs of knee OA are equally substantial. Knee OA is associated with nearly a two-fold increase in risk of needing sick leave in Sweden [35], while in Norway odds are twice as high for needing up to eight weeks sick leave, as well as for needing more than eight weeks leave [36]. The average number of actual missed work days (i.e. ‘absenteeism’) due to knee OA tends to be few (1-3 days per year) [37 38]. However, 80% of individuals with knee OA report they are not fully productive at work [38]. The majority of lost productivity is therefore due to reduced capacity while still at work (known as ‘presenteeism’) [38]. On average, in Western countries indirect costs average approximately $4,900 USD per year per person [34].  Taken together, total direct and indirect costs attributable to knee OA are estimated at $961 billion USD in Europe, and may be even higher in the U.S. on account of higher obesity and comorbidity rates [34]. This does not capture the intangible costs associated with living with pain, reduced functional capacity, and related symptoms like fatigue or depression that also impact quality of life. Thus, knee OA is associated with tremendous personal and societal burden.                                                  2 All monetary amounts reported in this section are converted to 2017 USD from either the year of publication or the year reported in the publication using online calculators (http://www.in2013dollars.com/) and converted from other denominations into USD where needed, https://www.google.ca/#q=convert+euros+to+dollars )    7   1.2 The contribution of the patellofemoral joint to knee OA is under-recognized Early anatomists described the knee as a hinge joint, capable of movement in a single sagittal plane [39]. Consistent with this perspective, research studies and clinical medicine traditionally focused on the tibiofemoral joint (Figure 1.2), the part of the knee responsible for sagittal plane movement (i.e. flexion/extension). The tibiofemoral joint consists of two condyloid articulations, corresponding to the lateral and medial condyles of the femur articulating with the lateral and medial condyles of the tibia, respectively. However, a third articulation also exists, that of the patella on the trochlear groove of the distal femur, called the patellofemoral joint. The knee is thus a compound joint consisting of three distinct yet inter-related articulations [40]. Understanding of knee OA has been compromised by a lack of patellofemoral joint research [41 42].      8            Figure 1.2 MR image of an asymptomatic knee, 55 year old man: a. patellofemoral joint; b. tibiofemoral joint; c. patellofemoral joint, axial view; d. lateral tibiofemoral compartment; e. medial tibiofemoral compartment.    1.2.1 Compartmental patterns of knee OA Standard frontal plane radiographs (e.g. postero-anterior [PA] view) are used to view the tibiofemoral joint. However, alternate views such as lateral or skyline are require to view the patellofemoral joint [43]. As early as 1992, researchers advised that epidemiological and clinical knee OA studies should routinely consider the patellofemoral joint, for example by obtaining multiple radiographic views of the knee [44]. This unfortunately has not consistently occurred. As a result, the prevalence and overall burden of knee OA (section 1.1) has likely been underestimated.   Compartmental prevalence of radiographic knee OA has been determined across several studies (see Table 1.1). Overall, OA is commonly isolated to the patellofemoral joint in individuals aged over 50 years with knee pain (community-dwellers or hospitalized), ranging from 11% to 35%. Combined patellofemoral and tibiofemoral OA is also very common, ranging from 19% to 88%. Longitudinally, in a sample aged over 50 years with knee pain and either no or mild knee OA   9  (n=414), three-years of follow-up revealed an annual incidence of patellofemoral OA of 9.6% compared to 7.2% at the tibiofemoral joint [45]. The most common sequence of development of radiographic knee OA was onset at the patellofemoral joint with subsequent progression to combined, multi-compartment knee OA [45].    A systematic review of 32 studies of radiographic patellofemoral OA estimated population-based prevalence at 25% of individuals aged 20 years or older, and 39% of symptomatic cohorts aged 30 years or older [46]. Estimates are even higher when MRI-defined structural features are included, with another systematic review reporting prevalence in patellofemoral OA features in about half of the general population [47]. Given the high prevalence of isolated patellofemoral OA, true prevalence of knee OA is likely substantially higher than current estimates that have not fully considered the patellofemoral joint [13]. Knee OA likely involves the patellofemoral joint (either isolated or combined) at least as often as the tibiofemoral joint [48 49].     10  Table 1.1 Estimates of compartmental prevalence of radiographic knee OA    Patellofemoral Medial tibiofemoral Lateral tibiofemoral Study Sample Views Isolated  Total  Isolated  Total  Isolated  Total  Ahlbäck 1968 [43]  370 knees in 281 hospital patients (80% women) with radiographic joint space narrowing PA Lateral Skyline  35% 48% 45% 57% 6% 10% McAlindon et al. 1992 [44] 273 community-dwelling adults aged ≥55 years with reported knee pain, 146 (54%) with radiographic OA (unilateral OA or symmetrical bilateral) AP Lateral MEN: 11%  WOMEN: 24%  19%   34%   21%   12%  31%   19%  2%   3%  5%   6% McAlindon et al. 1992 [44] 240 community-dwelling adults aged ≥55 years without knee pain, 40 (17%) with radiographic OA AP Lateral 4% 5% 7% 8% 1% 1% Ledingham et al. 1993 [50] 252 patients admitted to hospital for knee OA AP Lateral  24% 88% 10% 67% 1% 16% Duncan et al. 2006 [51] 777 adults reporting knee pain who were aged ≥50 years, 68% had radiographic knee OA PA Lateral Skyline 24% 64% 4% isolated TFJ / 44% total TFJ   1.2.2 Patellofemoral OA is under-diagnosed   The high prevalence of patellofemoral OA reported in the literature contrasts to current diagnostic patterns in clinical practice. A retrospective review of 57,555 general practice medical charts in the U.K. (patients aged 15 years or older) revealed 1,782 knee-related consultations. Of these, 303 were coded as disorders of the patellofemoral joint, and only 13 of these were diagnosed as patellofemoral OA [52]. It could be speculated that this is in part due to limited   11  awareness of patellofemoral OA by both patients and physicians, coupled with missed diagnoses on account of not ordering all three views when requesting radiographs.   1.2.3 Patellofemoral OA is clinically important  Patellofemoral OA is not only highly prevalent, but it is also clinically important. Patellofemoral OA is associated with at least as much pain and disability as tibiofemoral OA, and possibly more so. The odds of having pain are higher with patellofemoral osteophytes compared to tibiofemoral osteophytes [53 54]. In addition, patellar cartilage volume, but not tibial or femoral cartilage volume, was associated with self-reported pain, function and global scores of the Western Ontario McMaster Universities Osteoarthritis Scale (WOMAC) [55]. Pain, stiffness and reduced function are all associated with severity of patellofemoral OA [55 56]. Odds of having pain or reduced function may also be higher for combined OA compared to either compartment alone[57-60], yet pain with isolated patellofemoral OA is at least as severe as tibiofemoral OA, if not worse[59 61]. Finally, five to 10 years after anterior cruciate ligament reconstruction (ACLR), patellofemoral OA was more strongly associated with self-reported pain and function than tibiofemoral OA, and was also associated with several functional tests (while tibiofemoral OA was not associated with function) [62]. These findings highlight that patellofemoral OA is not only prevalent, but is also clinically important: patellofemoral OA may confer greater burden on individuals than tibiofemoral OA.  1.2.4 Patellofemoral OA is a research and public health priority It is well established that knee OA represents a substantial public health burden. The European League Against Rheumatism (EULAR) is a key international body guiding OA research and   12  management (http://www.eular.org/). EULAR has identified high priority research areas that include: improved understanding of the pathology of the earliest stages of OA; and identifying predictors of OA progression, particularly those that can assist with developing targeted interventions[63]. Since development of knee OA commonly starts in the patellofemoral joint [43 45 64], patellofemoral OA research addresses these global research priorities. Better understanding of patellofemoral OA and associated risk factors could lead to earlier diagnosis and to individualized treatments that could slow the progression of knee OA and improve quality of life.   1.3 Patellofemoral joint anatomy and function ‘Patella’ is derived from converging etymologies resulting from the British occupation of India (Latin patena, ‘small dish’ likely owing to the shape of the bone; and Hindi patella haddi, ‘farmer’s bone’ said to be related to the impact of this occupation on the kneecap) [65 66]. The patella is the largest sesamoid bone in the human body, estimated at approximately 45 mm in diameter and somewhat flat and triangular in shape [40 67 68]. The base of the patella is located paradoxically at the most proximal aspect of the patella, while the apex is located inferiorly (see Figure 1.3). The patella is positioned anterior and superior to the trochlear groove on the distal aspect of the femur (when the knee is in full extension), thus forming the patellofemoral joint. This joint is considered a modified plane joint [69], and the patella has considerable freedom of movement relative to the femur on account of the joint’s bony morphology and surrounding soft tissues. Movement of the patella is best described arthrokinematically. As the whole knee flexes and extends, the patella rotates (flexion/extension, frontal plane spin, medial/lateral tilt) and translates (medial/lateral, inferior/superior, anterior/posterior) (see Figure 1.4) to accommodate   13  the contours of the femoral trochlea as the relative positions of the patella, femur and tibia change. The patellofemoral joint shares its articular capsule with the medial and lateral tibiofemoral joint (thus forming a compound joint).       Figure 1.3 Image of patella, anterior view (left) and posterior view (right). a. base of patella; b. apex of patella; c. central median ridge; d. secondary ridge; e. lateral facet; f. medial facet; g. odd facet. Public domain3.             Figure 1.4 Patella rotations and translations a. patellar flexion/extension; b. medial/lateral spin; c. medial/lateral tilt; d. proximal/inferior translation; e. lateral/medial translation; f. anterior/posterior translation. Modified from illustrations by Vicky Earle. Reprinted with permission4.                                                   3 Palica, copyright holder. Modified from: https://commons.wikimedia.org/w/index.php?curid=65520; https://commons.wikimedia.org/w/index.php?curid=65513  4 Reprinted from Journal of Magnetic Resonance Imaging, Macri EM et al., “Patellofemoral and tibiofemoral alignment in a fully weight-bearing upright MR: Implementation and repeatability” Published Online First: doi:10.1002/jmri.25823, 2017, with permission from John Wiley and Sons   14  The patella is located within the tendon of the quadriceps femoris muscle, and one of its key functions is to increase the moment arm and thus torque of the quadriceps muscles about the knee joint [70]. In doing so, the patellofemoral joint is exposed to the weightbearing loads at the knee, particularly during flexed-knee weightbearing tasks including walking (half the body weight), climbing stairs (three to four times the body weight), and deep squatting or jumping (seven or more times the body weight) [71 72]. Possibly because of this weightbearing role of the patellofemoral joint, the articular cartilage on the posterior surface of the patella is the thickest articular cartilage in the human body, at up to five (5) mm [40 65 73]. The tibiofemoral joint bears as much or more weight than the patellofemoral joint (depending on the task)[73 74], however loading through the tibiofemoral joint is borne largely by the meniscus[75], perhaps explaining the thinner tibiofemoral cartilage (approximately 2.8 mm average thickness)[76].   The posterior surface of the patella is made up of several facets [40 67]. A central median ridge (also called the crista) runs vertically, dividing the patella broadly into a medial and lateral compartment (see Figure 1.3). A secondary ridge runs vertically and medial to the central median ridge, and divides the medial facet from the odd facet.   1.3.1 Patellofemoral alignment Alignment at the patellofemoral joint refers to the instantaneous position of the patella relative to the femur, at a given moment. In reality, bones move relative to each other over time, and in biomechanics this is described in terms of kinematics (i.e. movement without reference to the forces causing the movement). In lay terms, movement of the patella within the trochlear groove   15  is sometimes called patellar tracking. This section describes some of the features that can influence patellar tracking, which can be described in terms of alignment at a given moment.   Figure 1.5. Kinematics (patellar tracking) of the patellofemoral joint during knee flexion/extension.A., B, C. Movement of contact along the patella and femur during flexion extension. D, E. Path and contact areas of patella on femur at various angles of knee flexion. With permission5                                                    5 Reprinted from , Neumann DA, “Kinesiology of the musculoskeletal system”, Chapter 13., p 538, 2009, with permission from Elsevier   16  During movement, initial contact between the patella and femoral trochlea begins at the apex of the patella and occurs with the knee at approximately 20° of flexion [77]. As the knee moves into deeper flexion, the contact area moves proximally up the posterior patella, while contact of the medial and lateral patellar facets move inferiorly against the medial and lateral margins of the femoral trochlear groove. The contact surface area increases with deeper flexion angles up until approximately 60° - 90° flexion. Beyond 90°, the medial facet of the patella moves into the intercondylar notch of the femur, and thus contact is transferred onto the odd facet through to full flexion (about 135°). On account of the morphology of the knee, tracking of the patella in early flexion moves in a slight medial direction, then beyond 90° flexion it moves slightly laterally, thus tracking in a gentle “C” shape in the frontal plane throughout full knee excursion (flexion-extension) [77].  Soft tissues, including articular capsule, ligaments, retinacula, fatty tissue, and muscle, surround the knee complex to provide both passive and active restraint and stability to the patellofemoral joint. Key muscle groups that cross the knee joint include the quadriceps, hamstrings, and gastrocnemius muscles. Importantly, these muscles are all multiarticular (i.e. cross more than one joint), and can therefore influence positioning and movement at the knee but also at the ankle and hip joints. In addition, muscles that do not cross the knee can also influence movement at the knee. For example, weak or dysfunctional lumbopelvic muscles can influence positioning of the femur, such as adduction or internal rotation, which in turn can result in relative patellar lateral displacement or tilt [78 79]. In addition to muscle function, other factors can influence alignment either proximally (hip/pelvis/trunk) or distally (foot/ankle) that in turn affect patellofemoral alignment [78-81]. For example, someone who stands with pronated feet might present with an   17  internally rotated tibia [71 82]. The relatively large amount of movement possible at the patellofemoral joint, combined with the heavy influence of proximal, local, and distal factors, contributes to the dominant theory that biomechanical factors may increase the risk for onset and progression of knee OA [83].  1.4 What is OA? OA has been described in lay terms as a disease of ‘wear and tear’ [84]. Rather, it is more appropriate to describe the process as one of ‘wear and repair’ [85] because OA is a disease associated with loss of homeostasis [86 87].   Disruption of articular extracellular matrix is considered a ‘hallmark’ of OA [84] and its subsequent cartilage loss a ‘signature’ feature of OA [88]. The pathology underlying OA, however, is not limited to cartilage (or bone) as is commonly described. Rather, OA is a final common pathway representing total joint failure [89 90]. As such, it is accurate to conceptualize OA as a disease affecting synovial joints as ‘whole organs’ [84 91]. Thus knee OA involves cartilage and bone, but also meniscus, ligaments, joint capsule, synovium, muscles, nerves, blood vessels, bursae, fat pads and other periarticular structures [84 87 91].   1.4.1 Cartilage Cartilage is avascular, aneural, and made up mostly of water, collagen (mostly type II), aggrecans and other matrix molecules, and a small number of chondrocytes. Aggrecans are large aggregating proteoglycans (see Figure 1.6) that are highly negatively charged and thus function to draw in and maintain water content within the cartilage [92].     18         Figure 1.6 Aggrecans, by electron micrograph (A): proteoglycan aggregate. (B): model of a portion of the proteoglycan aggregate showing 6 aggrecan monomers. Micrograph by Joseph Buckwalter, University of Iowa; Model by Mark Sabo and Vincent Hascall, Cleveland Clinic Foundation. Reprinted with permission6.  A dominant hypothesis in OA pathogenesis is that chondrocytes and other cells such as synoviocytes respond to injury, inflammation, and/or altered metabolic processes, initiating cell-mediated repair and remodeling processes [84 91 93]. Subsequent matrix degradation occurs, including damage to aggrecans and leaking of aggrecan fragments into the synovial fluid. This leads to loss of ability to bind water appropriately within the cartilage, and edema and cartilage softening occurs [91 94]. Cartilage swelling or thickening occurs in younger patients two years after an injury to the anterior cruciate ligament [87 95]. Early cartilage changes are reversible if chondrocytes are able to repair the full extent of damage to the extracellular matrix. If not, however, apoptosis eventually takes place, and capacity to maintain metabolic homeostasis declines [96]. Since up to 95% of load support in healthy patellofemoral cartilage is provided by fluid pressurization, the altered fluid dynamics within the cartilage renders the extracellular matrix vulnerable to further mechanical insult [94 97]. Catabolic metabolism becomes the                                                  6 Reprinted from Journal of Biological Chemistry, Volume 276, Roseman S. “Reflections on Glycobiology”, pp 41527-41542, 2001, with permission from the American Society for Biochemistry and Molecular Biology.   19  dominant process, and irreversible histologic changes begin [84 96]. This includes superficial cartilage fraying, further loss of proteoglycan and collagen content, angiogenesis into the deep zone of cartilage, and subchondral bone remodeling [91 94 96]. Slowly, cartilage damage will progress to loss of cartilage volume [98].   1.4.2 Bone Bone remodeling in OA is theorized to be a response to changes in joint loading vectors [84 86]. On account of being highly vascular, bone adapts more quickly to load changes than cartilage. This adaptability is likely why bony changes are detected sooner than cartilage changes or joint space narrowing [84 86 96].   Osteophytes are bony spurs that form at joint margins through chondrogenesis followed by endochondral ossification. This is possibly a strategy to increase joint surface area and reduce overall joint load, or in an attempt to improve or restore stability of the joint [86 99 100]. In animal models, osteophyte growth begins in as little as two or three days following experimental procedure designed to initiate OA [86 99]. Central (or intra-articular) osteophytes, arising from the articular surface, are less common at the knee but may also develop and are thought to represent more severe OA [43 101-103].  Bone marrow lesions (BMLs) are often located near sites of bone microfracture or cartilage trauma. BMLs are defined on magnetic resonance imaging (MRI) as an ill-defined area of homogeneously increased signal intensity on fluid-sensitive sequences (or decreased signal intensity in T1-weighted sequences) [104 105]. Histologically, BMLs are made up of   20  lymphocytes, bone marrow necrosis (swelling or disintegration of fat cells), fibrosis (replacement of marrow by collagen fibres), or increased vascularization/bleeding [84 104-107]. BMLs likely represent an area of high bone turnover, signifying bony repair [105 108], and they may progress into cysts [107 109].  Several other common bony features are associated with OA. They are believed to represent pathological subchondral bone remodeling in areas responding to increased joint stress, and are often associated with areas of articular cartilage damage [110]. Subchondral attrition appears radiographically as hyperdensity, or on MRI as flattening or depression of the articular cortex [84 111 112]. Subchondral sclerosis appears radiographically as a thickened subchondral plate (hyperdensity) and on MRI appears as ill-defined low signal on fluid-sensitive images [43 113]. Both attrition and sclerosis are considered to represent more severe OA [96 114] and may be associated with changes in bone mineral density [100 113 115]. Subchondral cysts (or pseudocysts) are identified on imaging as well-defined lesions [87 110]. Cysts represent focal areas of bony necrosis, where phagocytosis of necrotic bone has left a fibrocartilaginous core, and increased nearby osteogenic activity contributes to a wall of sclerosis around the cyst core [43 110].  1.4.3 Other joint features Synovium plays an important role in knee OA. Synovitis (synovial thickening) and joint effusion (distension of the synovial cavity) are common findings in knee OA [84 96], and even predict incident radiographic knee OA [116]. Synovitis located in the parapatellar subregion in particular is associated with increased risk of knee pain [117]. Biochemical processes within the synovium   21  communicate with, and thus influence, processes occurring in cartilage and subchondral bone [86 96]. Synovial fibrosis may occur in advanced stages of OA [84].   The infrapatellar fat pad, or Hoffa’s fat pad, is intra-capsular and extra-synovial and plays a role, in conjunction with the synovium, in joint inflammation [118]. In patellofemoral OA, infrapatellar fat pad volume is 20% larger than in healthy controls [119]. However, fat pad volume in women with knee OA compared to healthy controls does not differ by OA status, but rather differs only by age [120]. In older women not selected on knee OA status, larger fat pad volume may be protective, associated with less cartilage volume loss over 2.6 years [121]. Thus, it is currently unclear how Hoffa’s fat pad changes in relation to knee OA. Size is theorized to be important either in a shock-absorbing capacity [121] or on account of its role in inflammation, with larger fat pads secreting greater quantities of adipokines [118 122].  Other soft tissues, such as meniscus and ligaments, may show matrix degeneration and collagen fiber disorganization in knee OA, similar to articular cartilage [84]. Patients with knee OA may also present with soft tissue damage, including ACL rupture, in the absence of a reported history of knee trauma, possibly secondary to the OA process [123].   Deposition of calcium pyrophosphate or basic calcium phosphate crystals may be seen with knee OA (both at the tibiofemoral and patellofemoral joints) [50 84 124], though it may present as an independent and possibly systemic condition that can resemble patellofemoral or knee OA (osteophytosis or bony attrition) [125]. These crystals may play a role in inflammation and   22  catabolic metabolism [124]. Studies suggest the presence of crystals in knee OA does not modify the phenotype of knee OA [125].  1.5 Diagnosing knee OA 1.5.1 Knee OA Clinically, knee OA is diagnosed radiographically in combination with clinical signs and symptoms. Radiographic tibiofemoral OA is typically evaluated with a PA view radiograph, while patellofemoral OA is evaluated with a skyline and/or lateral radiograph [126 127]. However, radiographic knee OA is often reported as tibiofemoral OA alone, which reflects commonly used radiographic grading systems that were developed for the tibiofemoral joint alone. Typical signs of radiographic OA include osteophytes, joint space narrowing (or joint space width), subchondral sclerosis, attrition and cysts [91 114 128]. In research, the most frequently used method for defining radiographic OA is the Kellgren and Lawrence (KL) scale (see Table 1.2) with a common cut-point for diagnosing OA of ≥ grade 2 [114 129 130]. In this scale, osteophytes are the main features used to diagnose knee OA [88 114 129]. This is problematic, given it remains unknown whether or not osteophytosis is exclusively pathologic [86]. Joint space narrowing serves as a proxy for cartilage loss since cartilage cannot be directly imaged radiographically [87 88]. Since cartilage loss is considered the ‘signature’ feature of OA, Felson et al. proposed a modified KL scale in which definite joint space narrowing is used to define grade 2 knee OA, with the addition of a new partial ‘grade 2/osteophyte’ that identifies a definite osteophyte without joint space narrowing [88]. Importantly, precise and reproducible joint space width measurement requires careful radio-anatomical positioning and sensitive measurement methods, particularly if measuring change over time [131].   23   Table 1.2 Kellgren and Lawrence grading system for knee OA [114 129]. Grade Kellgren & Lawrence 1957[114] Original description[129 132] 0 None No features of OA 1 Doubtful Doubtful JSN, possible OST lipping 2 Minimal Definite OST, possible JSN 3 Moderate Moderate multiple OST, definite JSN, some sclerosis and possible deformity of bone ends 4 Severe Large OST, marked JSN, severe sclerosis, definite deformity of bone ends OST = osteophyte; JSN = joint space narrowing  Another common diagnostic system is the Osteoarthritis Research Society International (OARSI) criteria (Table 1.3) [127]. This system grades each feature (osteophyte, joint space narrowing, attrition and sclerosis) separately, and it evaluates medial and lateral tibiofemoral compartments and plates separately [127]. Knee OA is defined as either joint space narrowing grade ≥2; osteophytes grade ≥2; or grade 1 joint space narrowing combined with a grade 1 osteophyte [133 134]. Using the OARSI system, diagnostic rates of tibiofemoral OA are higher than using the KL system, suggesting it may be a more sensitive diagnostic tool [133].  Table 1.3 Osteoarthritis Research Society International (OARSI) grading system for knee OA [127] Feature OARSI grade Marginal osteophytes 0 – 3+ Joint space narrowing 0 – 3+  Tibial attrition Absent/present Sclerosis Absent/present  Radiographic OA features are not consistently associated with clinical symptoms. Population-based epidemiology studies report tibiofemoral OA (KL ≥ grade 2) is up to three times more prevalent than symptomatic radiographic OA [13 135]. Since symptomatic OA will affect function and quality of life substantially more than asymptomatic OA, radiography as a stand-  24  alone diagnostic tool should be used cautiously. Clinical signs and symptoms should be considered along with radiological findings to diagnose clinical, or symptomatic, OA [2 128 136].  Clinical features for diagnosing knee OA include both symptoms and signs. The most recent diagnostic guidelines (2014) for clinical OA published by the National Institutes for Health and Care Excellence (NICE) are very simple: aged over 45 years, activity-related joint pain, and no morning joint-related symptoms or morning stiffness that is less than 30 minutes (https://pathways.nice.org.uk/pathways/osteoarthritis). Other diagnostic guidelines list symptoms as localized knee pain, brief morning stiffness (< 30 min), and decreased function; while signs include crepitus, bony enlargement, and restricted movement [128] (see Figure 1.7). The American College of Rheumatology defines clinical OA, with or without associated radiographic or laboratory tests, using a set of classification criteria (see Table 1.4) [136].      25   Figure 1.7 Major components of diagnosis of knee OA per EULAR recommendations. Reprinted with permission7  Table 1.4 American College of Rheumatology classification criteria for knee OA. Reprinted with permission8 Clinical and laboratory Clinical and radiographic Clinical§ Knee pain + at least 5 of 9: Knee pain + at least 1 of 3: Knee pain + at least 3 of 6: Age > 50 years Age > 50 years Age > 50 years Stiffness < 30 minutes Stiffness < 30 minutes Stiffness < 30 minutes Crepitus Crepitus Crepitus Bony tenderness + Bony tenderness Bony enlargement Osteophytes Bony enlargement No palpable warmth  No palpable warmth ESR < 40 mm/hour   RF < 1:40   SF OA      92% sensitive 91% sensitive 95% sensitive 75% specific 86% specific 69% specific ESR = erythrocyte sedimentation rate (Westergreen); RF = rheumatoid factor; SF OA = synovial fluid signs of OA (clear, viscous, or white blood cell count < 2,000/mm3). §Alternative for the clinical category would be 4 of 6, which is 84% sensitive and 89% specific.                                                  7 Reprinted from Annals of the Rheumatic Diseases, Volume 69, Zhang W et al., “EULAR evidence-based recommendations for the diagnosis of knee osteoarthritis”, pp 483-9, 2009, with permission from BMJ Publishing Group Ltd. 8 Reprinted from Arthritis and Rheumatism, Volume 29, Altman R et al., “Development of criteria for the classification and reporting of osteoarthritis: classification of osteoarthritis of the knee”, pp 1039-49, 1986, with permission from John Wiley and Sons   26  1.5.2 Patellofemoral OA An early OARSI OA atlas included the patellofemoral joint in 1995, using axial views of the patellofemoral joint to evaluate joint space narrowing, osteophytes, subluxations and subchondral sclerosis [126]. Unfortunately, when the atlas was revised in 2007 the patellofemoral joint was omitted [127]. Guidelines for identifying and quantifying OA have not been explicitly developed for the patellofemoral joint. Rather, methods developed for the tibiofemoral joint, such as KL or OARSI, have been adapted for use with the patellofemoral joint.   Several studies have attempted to determine which radiographic views are best for evaluating patellofemoral OA. Joint space narrowing is more reproducible using a skyline view radiograph compared to a lateral view [43 137-139]. The skyline view also enables localization of OA features to the medial or lateral compartment [137]. However, a lateral view may be better for viewing posterior osteophytes [138]. Using KL as a global score of patellofemoral OA (defining KL grade 1 as OA) reveals that neither view is superior in diagnosing radiographic patellofemoral OA [140]. Despite this, 4 – 7% of cases may be missed if only one view is acquired, and it is thus suggested that both lateral and skyline views be taken in addition to standard PA views in order to comprehensively evaluate the entire knee [138].   Not surprisingly, the classic clinical features reported for knee OA reflect clinical features of tibiofemoral OA, perhaps even late-stage OA, more so than they reflect patellofemoral OA [48 141 142]. It is not currently possible to diagnose patellofemoral OA on clinical features alone, and imaging thus remains necessary to confirm diagnosis [142].    27  While clinical features cannot be used to diagnose patellofemoral OA, characterizing its clinical presentation remains important. Clinical features of moderate to severe isolated patellofemoral OA (KL grade 3 or 4) may include swelling (or reported history of swelling), valgus malalignment, quadriceps weakness, and pain on compression of the patellofemoral joint [142]. Mild isolated patellofemoral OA (KL grade 2) is difficult to distinguish from a painful non-OA knee, but may include coarse crepitus and difficulty negotiating stairs [142]. Crepitus may distinguish patellofemoral OA from other causes of knee pain being treated by surgery (sensitivity 89%, specificity 83%) [140], and is associated with several patellofemoral MRI lesions in a population-based study with and without OA [143]. Pain descending stairs does not have adequate sensitivity and specificity (64% and 40% respectively) to be used diagnostically among people with or at high risk of developing knee OA, nor does anterior knee pain alone [144]. Other patellofemoral OA-related symptoms may include brief stiffness after prolonged sitting or lying (e.g. morning stiffness) [56], however this does not distinguish patellofemoral OA from multi-compartment knee OA or non-OA-related pain [142].  In the absence of clear diagnostic guidelines for clinical patellofemoral OA [145], a set of classification criteria are generally agreed upon in the literature to define patellofemoral OA in conjunction with radiographic findings. These include anterior knee pain (peri- or retro-patellar) that is made worse by tasks known to load the patellofemoral joint (e.g. ascending or descending stairs, getting in or out of a car or bath, or rising from sitting) [56 58 60 146-148].     28  1.5.3 The limitations of radiographic methods Radiographs are considered a criterion measure for diagnosing OA [96 128 149] and are therefore used extensively in both clinical and research settings, in part because they are affordable, fast, and easy to obtain [96]. However, there are several limitations to their use. For example, radiation exposure, while small for the knee, poses a risk to human tissue [150]. Another important consideration is that positioning of the knee joint within the X-ray beam influences the ability to reliably quantify osteophytes and joint space narrowing [96]. In addition, radiographs enable visualization of bone, but not other tissues involved in the OA process [96]. Finally, as reviewed above, definitions of OA vary substantially [3 88].  A definition of radiographic OA based on the presence of ‘definite’ osteophytes may be problematic. First, as discussed above (see section 1.5), it is not fully understood whether osteophytes consistently represent pathology [86]. Second, it is generally accepted that a ‘definite’ osteophyte marks the onset or prevalence of OA. However, osteophyte development lies on a continuum, and defining an osteophyte as ‘definite’ is arbitrary. For example, ‘doubtful’ osteophytes, noted as KL grade 1, may be identifiable as very small yet early osteophytes. In a prospective longitudinal study, over 60% of those with ‘doubtful’ osteophytes progressed to having ‘definite’ osteophytes at a 10 year follow up [151]. Thus, in the many studies that define ‘no OA’ as either KL grade 0 or 1, these studies may be erroneously including early OA in their ‘no OA’ comparison group, and in longitudinal studies may be misclassifying someone who progressed from grade 1 to grade 2 as someone with incident OA [45 152 153].     29  Defining radiographic OA based on joint space narrowing or width is also problematic. Also recommended as a key measure of both OA incidence and progression, this measure is a surrogate for cartilage thickness, which cannot be directly viewed radiographically [88 96]. Since early OA may be associated with cartilage swelling/thickening, joint space narrowing represents more advanced OA [87 91 95].   The use of radiographs as a primary diagnostic tool has resulted in OA being identified in a late stage [91 92 149 154] (see Figure 1.8). At this late stage, joint failure leaves few treatment options besides eventual total joint arthroplasty [15]. Unfortunately, arthroplasty is not a ‘cure’ for OA – up to 40% who undergo arthroplasty report no meaningful change or even worsening following surgery [155]. If OA disease modification is possible, it is necessary that OA be identified at a much earlier (and possibly reversible) stage than is afforded with radiography [92].      30   Figure 1.8 The natural history of OA, with current available assessment technologies placed along the disease continuum to demonstrate how early along the course of OA they can be useful. Reprinted with permission9  1.5.4 Magnetic resonance imaging (MRI) as a diagnostic tool In addition to radiography, imaging modalities used in OA research include computed tomography (CT), ultrasound, positon emission tomography, and MRI [149]. The latter is being used with increased frequency for both epidemiology and clinical research studies. Key benefits of using MRI includes its lack of ionizing radiation, and ability to view soft tissues [40 91 96 130]. In addition, parameters can be set to visualize different tissues or different aspects of tissues (such as cartilage morphology or cartilage biochemistry). MRI is not used routinely in                                                  9 Reprinted from Epidemiology of Aging, Newman AB and JA Cauley (eds.). Kwoh CK, “Epidemiology of osteoarthritis”, pp 523-36, 2012, with permission from Springer.     31  clinical settings for evaluating knee OA because clinically relevant diagnostic criteria have not yet been validated, and imaging is expensive, involves lengthy acquisition times, and requires time and skill for complex analysis of MRI features [96]. Despite this, its diverse applications have resulted in widespread use of MRI in OA research.  Broadly, MRI can be used for evaluating morphology (e.g. cartilage thickness, bone marrow lesions) or biochemical composition (e.g. collagen, extracellular matrix, water content) [156]. To evaluate cartilage morphology, semiquantitative features are scored on images with enhanced tissue contrast (since cartilage is characterized by low signal on T1- and intermediate on T2-weighted images). Generally, intermediate-weighted fast spin echo sequences with fat suppression are preferred [156] because they provide optimal contrast resolution. Biochemical (or quantitative) sequences, on the other hand, are obtained with sequences like T1ρ, T2 mapping or dGEMRIC (delayed gadolinium enhanced MRI of cartilage) [149 156]. The latter is a technique whereby gadolinium-diethylene triamine pentaacetic acid is injected as a contrast medium, and T1 mapping (most commonly) is performed once the medium has circulated adequately, to detect and quantify cartilage glycosaminoglycan content. These quantitative sequences may identify the earliest signs of OA, possibly detectable before morphological damage occurs [91]. Quantitative MRI is therefore being investigated as a potential imaging biomarker for disease onset.   1.5.4.1 Semiquantitative scoring Semiquantitative MRI scoring methods are designed to evaluate the knee as a whole organ, including all tissues possibly affected by OA [90 157 158]. Common systems include the whole   32  organ MRI score (WORMS) [90], the Boston Leeds osteoarthritis knee score (BLOKS) [158], and the MRI osteoarthritis knee score (MOAKS) [157]. With these methods, ordinal scores describe various morphometric states of different tissues within defined zones of the knee. These methods evaluate distinct features of OA rather than combining them into a global OA score.   Semiquantitative scoring has been used to evaluate knees with or without OA [64 159 160], and semiquantitative features such as BMLs and effusion/synovitis are associated with knee pain in individuals with knee OA [161]. However, semiquantitative MRI findings are also very common in asymptomatic knees [159 162] (in much higher frequencies than when OA features are evaluated with radiographs or ultrasound [47]). In a population-based sample of over 700 ambulatory adults aged ≥ 50 years, up to 97% of the subsample reporting knee pain had at least one semiquantitative MRI feature in the tibiofemoral joint, but so did 88% of the sample without pain [162]. In 70 young asymptomatic adults (aged 23 ±6 years), 21% had minor cartilage defects, 51% had patellofemoral BMLs, and 60% had patellofemoral osteophytes [159]. At first glance these studies call into question our understanding of OA. However, in fact it supports the concept of OA as a disease representing loss of homeostasis. The study findings suggest that bone and cartilage turnover appear to be a normal feature in knees across ages and health states. As discussed above (section 1.4), it is when catabolic processes begin to dominate that these features may become irreversible, representing OA onset. Semiquantitative MRI affords the opportunity to monitor these features both cross-sectionally and longitudinally, from healthy asymptomatic knees through to severe OA knees.     33  Although MRI is being employed increasingly to describe and quantify OA features, there is not yet consensus regarding a diagnostic threshold for knee OA using MRI. OARSI has proposed criteria for both tibiofemoral and patellofemoral OA using the MOAKS system (Table 1.5) [163], however these criteria have not yet been validated.  Table 1.5. Proposed OARSI definitions for MRI-defined tibiofemoral and patellofemoral OA. Reprinted with permission10. Accepted propositions for definition of OA on MRI after Delphi voting completion Preamble 1 MRI changes of OA may occur in the absence of radiographic findings of OA. 2 MRI may add to the diagnosis of OA and should be incorporated into the ACR diagnostic criteria including X-ray, clinical and laboratory parameters. 3 MRI may be used for inclusion in clinical studies according to criteria defined above but should not be a primary diagnostic tool in a clinical setting. 4 Certain MRI changes in isolation including cartilage loss, cartilage compositional change, cystic change and bone marrow lesions, ligamentous and tendinous damage, meniscal damage, and effusion and synovitis are not diagnostic of osteoarthritis. 5 No single MR finding is diagnostic of knee OA. 6 MRI findings indicative of knee OA may include abnormalities in all tissues of the joint: bone, cartilage meniscus, synovium, ligament and capsule. 7 Given the multiple tissue abnormalities detected by MRI in OA, diagnostic criteria are likely to involve several possible combinations of features. 8 Definite osteophyte formation is indicative of osteoarthritis.* 9 Joint space narrowing assessed by (non-weight bearing) MRI cannot be used as a diagnostic criterion. Definitions 10 A definition of tibiofemoral osteoarthritis on MRI would be: The presence of both group [A] features or one group [A] feature and two or more group [B] features Group [A] after exclusion of joint trauma within the last 6 months (by history) and exclusion of inflammatory arthritis (by radiographs, history and laboratory parameters): i) Definite osteophyte formation* ii) Full thickness cartilage loss Group [B]: i) Subchondral bone marrow lesion or cyst not associated with meniscal or ligamentous attachments ii) Meniscal subluxation, maceration or degenerative (horizontal) tear iii) Partial thickness cartilage loss (where full thickness loss is not present) iv) Bone attrition 11 Definition of PF OA requires all of the following involving the patella and/or anterior femur: i) A definite osteophyte ii) Partial or full thickness cartilage loss *The definition of a ‘definite osteophyte’ was not delineated in the Delphi process and requires further validation.                                                  10 Reprinted from Osteoarthritis and Cartilage, Volume 19, Hunter DJ et al., “Definition of osteoarthritis on MRI: results of a Delphi exercise”, pp 963-9, 2011, with permission from Elsevier Ltd.   34  MRI is a useful tool for evaluating bone, cartilage and other soft tissues, and is more sensitive than radiography in identifying very early OA [164 165]. However, in comparison to various reference standards (histology, arthroscopy, radiography), MRI specificity is generally higher than sensitivity [166], suggesting it may be more useful for ruling out than for ruling in an early OA diagnosis. Moreover, when compared to arthroscopic cartilage evaluation, MRI sensitivity and specificity are both lower when evaluating early cartilage lesions compared to more advanced lesions [167 168]. Also, use of arthroscopy as a reference standard is a form of incorporation bias; that is, MRI may indeed identify cartilage lesions with accuracy, but this may not equate to a clinical diagnosis of OA [169]. Given the high prevalence of cartilage lesions in healthy asymptomatic knees [159 162], it remains unclear to what extent MRI represents a valid imaging biomarker of early OA [170].  1.6 Risk factors  Evaluating the risk factors associated with, or contributing to, knee OA is clinically important. Risk factors can be identified that predict OA prevalence, incidence and/or worsening.  Important reasons for identifying risk factors are (i) to identify factors that may play a role in disease onset or progression; (ii) to identify target populations most susceptible to the disease, with a goal of prioritizing and tailoring diagnostic efforts; and (iii) to develop intervention strategies that target modifiable risk factors, with a goal of primary or secondary prevention.   Not surprisingly, risk factors at the tibiofemoral joint differ from those of the patellofemoral joint [89 171 172]. Summary tables are therefore presented separately for tibiofemoral OA (Table 1.6) and for patellofemoral OA (Table 1.7), though most studies did not evaluate whether OA was   35  isolated or combined. The tables report risk factors separately based on whether or not they are modifiable. This distinction is important because it identifies potential treatment targets.  Table 1.6 Risk factors for OA at the tibiofemoral joint.  Modifiable Non-modifiable Prevalence    BMI [89 171]  Knee injury (men) [89]  Age [12] Female sex [12 89 173] Ethnicity [12] Hand OA [171]   Incidence     BMI [174-177] Polyarticular OA [175] Knee injury [174 175 177] Occupation [177-179] Quadriceps weakness [96 180 181] Bone density [96] Vitamin D intake [96] Frontal plan alignment [96 182 183]  Age [175 176] Female sex [173-175] Genetics [96]  Worsening     BMI [184] Polyarticular OA [184 185] Vitamin D intake [184] Joint stiffness [184] Knee pain [185] Frontal plane alignment [182-186] Intense sport [96] Quadriceps weakness [181] Chronic joint effusion, synovitis, subchondral bone edema [96] Serum hyaluronic acid [185] Tumor necrosis factor α [185]  Prevalent patellofemoral OA [153] Age [184]    Risk factors are etiologically complex. For example, while obesity contributes to mechanical joint load, there is also a metabolic component as to how obesity confers risk of OA [187]. Fat tissue is responsible for moderating cell signalers such as adipokines, which play a role in inflammation [188]. This may explain why obesity increases risk of OA in non-weightbearing hands, despite risk of OA being higher in weightbearing joints [187]. In fact, knee joint load   36  during daily activities can be reduced up to four-fold per kilogram of weight lost in obese individuals with knee OA [189].   Table 1.7 Risk factors for OA at the patellofemoral joint*. Patellofemoral joint Modifiable Non-modifiable Prevalence    BMI [89 142 171 190]  Knee injury [89 191] Valgus alignment [142 190] Quadriceps strength [192] Forefoot varus deformity [193] Age [142 190] Male sex [142] Incidence    Occupation[177]  Worsening    BMI [171] Effusion [194 195] Quadriceps strength [196] Frontal plane alignment [98 197]   Female sex [171]  *Patellofemoral alignment and morphology have not been mentioned in this section, since they will be addressed comprehensively in Chapter 2  While load may increase OA risk, there remains uncertainty around specific loading activities [177 198]. This may be due to challenges with quantifying exposure, and accounting accurately for numerous confounders that influence cumulative joint load [177]. Activities that may affect joint load include occupation, household tasks, sport activity or history, physical activity, and combined lifetime activity [177-179 199]. Importantly, the relationship does not appear to be linear, and in fact moderate amounts of joint loading are necessary to maintain cartilage homeostasis and thus healthy joints [188]. Cumulative joint load or activity may also be a more relevant risk factor when in the presence of joint injury or obesity [179 198 200].   Risk factors for patellofemoral OA have not been well studied, thus Table 1.7 is not as comprehensive as Table 1.6. Associations are generally stronger in the lateral patellofemoral   37  compartment than medial [190 196 201], thus studies not investigating risk by compartment may miss important risk factors.   While Cicuttini reported premenopausal status as being cross-sectionally protective at the patellofemoral joint, estrogen replacement therapy and hysterectomy were not associated with patellofemoral OA, and age was not evaluated as a separate risk factor [171]. Peat et al. reported that female sex was protective for isolated patellofemoral OA in a cohort with knee pain, yet in population-based studies, patellofemoral OA is either more prevalent in women [89 144 202], or no different [49 201]. It is possible that patellofemoral OA is more prevalent in women in the medial patellofemoral compartment only [203], and since the medial compartment is not strongly associated with clinical outcomes [190 201], this could explain the higher prevalence in population-based studies compared to studies of symptomatic individuals.   Clinically, it is important to consider the overall risk profile of an individual, within the context of a biopsychosocial model, to identify those needing more careful assessment or to target specific modifiable risk factors (Figure 1.9). Intervention strategies might include weight management, strengthening or motor control exercise programs, or modifying the environment at work to reduce high load or high risk activities [177].     38              Figure 1.9 A conceptual model of OA, including modifiable and non-modifiable risk factors, plus aspects of the individual that contribute to OA outcomes. A biopsychosocial model also considers social and cultural contexts, such as family support, income, access to health care, and personal or cultural beliefs regarding health. Reprinted with permission11  Knee alignment and morphology have not been mentioned in this section, because they will be addressed comprehensively in Chapter 2 and through the remainder of this dissertation. However, it is worth acknowledging here that the ultimate aim of the work presented in this dissertation was to determine the role of knee alignment and morphology as potential risk factors for prevalent patellofemoral OA. The remainder of this chapter will continue to distinguish                                                  11 Reprinted from Nature Reviews Rheumatology, Volume 12, Roos EW et al, “Strategies for the prevention of knee osteoarthritis”, pp 92-101, 2016, with permission from Nature Publishing Group.     39  patellofemoral OA from tibiofemoral or knee OA, and provide a rationale for, and brief description of, the work presented in the following chapters.  1.7 Patellofemoral OA as a distinct phenotype 1.7.1 What is a phenotype? The term phenotype refers to the outward observable characteristics of a living organism that result from genetics in combination with environment influences. As it pertains to clinical phenomena, it can be thought of in terms of ‘typing’ conditions based on observable characteristics (Greek phainein, 'to show') [204]. Knee OA is a heterogeneous disease with a common end pathway of total joint disease [89 205 206]. Because of this, it may be fruitful to stratify knee OA into subsets based on pathoanatomic features. The intention is to identify, develop or improve tailored palliative or disease-modifying strategies for specific knee OA subsets [96 205 206].   There are numerous ways to divide OA into phenotypes. For example, it can be described in terms of: (i) mechanism of onset (e.g. joint trauma, obesity, genetic predisposition) [205]; (ii) dominant tissue involved (subchondral bone, cartilage, synovium); or (iii) dominant driver of pain or symptoms (mechanical, inflammatory, central or peripheral sensitization) [96 205 206]. It can also be described by dominant location of the disease, such as tibiofemoral, patellofemoral, multi-compartment, or polyarticular (e.g. knee and hand). For the purposes of this dissertation, I have adopted the latter definition, and in the following section will justify the clinical usefulness of defining patellofemoral OA as a distinct phenotype.    40  1.7.2 Patellofemoral OA is unique from tibiofemoral OA Whereas knee OA is characteristically different from hip or hand OA, it is clinically useful to further divide the knee into its constituent compartments.   As described above, patellofemoral OA is highly prevalent (see Section 1.2.1) and is associated with at least as much pain and disability as tibiofemoral OA, and possibly more so (see Section 1.2.1). Other features that may be unique to patellofemoral OA (or worse compared to tibiofemoral OA) include patellofemoral crepitus [143 207]; history of patellofemoral pain [143]; anterior knee pain that is increased with ascending and descending stairs but absent when walking on level ground [58 144]; gait impairments that differ from those seen with tibiofemoral joint involvement [148 208-212]; and impaired quadriceps and hip muscle function [58 201 208 209 211-213]. Patellofemoral OA also represents a distinct phenotype on account of its unique risk factor profile (see Section 1.6).   Patellofemoral OA precedes multi-compartment OA in a number of cases [45 64 153], and may therefore represent an earlier stage of disease and a possible target for early intervention [45]. Of the few clinical trials done to date, targeted patellofemoral OA interventions such as taping, bracing and physiotherapy, offer short-term efficacy in improving pain, function and disability [146 147 214 215]. Defining patellofemoral OA as a distinct phenotype may be clinically useful in that effective management of patellofemoral OA may differ from that of tibiofemoral OA [41].    41  1.7.3 Patellofemoral OA may be linked to patellofemoral pain Patellofemoral pain is a condition characterized by diffuse anterior knee pain that is made worse by activities that load the patellofemoral joint, such as squatting, stair climbing or prolonged sitting [48 216-218]. It commonly affects more women than men, and unlike patellofemoral OA, it presents in adolescents and young adults [219] as well as middle-age and older individuals [220 221]. The etiology is likely multifactorial, a combination of joint overload through overtraining in the presence of mechanical impairments and increased susceptibility [222]. It may be associated with crepitus, tenderness on patella facet palpation, catching, or mild swelling [48 218]. Other features may include patellar malalignment [223], dynamic knee valgus [224 225], hip and quadriceps muscle impairments [223 226-229], reduced lower extremity flexibility [223], or altered foot biomechanics [223 224].  Clinical findings are remarkably similar in both patellofemoral pain and patellofemoral OA (e.g. anterior knee pain made worse with tasks that increase load in the patellofemoral joint; crepitus, lower extremity muscle weakness – see section 1.5.2). Thus, it has been proposed that these two conditions lie on a disease continuum [82 230 231]. Of 63 girls presenting to a clinic at mean age 16 with patellofemoral pain, 77% had persistent pain up to 20 years later [232], highlighting the chronic nature of patellofemoral pain. Retrospectively, 22% of 118 patients with patellofemoral OA undergoing patella arthroplasty reported a history of anterior knee pain in their adolescence or early adulthood [233]. In a cross-sectional study of 224 middle-aged women (mean 54 ±10 years) with chronic patellofemoral pain, 69% had patellofemoral OA (98 combined, 57 isolated) [220]. In a similar study of 80 younger individuals (mean 36±7 years), almost half had early radiographic OA (KL Grade ≥1), most commonly isolated to the patellofemoral joint [234].    42   Though longitudinal studies are still needed for confirmation, evidence supports a link between patellofemoral pain and OA. With the evidence base for patellofemoral pain considerably well established, this literature may inform clinical and experimental approaches to evaluate and manage patellofemoral OA. It is plausible that somewhere along the continuum from patellofemoral pain to OA lies the critical point where homeostasis is lost and damage becomes irreversible. Investigating the two conditions as a single disease spectrum may help to uncover the nature and etiology of the phenotype of patellofemoral OA.  1.8 Patellofemoral joint mechanics and OA  As mentioned above (Section 1.4.1), the earliest stages of OA are characterized as a cellular response to injury or altered chemical environment (e.g. inflammation) in which homeostasis (i.e. repair, remodeling) is not maintained [84 91]. At the knee joint, mechanical injury can be described in terms of joint stress. Injury can occur either as a result of abnormal joint stress in a healthy joint, or normal joint stress in a susceptible joint (e.g. in the presence of underlying joint abnormalities) [82 87 235]. The injurious mechanical insult may be acutely traumatic or chronically repetitive and/or sustained [236 237]. Since cartilage is aneural, tissue level micro-injury may go unnoticed by the individual despite local cell-mediated responses. In the patellofemoral joint, two basic mechanical conditions lead to increased joint stress [48 238]. The first condition involves increased joint reaction force, which is influenced by body weight, and type or intensity of activity (e.g. weight bearing with knees flexed, as in squatting or jumping) [238 239]. The second condition is reduced joint contact area, which may be influenced by joint morphology (e.g. bony geometry that reduces joint congruency) [240] or patellofemoral   43  alignment (e.g. patella not centered within the trochlear groove) [93 238 239 241 242]. Alignment can be altered by structural features (e.g. tight lateral retinaculum or iliotibial band, or joint morphology) [239 243], or by neuromuscular function or impairment. For example, hip muscle impairment is associated with dynamic valgus, an aberrant movement pattern characterized by femoral adduction/internal rotation and tibial external rotation [209 244-247].   Joint stress may be elevated in individuals with patellofemoral pain. Compared to healthy controls, it is higher in level walking, and increases with faster speed, even in the presence of reduced joint reaction forces, highlighting the importance of contact area as a key contributor to joint stress in patellofemoral pain [238]. Joint stress is also higher when performing a two-legged squat to 15° or 45° in patellofemoral pain compared to controls (see Figure 1.10) [242]. Furthermore, existing patellofemoral joint structural damage may influence joint stress and render the joint more susceptible to further damage. For example, partial- or full-thickness cartilage lesions may cause uneven load distribution across the joint, and even mildly damaged cartilage is less able to withstand load [97 248]. Abnormal joint loads might therefore create an initial injury, or they might cause worsening of existing lesions [82].  Altered joint stress is likely the common underlying mechanism of many risk factors for both patellofemoral pain and patellofemoral OA (e.g. BMI, knee injury/trauma, muscle imbalances) [82 238]. However, joint stress cannot be directly measured in vivo. Instead, researchers and clinicians measure surrogate features that may contribute to joint stress. One such approach involves in vivo measurement of patellofemoral alignment and/or trochlear morphology, since these features are believed to be associated with joint congruency or stability, load distribution,   44  and/or joint contact area. Alignment and morphology measures are best evaluated using imaging modalities [93].         Figure 1.10 Patellofemoral joint contact pressure profiles of a control (left) and a patellofemoral pain (PFP, right) participant, standing with knees flexed to 15° and 45°. (L = lateral, M = medial). Reprinted with permission12  1.9 Using MRI to measure patellofemoral alignment and morphology MRI is one of several imaging modalities used to evaluate knee joint alignment and trochlear morphology [249]. Other commonly used methods include radiography [250], computed tomography (CT) [251], and fluoroscopy [252]. MRI is an appealing choice for several reasons. As previously stated, MRI does not emit ionizing radiation; and offers direct visualization of soft tissues (cartilage, ligament, muscle, meniscus) [40 91 96 130]. In addition, whereas skyline radiographs only show the anterior portion of the knee joint, MRI constructs axial plane images of the entire knee joint, including the posterior condyles. It is also possible to evaluate images by                                                  12 Reprinted from Osteoarthritis and Cartilage, Volume 19, Farrokhi S et al., “Individuals with patellofemoral pain exhibit greater patellofemoral joint stress: a finite element analysis study”, pp 287–94, 2011, with permission from Elsevier.   45  slice (i.e. tomographically), rather than viewing the entire knee as a two dimensional projection as in standard plane radiographs, with associated projection (aka magnification) error [93]. MRI thus provides more stable alignment estimates and better detection of trochlear dysmorphia than radiographic methods [253].  Many factors influence the quality of MR images and the extent to which images accurately reflect patellofemoral alignment or morphology [167]. A higher quality image will have adequate spatial resolution, signal to noise ratio, and tissue contrast, with efforts taken to optimize the scan plane and minimize artifacts [167]. Image quality is influenced by hardware considerations (type and strength of the scanner, type of knee coil) as well as sequence parameters (e.g. sequence protocol, fat suppression techniques) [156]. Standard clinical images are taken in non-weightbearing and supine with the knee relaxed near full extension. Acquiring images under other conditions may require specialized rigs, coils, and/or scanners. Conditions that can be altered include joint position (i.e. angle of knee flexion, supine vs. standing) [254]; amount of load (non-weight bearing, partial or full weight bearing) [255]; direction of load (axial loading through the feet vs. torque through the lower leg) [79 255]; and movement (static vs. dynamic) [256].   1.10 Theme of dissertation and study objectives The theme of this dissertation is the relationship of knee alignment or trochlear morphology to patellofemoral OA. The main objective was to use MRI to quantify the cross-sectional associations between measures of patellofemoral alignment, tibiofemoral alignment, or morphology with patellofemoral OA. My primary overarching hypothesis was that knee   46  alignment (patellofemoral and tibiofemoral) and trochlear morphology are associated with prevalent patellofemoral OA.   1.10.1 Study #1: Systematic review In my first study (Chapter 2), I conducted a systematic review of the literature to determine what is currently known about the relationship of knee alignment or morphology to presence, severity, onset or progression of patellofemoral OA, using any kind of imaging modality.  1.10.2 Study #2: Alignment and morphology one year following ACL reconstruction In my second study (Chapter 3), I investigated the relationship of patellofemoral alignment and morphology to early post-traumatic patellofemoral OA in a sample of individuals one year following anterior cruciate ligament reconstruction (ACLR). I hypothesized that following ACLR, alignment and morphology are associated with early patellofemoral OA, and that the nature of these associations are similar to those seen in non-traumatic patellofemoral OA.  1.10.3 Study #3: Framingham community cohort: reference values and dose-response patterns In my third study (Chapter 4), I first determined reference values for patellofemoral alignment and morphology in a subsample of community-dwelling adults aged ≥50 years with no patellofemoral OA and no knee pain. To accomplish this I used an existing population-based cohort, the Framingham community cohort. Using the full sample (both with and without patellofemoral OA), I next explored the dose-response pattern of alignment and morphology   47  measures to the odds of having prevalent patellofemoral OA across the range of values of each alignment or morphology measure.   1.10.4 Study #4: The “MRI open-bore for osteoarthritis Vancouver” (MOOV) study  In my fourth study (Chapter 5 and Chapter 6), I developed methods for evaluating patellofemoral alignment in three dimensions (3D) in individuals standing and squatting. I then evaluated the reliability of these methods in individuals with patellofemoral OA and in asymptomatic controls. In a cross-sectional study, I then evaluated 3D patellofemoral alignment in individuals with patellofemoral OA and matched asymptomatic controls. I hypothesized that (i) the patella would be more laterally displaced and tilted, and more proximally translated, when standing compared to supine in individuals with and without patellofemoral OA, (ii) that malalignment would be more pronounced in the OA group in standing compared to supine, that (iii) the patella would be more laterally displaced and tilted, and more proximally translated, in patellofemoral OA compared to matched controls in all positions, and (iv) that between-group differences would be greater when transitioning from two-legged stance to one-legged stance.    48  Chapter 2 Is tibiofemoral or patellofemoral alignment or trochlear morphology associated with patellofemoral osteoarthritis? A systematic review.  2.1 Background As outlined in Chapter 1 (section 1.8), altered knee (patellofemoral or tibiofemoral) alignment, or trochlear morphology, may contribute to incidence and/or progression of knee OA by increasing joint stress beyond tissue capacity [41 83 87 93 191 235 238-240 242]. For example, a high riding patella (‘patella alta’) or laterally displaced or tilted patella may reduce joint contact area, resulting in elevated joint stress. Malalignment may also alter the location and direction of load at the joint, for example a laterally displaced patella might transfer greater load onto the lateral patellofemoral compartment [81 242]. Bony shape, such as a shallow trochlear groove, might reduce passive joint stability resulting in altered load patterns or reduced contact area.   Several studies have used radiographs, CT, or MRI to assess the relationship between alignment or morphology and features of patellofemoral OA [257-260]. However, results of those studies were not always consistent. For example, Davies-Tuck et al. [260] reported in a cross-sectional study of knee OA that individuals with larger (i.e. shallower) sulcus angles had greater patellofemoral joint cartilage volume. In contrast, Jungmann et al. [261] found those with larger sulcus angles had smaller cartilage volumes. Thus, the clinical relevance of these measures as potential risk factors for patellofemoral OA remains unclear.     49  2.1.1 Study aims The aim of this study was to evaluate the associations of knee (patellofemoral and tibiofemoral) alignment or morphology to presence, severity, onset or progression of patellofemoral OA, using a systematic review.  2.2 Methods 2.2.1 Study design and protocol registration The protocol for this systematic review was prospectively registered with PROSPERO (#CRD42014007382) [262]. I followed the PRISMA guidelines to report this review [263].   2.2.2 Ethics Ethical approval was not required for this study because the study did not recruit participants.  2.2.3 Eligibility All studies were included that were English-language quantitative peer-reviewed publications that met criteria under three broad subheadings: participants, variables of interest, and outcome measures.  2.2.3.1 Participants  Studies that included participants with a patellofemoral OA diagnosis or who had features associated with patellofemoral OA using common imaging methods, either alone or with concurrent tibiofemoral joint involvement. I included studies that did not specify a diagnosis of patellofemoral OA for inclusion into their study (particularly studies with a sample comprised of   50  various musculoskeletal knee conditions) if the mean/median age was at least 50 years, or the lower limit of range was of at least 40 years. This age criterion helped ensure patellofemoral joint features were likely attributable to an OA process, and facilitated comparisons to knees of a similar age group.   2.2.3.2 Variables of interest  Studies included estimates of patellofemoral or tibiofemoral alignment or trochlear morphology using radiographs, CT or MRI, including (but not limited to) patellar displacement, patellar tilt, patellar height, frontal plane tibiofemoral alignment, and trochlear dysplasia. We did not exclude studies based on the imaging method employed. If a clinical intervention was used in the study, baseline data were used to avoid confounding.  2.2.3.3 Outcome measures  Any patellofemoral OA feature seen on imaging. Examples include osteophytes or joint space narrowing (using radiographs); and cartilage volume, cartilage damage, or bone marrow lesions (using MRI).  2.2.4 Search strategy To ensure a systematic review did not already exist, I searched the Cochrane Database of Systematic Reviews (CDSR, via OvidSP), DARE (via OvidSP), PROSPERO, and PEDro. I conducted the initial search in February 2014 and updated this in January 2015: MEDLINE (OvidSP, 1946 -), CINAHL (EBSCO, 1982-), EMBASE (OvidSP, 1974-), PEDro, Cochrane Central Register of Controlled Trials (CENTRAL, OvidSP), SPORTDiscus (EBSCO, 1837-),   51  Web of Knowledge (Web of Science platform, 1864-), Google Scholar, and conference proceedings (Papersfirst and Proceedingsfirst, 1993-). I hand-searched reference lists of retrieved articles and relevant reviews; contacted authors for additional references; and searched clinicaltrials.gov for registered studies. I screened citing articles of retrieved articles via Google Scholar and Web of Science. All potential references were imported and deduplicated in Endnote X7 (Thomson Reuters, Carlsbad, California, U.S.).   I developed search strings with input from a co-author (K. Crossley) and a reference librarian (C. Beck). I used three broad concepts: patellofemoral osteoarthritis; alignment or morphology; and imaging. Subject headings were exploded where appropriate, and scope notes examined for additional possible terms. I developed the first search strategy in Medline then adapted it for the other databases using their respective indexing vocabularies and syntax (see Appendix A).   A co-investigator (J. Stefanik) and I screened all titles and abstracts independently to identify potential studies. Where consensus could not be reached, a third reviewer (K. Crossley) was consulted. Full articles were then retrieved for final screening using the same procedure.  2.2.5 Data extraction A co-investigator (J. Stefanik) and I extracted all data independently. Data extracted included: age, sex, BMI, target population, sample size, follow-up period if longitudinal, alignment and morphology measures, and patellofemoral OA features. I contacted authors in the event of missing information.    52  2.2.6 Quality assessment There is no agreed upon criterion measure to assess quality in observational studies. Therefore, I used a 14-item checklist to evaluate risk of bias (see Appendix B). This tool was adapted from Downs and Black [264] and informed by previously published modifications to the tool [265 266], recent systematic reviews of quality assessment tools for observational studies [267 268] and the STROBE guidelines for reporting of observational studies [269].   I removed several items from the original index due to poor performance of the items in the original study (items 9, 11, 12, 13, 14, 17, 19, 26) [264]. I further removed items that were designed for evaluating intervention studies and therefore were not relevant for the current systematic review (4, 8, 21, 22, 23, 24, 26). Finally, I removed item 10 (reporting exact p-values) in favor of item 7 (reporting measures of variability), and removed item 16 (reporting if data dredging occurred) because exploratory research would be appropriate at this stage of the research cycle, and because results were only subsequently analysed and reported in this systematic review if a measure was used in at least three different studies (see Section 2.2.7). Finally, I added three items from Langham et al.’s index [265]: item 3 ‘Is the patient sample representative of patients treated in routine clinical practice?’; item 4 ‘Is there information on possibility of selection bias present in study?’; and item 17 ‘Was a sample size calculation reported?’.   A co-investigator (J. Stefanik) and I independently appraised study quality. Disagreements were resolved by a third arbitrator (K. Crossley). Studies with less than half of the quality assessment   53  tool items endorsed (i.e. total score of <7) were deemed to have a high risk of bias and hence, excluded from further analysis [229].  2.2.7 Statistical analyses Measures of central tendency and distribution of all alignment or morphology measures were extracted, and presented as a function of patellofemoral OA prevalence or severity where available. Only measures used in at least three studies were analysed and reported. We planned a meta-analysis in the event that measures could be statistically pooled across multiple studies. A best evidence synthesis, in which evidence is prioritized based on levels of evidence [270], was undertaken where meta-analysis was not possible. I synthesized the evidence using methods based on Tanamas et al [186]. Specifically, I rated the synthesized evidence as ‘strong’ if there were multiple high-quality cohort studies; ‘moderate’ if there was either one high-quality cohort study and more than two high-quality case-control studies, or more than three high-quality case-control studies; ‘limited’ if there were either one or two case-control studies, or multiple cross-sectional studies; and ‘insufficient’ if there was not more than one cross-sectional study (see Table 2.1). Evidence was summarized as ‘conflicting’ if findings in <75% of the trials were inconsistent. I further defined study size as small (less than 25 participants), medium (25 to 100), or large (more than 100 participants).     54  Table 2.1 Best evidence synthesis guidelines. Based on Tanamas et al.[186] Level of evidence Study size Strong Generally consistent findings in multiple high-quality cohort studies  Large >100 participants Moderate Generally consistent findings in 1 high-quality cohort study and >2 high-quality case-control studies; or in >3 high-quality case-control studies    Limited Generally consistent findings in a single cohort study; 1 or 2 case-control studies; or multiple cross-sectional studies  Medium 25 – 100 participants Conflicting Inconsistent findings in <75% of the trials    Insufficient No studies or 1 cross-sectional study Small <25 participants   2.3 Results 2.3.1 Search strategy The search revealed 9,500 citations. Limiting to English-language titles and removing duplicates rendered 5,994 citations. After searching titles and abstracts, we obtained 35 full text articles. Nineteen studies met our criteria, and one more was added following reference list reviews. Four studies had insufficient quality for inclusion [271-274], leaving 16 papers in total (see Figure 2.1).  2.3.2 Study characteristics Of the 16 studies included, five [98 197 257 260 275] were longitudinal (follow-up times 1.5 to 3 years), and the remaining 11 were cross-sectional (see Table 2.2). The 16 publications investigated 12 study samples. Six of the studies were nested within four large multi-centre cohort studies [257-259 261 275 276]. The target population was patellofemoral OA in two studies [146 277]; knee OA (or at high risk of developing knee OA) in 11 papers [98 197 251   55  258-261 275 276 278 279]; and various knee conditions in three studies [257 280 281]. No studies investigated patellofemoral OA onset as an outcome.   Figure 2.1 Flow chart for study inclusion. Reprinted with permission13.                                                   13 Reprinted from Arthritis Care and Research, Volume 68, Macri EM et al., “Is tibiofemoral or patellofemoral alignment or trochlear morphology associated with patellofemoral osteoarthritis? A Systematic Review”, pp 1453-70, 2016, with permission from John Wiley and Sons.   56  Table 2.2 Study characteristics Listed alphabetically by first author, except where multiple papers are from a single study. Paper Target population n (% women)  Age x (SD) BMI x(SD) Knee assessed PF OA  definition Follow-up Ali 2010[281]  various knee conditions in outpatient orthopaedic clinic, >40 yo 51 (49) range (40,87) ? symptomatic none - Cahue 2004[197] MAK, community, symptomatic TFJOA KL ≥ Gr. 2 in 1 or both knees 217 (71) 68.4 (10.8) 30.6 (6.1) both OARSI  OST ≥2 & JSN≥2 18 m Elahi 2000[278]  same sample as Cahue et al. 2004 292 (71) 66 (11) 31 (7) dominant  (see Cahue) (see Cahue) Crossley 2009[146] community, symptomatic PFOA vs. healthy, ≥40yo 28 (79) 55.1 (7.2) 27.3 (4.1) most symptomatic (OA), right (control) OARSI  OST ≥2, or OST1 +JSN≥1 - Davies-Tuck 2008[260] outpatient & community, ACR clinical & radiological knee OA* 100 (61) 63 (10) 29 (5) symptomatic or least symptomatic if bilateral OA  OST ≥2 (Burnett) 24 m Teichtahl 2008 [260] same sample as Davies-Tuck et al, 2008[38] 99  (60)    OST≥2 23 (2) m Hunter 2007 [257] Health ABC, community, knee pain onset ≤1y prior to baseline, 70-79 yo 2y prior 595 (60) 73.6 (2.9) 28.8 (4.9) both OARSI OST ≥2, or OST1 +JSN≥2 36 m Im 2013 [279] outpatients, knee pain & KL≥1 all compartments 251 (94) 66.8 (9.6) 26.1 (3.7) right, unless left only affected KL≥1 - Jungmann 2013 [261] OAI, symptomatic knee OA, progression sub-sample 135 (80) 64.1 (9.4) 28.3 (4.3) right none - Kalichman et al, 2007 [276] BOKS, subset with ACR clinical & radiological knee OA 213 (41) 66.7 (9.3) 31.4 (5.5) most symptomatic OST ≥2 - Kalichman 2007 [259]  same sample as Kalichman et al. 2007[43] [44]      - Mofidi 2014 [277] Pre-operative, isolated PFOA 34 (82) 61.2 (9.8) 29.5 (6.7) pre-operative ? - Otsuki 2014 [251] Pre-operative, varus knee OA 85 (75) 73.8 (6.4) 25.9 (3.5) pre-operative JSN ≥2 - Stefanik 2010 [275] MOST, 50 – 79 yo with or high risk of knee OA 907  (63) 62 (8) 30.0 (4.8) random OST≥2 or JSN ≥2 30 m Stefanik 2012 [258] same sample as Stefanik et al. 2010[41] 881  (63)     - Tsavalas 2012 [280] MR department, various knee conditions 201  (76) med 65 r 52,82 ? symptomatic≠  ICRS≥1 - Acronyms: ACR (American College of Rheumatology); BML (bone marrow lesions); BO (bisect offset); BOKS (Boston Osteoarthritis of the Knee Study); CA (Merchant’s congruence angle); GH (general hospital); ICRS   57  (International Cartilage Repair Society); JSN (joint space narrowing); KL( Kellgren & Lawrence);; m (months); MAK (Mechanical Factors in Arthritis of the Knee cohort); med (median); MR (magnetic resonance imaging); OAI (Osteoarthritis Initiative); OARSI (Osteoarthritis Research Society International); PFOA (patellofemoral osteoarthritis); r (range); TFJOA (tibiofemoral joint osteoarthritis); TKR (total knee replacement); WORMS (whole organ magnetic resonance imaging score); yo (years old). *not clear how this was determined, as max 82% of sample has definite osteophytes reported in paper ^reports 67 had lateral PFOA, 19 had PFOA defined as OST + JSN, though only mention “asymmetric PFOA” and do not clarify how many had ‘symmetric PFOA’. Also only assessed dominant leg, so may not have assessed the OA knee +Mixed description ISR vs. modified ISR, most likely ISR reported in these 2 studies  ≠ Reports one or both knees had MRIs, but analysis appears to be done by person and not by knees – unclear which knee was analyzed  In total, 2,892 individuals (66% women) were included: sample size ranged from 28 [146] to 907 [275] and the proportion of women ranged from 41% [259 276] to 94%[279]. Two studies reported a mean age of 74 years [251 257]; 12 papers had a mean age in the 60s [98 197 258-261 275-280]; one study’s mean age was 55 years [146]; and one study [281] did not report mean age (range 40-87 years). Mean BMI ranged from 25.9 kg/m2 [251] to 31.4 kg/m2 [259 276].    Where a radiographic threshold for patellofemoral OA diagnosis was specified, two papers used osteophytes [259 276]; and seven papers used osteophytes and/or joint space narrowing [98 146 257 258 260 275 279]. One study used MRI to define patellofemoral OA based on cartilage damage [280], and one study used computed tomography (CT) to define patellofemoral OA on a four-point grading system [251]. Several studies included knee pain to define knee OA [98 257 259 260 276], but only one specified anterior knee pain [146].  Alignment or morphology was measured using radiography in eight papers [98 197 257 260 275 277-279]; MRI in seven papers [146 258 259 261 276 280 281]; and CT in two papers [251 277]. Frontal plane tibiofemoral alignment was measured (all using radiographs) in five studies [98 197 251 278 279]. Sagittal plane alignment (patellar height) was measured in six papers, two on   58  radiograph [275 277] and four on MRI [259 276 280 281]. The remaining alignment measures were evaluated in an axial plane in six papers, one on radiograph [257], four on MRI [146 259 276 280], and one on CT [251]. Morphology was measured in ten papers, three on radiograph [257 260 277], six on MRI [258 259 261 276 280 281], and one also on CT [277] (see Appendix C).  2.3.3 Quality assessment Table 2.3 presents quality assessment findings, including the four lower quality studies that were removed from further analysis. Due to substantial study design heterogeneity, I did not pool results across studies, and instead conducted a best evidence synthesis.     59  Table 2.3 Quality assessment scores, in order of publication year.  Listed by first author (nb studies highlighted in grey were not included in Best Evidence Synthesis) Paper Year 1* 2 3 4 5 6 7 8 9 10 11 12 13 14 Total 'yes' Mofidi [277] 2014               10 Otsuki [251] 2014               11 Im [279] 2013            11 Jungmann [261] 2013         10 Stefanik [258] 2012             11 Tsavalas [280] 2012            11  Ali [281] 2010         7 Stefanik [275] 2010          10 Crossley [146] 2009           11 Davies-Tuck [260] 2008            12 Teichtahl [98] 2008            13 Hunter [257] 2007             11 Kalichman [276] 2007         12 Kalichman [259] 2007            12 von Eisenhart-Rothe [271] 2006          4 Cahue [197] 2004            9 Elahi [278] 2000            7 Harrison [272] 1994            3 Iwano [273] 1990         6 Rose [274] 1982              4 Endorsements by item (included studies only)  16 12 8 16 15 11 8 14 15 14 13 12 3 14  *Items: 1. Is the hypothesis/aim/objective of the study clearly described? 2. Are the characteristics of the patients included in the study clearly described? 3. Is the patient sample representative of patients treated in routine clinical practice? 4. Is there information on possibility of selection bias present in study? 5. Are the main outcomes to be measured clearly described in the Introduction or Methods section? 6. Were the main outcome measures used accurate (valid and reliable)? 7. Was an attempt made to blind those measuring the main outcomes of the intervention? 8. Are the main findings of the study clearly described?  9. Does the study provide estimates of the random variability in the data for the main outcomes? 10. Were the statistical tests used to assess the main outcomes appropriate?  11. Are the distributions of principal confounders in each group of subjects to be compared clearly described?  12. Was there adequate adjustment for confounding in the analyses from which the main findings were drawn? 13. Was a sample size calculation reported? 14. Did the study have sufficient power to detect a clinically important effect where the probability value for a difference being due to chance is less than 5%?      60  2.3.4 Axial plane patellofemoral alignment: limited evidence Numerous methods are used to assess axial plane alignment (see C.1), though generally they measure two aspects of lateral patellar malalignment: displacement and tilt. Overall, evidence is limited in the axial plane, with one large cohort study [257], three large cross-sectional studies [259 276 280] and two medium cross-sectional studies [146 251] (see Table 2.4).   Table 2.4 Axial plane alignment Listed alphabetically by first author, except where multiple papers are from a single study. Paper n* PFOA feature outcome Alignment assessed Results†                            p^ Point estimates            (distribution)           Additional descriptions Crossley 2009 [146] 28 OARSI y/n (PFOA vs. no OA)  LD (%)  BO (%)   LPTA (°)† Displacement: 17.04 (7.96)      v 5.08 (4.71) 71.55 (13.08)      v 51.69 (5.03) Tilt: 16.06 (9.47)      v 17.36 (2.60)   <0.001  0.000   0.626  Hunter 2007 [257]  595 change in JSN (36 months)   BO              PTA (°)  Displacement, 36 months: Lateral: 0.7 (0.3, 1.4) 0.6 (0.3, 1.3 0.4 (0.2, 0.8)    Medial:  1.2 (0.6, 2.4) 1.2 (0.6, 2.4) 2.2 (1.1, 4.5)   Tilt, 36 months: Lateral: 0.6 (0.3, 1.3) 0.6 (0.3, 1.3) 1.1 (0.6, 2.2)   Medial:  0.7 (0.4, 1.3) 0.5 (0.3, 1.0)    0.02       0.03       0.70      <0.0001    2nd (0.38 – 0.43) 3rd (0.43-0.47) 4th (0.47 – 0.60, most medial) quartile, odds of JSN progression   2nd 3rd 4th quartile, odds of JSN progression    2nd (3-4) 3rd (5-6) 4th (7-18) quartile, odds of JSN progression   2nd 3rd   61  Paper n* PFOA feature outcome Alignment assessed Results†                            p^ Point estimates            (distribution)           Additional descriptions 0.2 (0.1, 0.4) 4th quartile, odds of JSN progression Kalichman 2007 [276] 213 OST, JSN y/n  BO                     LPTA   Displacement: Lateral: 2.2 (0.8, 6.0) 4.2 (1.6, 11.3) 8.3 (3.1, 22.3)   0.9 (0.5, 1.5) 1.3 (0.8, 2.3) 3.1 (1.8, 5.3)  Medial: 0.9 (0.3, 2.3) 0.7 (0.3, 1.9) 0.2 (0.1, 0.6)   0.9 (0.3, 2.3) 0.7 (0.3, 1.9) 0.2 (0.1, 0.6)  Tilt: Lateral: 0.5 (0.2, 1.0) 0.3 (0.1, 0.7) 0.1 (0.04, 0.3)   0.4 (0.2, 0.6) 0.5 (0.3, 0.9) 0.3 (0.2, 0.5)   Medial: 1.5 (0.5, 4.3) 1.7 (0.6, 4.8) 2.2 (0.8, 5.8)   0.8 (0.5, 1.4) 0.9 (0.6, 1.6) 1.4 (0.9, 2.4)   <0.0001     <0.0001     0.0026     0.888      <0.0001     <0.0001      0.0259      0.1080   2nd (54.76 – 60.42) 3rd (60.47-66.67)   4th (66.67 – 100) quartile, odds of JSN  2nd 3rd 4th quartile, odds of OST  2nd 3rd 4th quartile, odds of JSN  2nd 3rd 4th quartile, odds of OST   2nd (14-17) 3rd (18 - 21)  4th (22 - 35) quartile, odds of JSN  2nd 3rd 4th quartile, odds of OST   2nd 3rd 4th quartile, odds of JSN   2nd 3rd 4th quartile, odds of OST Kalichman 2007 [259] 213 WORMS cartilage morphology, BML  BO     Displacement: Lateral: 0.8 (0.5, 1.4) 1.1 (0.7, 2.0) 3.4 (1.9, 6.1)    <0.0001      2nd  3rd  4th quartile, odds of cartilage damage   62  Paper n* PFOA feature outcome Alignment assessed Results†                            p^ Point estimates            (distribution)           Additional descriptions                 LPTA   0.9 (0.3, 2.5) 1.7 (0.7, 4.3) 3.2 (1.3, 7.7)  Medial: 0.9 (0.5, 1.6) 1.2 (0.7, 2.1) 1.5 (0.7, 2.1)   1.1 (0.5, 2.4) 0.6 (0.3, 1.4) 0.6 (0.3, 1.4)  Tilt: Lateral: 0.7 (0.4, 1.2) 0.8 (0.5, 1.5) 0.3 (0.2, 0.5)   0.4 (0.2, 0.8) 0.2 (0.1, 0.5) 0.1 (0.05, 0.3)  Medial: 1.3 (0.8, 2.2) 1.2 (0.7, 2.0) 0.7 (0.4, 1.2)   1.1 (0.5, 2.5) 0.6 (0.3, 1.5) 1.0 (0.5, 2.3)  <0.0001     0.0756     0.13      <0.0001     <0.0001     0.1553     0.72  2nd  3rd  4th quartile, odds of BML  2nd  3rd  4th quartile, odds of cartilage damage  2nd  3rd  4th quartile, odds of BML   2nd  3rd  4th quartile, odds of cartilage damage  2nd  3rd  4th quartile, odds of BML  2nd  3rd  4th quartile, odds of cartilage damage  2nd  3rd  4th quartile, odds of BML Otsuki 2014 [251] 85 JSN (Iwano) [273] PTA 12.2 (3.5, 42.9)  0.038 0.002 0.015 More lateral tilt in JSN Grade 2 (~10), Grade 3 (~12),  Grade 4 (~1831) v Gr. 1 (~7.5) (derived from graph, actual values not reported)   63  Paper n* PFOA feature outcome Alignment assessed Results†                            p^ Point estimates            (distribution)           Additional descriptions Tsavalas 2012 [280] 201 ICRS (any vs. no cartilage damage), OST, BML  LPD (mm)      LPFTA  (°)  Displacement: 0.2 (3.3) v -1.2 (2.7) r = 0.25  AUC 0.63 (0.56, 0.70)  Tilt: 5.8 (5.4) v 8.3 (4) r = -0.24  AUC 0.65 (0.58, 0.71)  0.0009 0.0002     0.0003 0.0005   ↑ lateral displacement associated with ↑ cartilage damage AUC indicates LPD ‘poor’ OA marker  ↑ lateral tilt associated with ↑ cartilage damage AUC indicates LPFTA ‘poor’ OA marker Acronyms: BL (baseline); BML (bone marrow lesions); BO (bisect offset); ICRS (International Cartilage Repair Society); iPFOA isolated PFOA; JSN (joint space narrowing); LD (lateral displacement); lPFOA (lateral compartment); LPFTA (lateral patellofemoral tilt angle - Laurin method); LPTA (lateral patellar tilt angle); mPFOA (medial compartment); OST (osteophytes); OARSI (Osteoarthritis Research Society International); PFOA (patellofemoral osteoarthritis); PTA (patellar tilt angle);WORMS (whole organ magnetic resonance imaging score)  *Sample size reported for full cohort or older age group only where two age groups recruited  †x(SD), OR(95% CI) or otherwise stated ^boldface indicates statistically significant  †NB larger numbers = less lateral tilt for this measure (remaining measures of lateral tilt, larger number = greater lateral tilt)  Five cross-sectional papers [146 251 259 276 280] (four studies) investigated three measures of patellar displacement and three of tilt. Greater lateral patellar displacement was associated with more severe patellofemoral OA or OA features. Two papers (one study) [259 276] that reported laterality of features found consistent and strong effects regarding lateral displacement in the lateral patellofemoral compartment. Findings in the medial patellofemoral compartment, however, were less clear, with a nonsignificant linear trend of greater displacement associated with higher odds of cartilage damage, a nonsignificant protective effect on BMLs, and possible protective effect on both joint space narrowing and osteophytes [259 276]. Greater lateral patellar tilt was associated with more severe patellofemoral OA or OA features in all but one study. Direction of tilt, as with displacement, appears to influence laterality of patellofemoral OA   64  features, with greater lateral tilt at higher odds of all lateral patellofemoral compartment features. The relationship between lateral tilt and medial patellofemoral OA features was less clear.   Longitudinally, at a three year follow-up, the most laterally displaced patellae had higher odds of lateral, and lower odds of medial, patellofemoral compartment joint space narrowing progression [257]. The most laterally tilted patellae had lower odds of medial patellofemoral compartment joint space narrowing progression, but no clear association with lateral joint space narrowing progression.  2.3.5 Sagittal plane patellofemoral alignment: limited evidence In the sagittal plane, a high positioned patella may increase the likelihood of reduced patellofemoral joint contact area and loss of stability within the trochlear groove, potentially leading to OA. Evidence is limited in the sagittal plane, with one large cohort study, three large cross-sectional papers (two samples), and two medium cross-sectional studies (see Table 2.5). In this review Insall Salvati (or modified Insall Salvati) ratio was the most common method for measuring patella height.  Table 2.5 Sagittal plane alignment. Listed alphabetically by first author, except where multiple papers are from a single study. Paper n* PFOA feature outcome Alignment assessed  Results†                            p^ Point estimates             (distributions)           Additional descriptions Ali 2010 [281] 51 ICRS cartilage defects (3/4 v 0) ISR   1.08(0.11) v 0.98(0.14)  0.14    Kalichman 2007 [276] 213 OST, JSN y/n  ISR+    Lateral: 1.6 (0.7, 3.7) 1.4 (0.6, 3.2) 2.8 (1.2,6.4)  0.01    2nd (0.88 – 0.98) 3rd (0.98 – 1.12)   65  Paper n* PFOA feature outcome Alignment assessed  Results†                            p^ Point estimates             (distributions)           Additional descriptions   1.7 (1.0, 2.9) 1.2 (0.7, 2.1) 1.7 (1.0, 2.84)  Medial: 2.0 (0.7, 6.0) 2.1 (0.7, 6.2) 2.5 (0.9, 7.1)  1.3 (0.8, 2.2) 1.2 (0.7, 2.0) 1.1 (0.7, 1.9)   0.01     0.13    0.60 4th (1.13- 1.71) quartile, odds of JSN   2nd  3rd  4th quartile, odds of OST  2nd 3rd 4th quartile, odds of JSN  2nd  3rd  4th quartile, odds of OST Kalichman 2007 [259]  WORMS cartilage morphology & BMLs  ISR+  Lateral: 1.1 (0.7, 1.9) 1.7 (1.0, 2.9) 2.0 (1.2, 3.6)    0.9 (0.4, 2.4) 0.6 (0.3, 1.7) 2.5 (1.1, 5.4)  Medial: 1.4 (0.8, 2.4) 2.0 (1.2, 3.3) 2.0 (1.2, 3.4)   1.1 (0.5, 2.4) 0.9 (0.4, 2.1) 1.1 (0.5, 2.5)  0.0068      0.003     0.0109     0.0016     2nd (0.88 – 0.98) 3rd (0.98 – 1.12)  4th (1.13 – 1.71) quartile, odds of cartilage damage  2nd  3rd  4th quartile, odds of BML   2nd  3rd  4th quartile, odds of cartilage damage  2nd  3rd  4th quartile, odds of BML  Mofidi 2014 [277] 34 iPFOA v varus knee OA  ISR 1.3 (0.3) v 1.0 (0.2) 0.0001  Stefanik 2010 [275] 907 WORMS progres-sion ≥0.5 for cartilage damage, BMLs  (BL and 30 months)  ISR           Lateral: 1.2 (0.8, 1.7) 1.3 (1.0, 1.9) 2.4 (1.7, 3.3)    1.2 (0.8, 1.8) 1.4 (0.9, 2.1) 2.9 (2.0, 4.3)   <0.0001      <0.0001     2nd (0.99 – 1.09) 3rd (1.10 – 1.21)  4th (1.22-1.66) quartile, odds of cartilage damage  2nd  3rd  4th quartile, odds of BML   66  Paper n* PFOA feature outcome Alignment assessed  Results†                            p^ Point estimates             (distributions)           Additional descriptions                          1.3 (0.8, 2.2) 1.4 (0.8, 2.2) 3.5 (2.3, 5.5)  Medial: 1.0 (0.7, 1.4) 1.2 (0.9, 1.6) 1.5 (1.1, 2.0)   1.1 (0.8, 1.6) 1.5 (1.1, 2.1) 1.3(0.9, 1.8)   1.5 (1.0, 2.4) 1.4 (0.9, 2.2) 2.2 (1.4, 3.4)  Lateral, 30 months: 0.9 (0.5, 1.6) 1.5 (0.9, 2.5) 2.1 (1.2, 3.5)  1.6 (0.9, 3.0) 1.7 (0.9, 3.2) 2.3 (1.2, 4.3)  Medial, 30 months:  1.9 (1.0, 3.4) 2.1 (1.2, 3.7) 2.0 (1.1, 3.6)   2.0 (1.0, 3.8) 1.7 (0.9, 3.3) 1.4 (0.7, 2.8)  <0.0001     0.0085     0.04     0.0001     0.0009    0.0009     0.005     0.45   -2nd  3rd  4th quartile, odds of SBA  2nd  3rd  4th quartile, odds of cartilage damage  2nd  3rd  4th quartile, odds of BML  2nd  3rd  4th quartile, odds of SBA   2nd  3rd  4th quartile, odds of cartilage damage 2nd  3rd  4th quartile, odds of BML  2nd  3rd  4th quartile, odds of cartilage damage  2nd  3rd  4th quartile, odds of BML Tsavalas 2012 [280] 201 ICRS (any cartilage damage vs. none) IISR  1.01 (0.12) v 0.98 (0.11)   0.1686  Acronyms: BL (baseline); BML (bone marrow lesions); ICRS (International Cartilage Repair Society); iPFOA isolated PFOA; ISR (Insall-Salvati Ratio); JSN (joint space narrowing); OST (osteophytes); OARSI (Osteoarthritis Research Society International); PFOA (patellofemoral osteoarthritis); SBA (subchondral bone attrition);WORMS (whole organ magnetic resonance imaging score)  *Sample size reported for full cohort or older age group only where two age groups recruited  +Mixed description ISR vs. modified ISR, most likely ISR reported in these 2 studies †x(SD), OR(95% CI) or otherwise stated     67  In four papers (three studies) investigating knee OA, higher positioned patellae were associated with patellofemoral OA features cross-sectionally [259 275-277]. In terms of laterality, odds ratios were consistent and significant in the lateral patellofemoral compartment across all measures of patellofemoral OA features, while at the medial patellofemoral compartment some odds ratios were smaller or less precise. In the two studies investigating samples with various knee-related clinical conditions [280 281] patellar height was no different in those with cartilage defects compared to those with no defects.   The only longitudinal study found that at 30 months, higher patellae had higher odds of cartilage damage and BML progression for both the lateral and medial patellofemoral compartments [275].  2.3.6 Frontal plane tibiofemoral alignment: strong evidence Frontal plane tibiofemoral alignment may influence loading through the patellofemoral joint via the alignment and action of the extensor mechanism. In this review, three papers (two studies) [197 278 279] used long limb radiographs to measure mechanical axis (hip-knee-ankle); and two [98 251] measured anatomical axis (tibiofemoral angle) with either long limb or standard radiographs to describe knees as either varus or valgus (see Table 2.6). With two cohort (one large, one medium) and three cross-sectional studies (one large, two medium) investigating four samples with symptomatic knee OA, there was strong evidence of a relationship between frontal plane alignment and patellofemoral OA features.     68  Table 2.6 Frontal plane tibiofemoral alignment. Listed alphabetically by first author, except where multiple papers are from a single study. Paper n* PFOA feature outcome Alignment assessed Results†                        p^ Point estimates          (distribution)           Additional descriptions Cahue 2004 [197]  217 ↑ JSN (OARSI), 18 months  HKA 18 months: 1.98 (1.15, 3.43)    1.85 (1.00, 3.44)    1.25 (0.79, 1.96)    1.64 (1.01, 2.66)         BL varus(<0°), odds of any PF JSN progression  BL varus, odds of isolated medial PF JSN progression  BL valgus (>0°), odds for ‘any’ JSN progression  BL valgus, odds of isolated lateral PF JSN progression. Elahi 2000 [278]  292 OARSI OST ≥2 & JSN≥2, asymmetric y/n  HKA  ≤0.0006    <0.02    0.0066 iPFOA & combined OA more likely valgus than isolated TFJOA  lPFOA more likely valgus & mPFOA more likely varus  lateral JSN more likely valgus & medial JSN more likely varus Im 2013 [279]  251 JSW, KL 1-4 HKA   β -0.15 (-0.11,0.19)   β 6.11(4.46, 7.78) β 3.94(2.36, 5.52) β 2.84 (1.21, 4.47) 0.131   <0.001 varus NA to PF JSW  PFJ KL Gr. 4 ↑ varus v KL 1, KL 2, or  KL3  Otsuki 2014 [251]  85 JSN (Iwano) [36] TFA   0.015 <0.001 0.003 TFA more varus in JSN Grade 2 (~182.5),  Grade 3 (~185),  Grade 4 (~183) v Gr. 1 (~181) (derived from graph, actual values not reported) Teichtahl 2008 [98]  99 cartilage volume annual change,  23 months  TFA  β 3.5 (-24.6, 31.7)   23 months: β 2.8 (-23.1, 28.6)    β 0.7 (-6.0, 4.5)  0.80    0.83    0.79  BL TFA NA with BL patella cartilage volume   BL TFA NA with follow-up patella cartilage volume  BL TFA (180.8[5.8]°) NA with annual   69  Paper n* PFOA feature outcome Alignment assessed Results†                        p^ Point estimates          (distribution)           Additional descriptions     β -23.4 (-38.7,-8.1)      0.003  change in patella cartilage volume (142[157] mm3)   Patella cartilage volume loss of 23.4 (38.7, 8.1) mm3 per 1° change (↑ valgus) Acronyms: BL (baseline); BML (bone marrow lesions); BO (bisect offset); HKA (hip-knee-ankle angle); ICRS (International Cartilage Repair Society); iPFOA isolated PFOA; ISR (Insall-Salvati Ratio); JSN (joint space narrowing); KL( Kellgren & Lawrence); LD (lateral displacement); lPFOA (lateral compartment); LPFTA (lateral patellofemoral tilt angle - Laurin method); LPTA (lateral patellar tilt angle); mISR (modified Insall-Salvati Ratio); mPFOA (medial compartment); NA no association; NS (not significant); OST (osteophytes); OARSI (Osteoarthritis Research Society International); PFOA (patellofemoral osteoarthritis); PTA (patellar tilt angle);  PTI (patellotrochlear index); SBA (subchondral bone attrition); TFA (tibiofemoral angle); TFJOA (tibiofemoral joint osteoarthritis); WORMS (whole organ magnetic resonance imaging score)  *Sample size reported for full cohort or older age group only where two age groups recruited  +Mixed description ISR vs. modified ISR, most likely ISR reported in these 2 studies †mean(SD), OR(95% CI) or otherwise stated  In samples with combined (both patellofemoral and tibiofemoral) knee OA, two studies [251 279] found higher grades of patellofemoral joint space narrowing were cross-sectionally associated with greater varus angles, while one study [98] found no cross-sectional association between frontal plane alignment and cartilage volume. The only study in this review that also included isolated patellofemoral OA within their sample [278] showed isolated patellofemoral OA knees were more likely valgus compared to combined or isolated tibiofemoral OA. This was also the only study to evaluate patellofemoral compartment laterality, and those with lateral patellofemoral OA were more likely valgus than those with medial patellofemoral OA, and vice versa.   Two studies investigated frontal plane alignment longitudinally [98 197]. At 18 months follow-up, Cahue et al. [197] found 55% patellofemoral joint space narrowing progression, with   70  baseline varus knees at higher odds of isolated medial or ‘any’ progression, while valgus knees had higher odds of lateral joint space narrowing progression. In the second longitudinal study [98], at 23 months follow-up, baseline frontal plane alignment did not predict patellofemoral cartilage volume change, but for every degree change towards valgus there was a mean loss of 23.4 mm3 (8.1, 38.7) patellar cartilage volume.  2.3.7 Trochlear morphology: strong evidence Measures of trochlear morphology are generally designed to evaluate structural stability offered by the trochlear groove to the patella, and the most commonly used measure in this review was sulcus angle. Nine papers (eight studies) used five different measures of trochlear morphology. With two cohort studies (one large and one medium) and six cross-sectional studies (four large and two medium), synthesized results provide strong evidence of a relationship between measures of trochlear dysplasia and increased severity of patellofemoral OA features.   Table 2.7 Trochlear morphology. Listed alphabetically by first author, except where multiple papers are from a single study. Paper n* PFOA feature outcome Morph-ology measure Results†                                  p^ Point estimates                (distribution)                     Additional descriptions Ali 2010 [281] 51 ICRS severe cartilage defects (3/4) vs. none (0) SA (°) LTI (°) TD(mm) TFR (%)  154 (16) v 148 (9)   21.5 (5.7) v 21.5 (2.8) 3.4 (2.5) v 4.3 (1.7) 63 (12) v 62 (15)  0.21 0.15 0.23 0.29     Davies-Tuck 2008 [260]  100 change in PF cartilage volume (mm3), 24 months   SA Total patella cartilage: 15.4 (1.8, 29) ↑ in total cartilage volume per 1° ↑ in SA Lateral patella cartilage: 7.2 (-1.2,15.5) ↑ per     1° ↑ in SA Medial patella cartilage:  0.03    0.09   0.003     71  Paper n* PFOA feature outcome Morph-ology measure Results†                                  p^ Point estimates                (distribution)                     Additional descriptions 9.1 (3.1,15.0) ↑ per 1° ↑ in SA 24 months: -0.6 (-5.2, 4) total -0.3(-3.1,2.5) lateral -0.2(-2.3, 1.9) medial  >0.8 >0.8 >0.8 Hunter 2007 [257] 595 change in JSN, 36 months   SA Lateral, 36 months: 0.6 (0.3, 1.3) 0.7 (0.3, 1.6) 2.1 (1.0, 4.4)   Medial, 36 months: 3.4 (1.6, 7.3) 3.0 (1.4, 6.4) 1.5 (0.6, 3.7)   0.04      0.07  2nd (126 - 129°) 3rd (130 - 135°)  4th (136 – 156°) quartile, odds of JSN progression  2nd  3rd  4th quartile, odds of JSN progression Jungmann 2013 [261]  135 WORMS, cartilage volume SA  (>170° vs. normal)         TD  (<3 mm vs. normal)        TFR  (<0.4 vs. normal) Total PF WORMS:  12.2 (1.1) v 8.6 (0.6) Patella cartilage damage:  4.8 (0.3) v 3.8 (0.2) Patella BME: 19/22 v 66/113 Patella subchondral cysts:  6/22 v 11/113 Patella cartilage volume:  795 (158) mm3 v 1,259 (94) mm3 Total PF WORMS: 11.2 (0.5) v 5.7 (0.6) Patella cartilage damage: 4.5 (0.2) v 3.1 (0.2) Patella BME: 65/85 v 20/50 Patella subchondral cysts: 17/85 v 0/50 Patella cartilage volume:  900 (72) mm3 v 1671 (95) mm3  Total PF WORMS:  12.3 (0.9) v 8.3 (0.5) Patella cartilage damage: 4.9 (0.3) v 3.7 (0.2) Patella BME: 24/30 v 61/105 Patella subchondral cysts: 4/30 v 13/105 Patella cartilage volume:  1,004 (137) mm3 v 1,237 (73) mm3  0.003  <0.05  <0.05  <0.05  <0.001   <0.001  <0.05  <0.05  <0.05  <0.001   <0.001  <0.05  <0.05  <0.05  0.26      72  Paper n* PFOA feature outcome Morph-ology measure Results†                                  p^ Point estimates                (distribution)                     Additional descriptions Kalichman 2007 [276] 213 JSN, OST y/n SA Lateral: 1.5 (0.7, 3.3) 1.6 (0.7, 3.6) 1.4 (0.6, 3.2)   1.6 (1.0, 2.7) 1.8 (1.1, 3.1) 1.5 (0.9, 2.6)  Medial: 1.4 (0.5, 4.0) 1.7 (0.6, 4.9) 3.2 (1.2, 8.7)   1.5 (0.9, 2.4) 1.7 (1.0, 2.9) 1.7 (1.0, 2.8)  0.12     0.08     0.016     0.051  2nd (114-119°) 3rd (120-124°) 4th (125-155°) quartile, odds of JSN   2nd  3rd  4th quartile, odds of OST  2nd  3rd  4th quartile, odds of JSN  2nd  3rd  4th quartile, odds of OST Kalichman 2007 [259]  WORMS cartilage, BML SA Lateral: 2.1 (1.2, 3.5) 1.8 (1.0, 3.0) 2.8 (1.6, 4.8)    4.6 (1.8, 2.1) 1.8 (0.6, 5.2) 3.6 (1.4, 9.6)  Medial: 1.7 (1.0, 2.8) 1.6 (0.9, 2.6) 2.4 (1.4, 4.1)   1.6 (0.7, 3.8) 1.2 (0.5, 3.1) 2.0 (0.9, 4.7)  0.0009      0.007     0.0016     0.06  2nd (114 - 119°)  3rd (120-124°) 4th (125-155°)  quartiles,  odds of cartilage damage  2nd  3rd  4th quartile, odds of BML  2nd  3rd  4th quartile, odds of cartilage damage  2nd  3rd  4th quartile, odds of BML Mofidi 2014 [277] 34 iPFOA vs. varus knee OA SA 134 (7) v 125 (7) 0.001  Stefanik 2012 [258]  881 WORMS cartilage damage, BML SA       Lateral: 1.0 (0.7, 1.4) 1.1 (0.8, 1.5) 1.5 (1.1, 2.1)     0.004       2nd (125.14-131.12) 3rd (131.13-136.54) 4th (>136.54°) quartile, odds of cartilage damage    73  Paper n* PFOA feature outcome Morph-ology measure Results†                                  p^ Point estimates                (distribution)                     Additional descriptions                LTI                     TA (°)            1.0 (0.7, 1.4) 1.0 (0.7, 1.5) 1.6 (1.1, 2.3)  Medial: 1.1 (0.8, 1.4) 1.0 (0.7, 1.3) 1.3 (1.0, 1.7)   1.1 (0.8, 1.6) 1.0 (0.7, 1.4) 1.1 (0.7, 1.5)  Lateral: 1.7 (1.2, 2.4) 1.9 (1.4, 2.7)  2.6 (1.9, 3.7)    1.1 (0.8, 1.7) 1.5 (1.0, 2.3) 2.3 (1.5, 3.3)  Medial: 1.3 (1.0, 1.8), 1.5 (1.1, 2.0), 1.6 (1.2, 2.1)   1.3 (0.9, 1.8) 1.1 (0.8, 1.6) 0.9 (0.6, 1.3)  Lateral: 1.3 (0.5, 1.6) 1.4 (0.9, 2.5) 2.0 (1.2, 3.5)    1.3 (0.9, 1.9) 1.4 (1.0, 2.0), 1.5 (1.1, 2.2)  Medial: 1.2 (0.9, 1.6) 1.0 (0.7, 1.3) 1.2 (0.9, 1.6)  0.007     0.08     0.84     <0.0001      <0.0001     0.004     0.39     0.0002      0.01     0.37    2nd  3rd  4th quartile, odds of BML  2nd  3rd  4th quartile, odds of cartilage damage  2nd  3rd  4th quartile, odds of BML  2nd (24.92 – 28.11) 3rd (21.70 – 24.91) 4th (<21.7) quartile, odds of cartilage damage  2nd  3rd  4th quartile, odds of BML  2nd  3rd  4th quartile, odds of cartilage damage  2nd  3rd  4th quartile, odds of BML  2nd (2.96 – 4.94) 3rd (0.87 – 2.95) 4th (<0.8) quartile odds of cartilage damage  2nd  3rd  4th quartile odds of BML  2nd  3rd  4th quartile, odds of cartilage damage   74  Paper n* PFOA feature outcome Morph-ology measure Results†                                  p^ Point estimates                (distribution)                     Additional descriptions  1.1 (0.8, 1.6) 1.0 (0.7, 1.4) 0.8 (0.6, 1.1)  0.08  2nd  3rd  4th quartile, odds of BML Tsavalas 2012 [280] 201 ICRS (any damage vs. none) SA      TD    141.1 (7.7) v 134 (6.4) r = 0.44  AUC 0.76 (0.69, 0.81)   5.8 (1.4) v 6.9 (1.2) r = -0.418  AUC 0.75 (0.68, 0.80)  <0.0001 <0.0001     <0.0001 <0.0001   ↑SA associated with ↑ cartilage damage   AUC indicates SA good PFOA marker   ↓TD associated with ↑ cartilage damage  AUC indicates TD good PFOA marker  Acronyms: BME (bone marrow edema); BML (bone marrow lesions); ICRS (International Cartilage Repair Society); JSN (joint space narrowing); iPFOA (isolated patellofemoral osteoarthritis); KL( Kellgren & Lawrence); LTI (lateral trochlear inclination); OARSI (Osteoarthritis Research Society International); PFOA (patellofemoral osteoarthritis); SA (sulcus angle); TD (trochlear depth); TFA (tibiofemoral angle); TFR (trochlear facet ratio); WORMS (whole organ magnetic resonance imaging score). *Sample size reported for full cohort or older age group only where two age groups recruited  †x(SD), OR(95% CI) or otherwise stated ^boldface indicates statistically significant    There were generally consistent findings that shallower trochleae, regardless of outcome measure, were cross-sectionally associated with indicators of worse patellofemoral OA such as greater cartilage damage, lower cartilage volume, higher prevalence / worse bone marrow lesions, joint space narrowing or osteophytes. These findings were consistent across samples with knee OA [258 259 261 276 280] and those with isolated patellofemoral OA compared to tibiofemoral OA controls [277]. In samples with ‘various knee-related conditions’, results were similar in direction although less precise and thus did not consistently achieve statistical significance [280 281]. Papers evaluating patellofemoral compartment laterality in knee OA [258 259 276] showed those with shallower trochleae had worse cartilage damage in both lateral and   75  medial patellofemoral compartments, although odds of BMLs were greater on the lateral patellofemoral compartment compared to the medial. Lateral trochlear inclination was only used in one study investigating knee OA [258], and it demonstrated larger effect sizes than both sulcus angle and trochlear angle. In contrast to these overall findings, Davies-Tuck et al. [260] found wider sulcus angles were cross-sectionally associated with higher medial patellofemoral cartilage volume and total patella cartilage volume when investigating compartment specific patellofemoral OA in people with knee OA.  Two studies investigated sulcus angle in samples with knee OA longitudinally [257 260]. Davies-Tuck found no association between baseline sulcus angle and change in patellar cartilage volume over two years. At three years, Hunter et al. [257] found those in the middle (2nd and 3rd) quartiles of sulcus angle had higher odds of medial patellofemoral compartment joint space narrowing progression compared to those with the narrowest sulcus angle. Thus, over a two to three year period there is no clear relationship between trochlear morphology and patellofemoral OA progression.  2.4 Discussion In this systematic review, there was strong evidence of a relationship between knee alignment or trochlear morphology and patellofemoral OA features in people with knee OA (patellofemoral, tibiofemoral, or combined) or other knee conditions. However, most studies did not specifically recruit participants diagnosed with patellofemoral OA. Results are based largely on cross-sectional studies. Follow-up times in longitudinal studies were short (at most three years), thus ability to establish causality between alignment or morphology and patellofemoral OA is limited.    76  The results of this review are consistent with current theories of knee OA. It is possible that malaligned patellofemoral joints could result in reduced patellofemoral joint contact area, resulting in increased joint stress, which could lead to incidence or progression of joint disease. Rare instances of conflicting results in this review could be explained by methodological differences such as sample recruitment criteria, sample size, choice of patellofemoral outcome measure, or choice of imaging modality.  For example, Kalichman et al. [276] found sulcus angle had smaller (and non-significant) odds ratios when using lateral patellofemoral joint space narrowing on radiographs as an outcome compared to when using lateral cartilage damage on MRI in the same sample [259]. Joint space narrowing on radiograph is a proxy for cartilage thickness, and may be less sensitive than MRI methods [98], potentially explaining the different findings.  2.4.1 Methodological considerations Heterogeneity and methodological variations limit generalizability across studies. Methods for estimating alignment, morphology and patellofemoral OA features varied substantially across studies. Image acquisition methods also varied, for example radiographs varied by acquisition plane, knee flexion angle (0° to 50°), and weightbearing status. This is relevant for patellofemoral OA since pain tends to increase in weight bearing at 20° – 30° flexion as the patella engages with the trochlea [77], and alignment differs with weight bearing conditions, possibly due to pain avoidance strategies [282] or because of muscular recruitment patterns [283].     77  Choice of reference for comparison and/or lack of a healthy comparison group may have affected results or limited interpretation of study findings. For example, converting continuous measures into categories may result in loss of fidelity [284] and has no biological basis for group distinction [186 285]. Using the first category as a reference (most or least of a given feature) does not necessarily represent healthy or typical [186].   Alignment and morphology measures from the present study cannot be easily compared to values for healthy or asymptomatic knees. Several studies have reported values for participants with asymptomatic knees [146 286-289]. However, these values may vary substantially based on imaging modality used, slice taken (in the case of MRI/CT), position of joint, weight bearing status, and the alignment method used. Many cut-points used in the literature to define ‘abnormal’ do not explain or justify how they were derived [281 290], and none have been validated for patellofemoral OA.  Crossley et al. [146] was the only study included in the present systematic review that recruited a healthy comparison group. It was therefore the only study where effect sizes for alignment could be estimated between those with and without patellofemoral OA. This study found a Cohen’s d in lateral displacement of 1.83 and in bisect offset of 2.00, with the patellofemoral OA group more laterally displaced in both measures. The Cohen’s d in lateral patellar tilt angle was 0.19 and the difference between the two groups was not significant.     78  2.4.2 Limitations There are several limitations to this review.  Pooling of the data was not possible due to study design heterogeneity. In addition, publication bias [291] is impossible to avoid, however my search strategy was extensive and included searching of grey literature in an attempt to mitigate this risk. I limited this review to English language papers due to lack of resources for translating.   Quality assessment tools for observational studies with adequate validity and reliability have not yet been developed and thus may not sufficiently rate bias within and among studies included in this review [264 268], nor are they capable of teasing out content-specific nuances in methodological quality. For example, in the frontal plane, mechanical axis measured on long limb radiographs are generally accepted as being superior to anatomic axis measures using standard AP radiographs.   Finally, there are limitations to using ‘levels of evidence’ to summarize the state of the literature [292], since this taxonomy relies solely on the overall study design (ranking cohort, case-control, and cross-sectional) with no regard for internal validity of those studies. The present review reveals strong evidence of a relationship between patellofemoral OA features and both trochlear morphology and frontal plane tibiofemoral alignment. However, in the sagittal and axial planes, evidence was labelled ‘limited evidence’ on account of only a single longitudinal study, yet there were six publications for both planes. These included cross-sectional studies of large sample sizes nested within well-designed cohort studies that had higher quality assessment scores than the studies included in the frontal plane, for example. Further, these studies consistently showed a cross-sectional association between patellofemoral OA features and alignment in the sagittal   79  and axial planes. Thus, the levels of evidence terminology may be both confusing and inadequate.  2.4.3 Clinical implications With overall strong evidence of a relationship between knee alignment or trochlear morphology and patellofemoral OA, targeted clinical interventions with potential to affect knee alignment may alter the course of patellofemoral OA in those with malalignment or trochlear dysmorphia. Importantly, knee malalignment may be influenced by muscular strength and motor control [78 209 293], or passive interventions like taping [146] or bracing [214]. For example, hip abductor muscle weakness [209] may cause femoral internal rotation and adduction, leading to relative lateral patellar displacement and tilt [78 79]; and quadriceps muscle weakness [192 213] can influence patellar positioning through similar mechanisms. Given the strong evidence in this systematic review of a relationship between altered alignment or morphology and patellofemoral OA presence or severity, it is advisable that in clinical settings, patellofemoral OA patients be evaluated for malalignment as well as potential contributing factors such as muscular weakness or altered motor control.      80  Chapter 3 Lateral displacement, sulcus angle and trochlear angle are associated with early patellofemoral OA following ACL reconstruction  3.1 Background In Chapter 2, a systematic review revealed strong evidence of a relationship between altered alignment or morphology and patellofemoral OA presence or severity [294]. Patellofemoral alignment may be modifiable [146 147 293 295], and since early intervention may offer the greatest potential for modifying disease outcomes [91], quantifying patellofemoral alignment or morphology in early OA could guide more effective intervention strategies. The relationship of patellofemoral alignment or morphology to early patellofemoral OA has not been specifically evaluated.   Anterior cruciate ligament (ACL) injury and/or reconstruction (ACLR) provides a good model for evaluating early OA, since post-traumatic OA is instigated by a specific event and develops relatively quickly following injury or surgery [27]. OA is also prevalent following ACL injury, occurring in approximately half of all ACL injured knees within 10 – 20 years of trauma, regardless of whether or not ACLR is performed [191 296 297]. Further, the most common early manifestation of knee OA following ACLR is isolated to the patellofemoral joint (12% and 20% at one and 10 years post-ACLR, respectively) [27 62]. Prevalence of patellofemoral OA does not differ significantly when ACLR is performed using bone-patellar tendon-bone (BPTB) graft compared to hamstring tendon (HT) graft, though overall knee OA rates may be higher[298 299].    81  ACL injury or ACLR can lead to substantial loss of health-related quality of life [28], and appears to be compromised even further in those who go on to develop post-traumatic OA [62 300-302]. Importantly, the impact of OA on younger people may be greater, affecting their ability to participate fully in occupations and other meaningful activities, and lowering quality of life for a substantially longer period of time [303].   Altered knee biomechanics may accelerate development of post-traumatic patellofemoral OA [62 191] since knee biomechanics, including patellofemoral maltracking, are altered following ACL injury and reconstruction (regardless of whether BPTB or HT is used) [62 191 304-307]. If patellofemoral OA following ACLR constituted a ‘malalignment phenotype’, this population might benefit from distinct intervention strategies aimed at secondary prevention. This provides the rationale to evaluate the relationship between alignment or morphology and patellofemoral OA early after ACLR.  3.1.1 Study aims In this study, I investigated measures of patellofemoral alignment and trochlear morphology in relation to radiographic features of patellofemoral OA one year following ACLR. Based on the findings of the systematic review in Chapter 2, I hypothesized that in patellofemoral OA, the patella would be positioned more proximal and would be more laterally displaced and tilted; and the trochleae would be shallower, compared to ACLR knees without patellofemoral OA.    82  3.2 Methods 3.2.1 Study design This cross-sectional study was an ancillary analysis of an existing cohort [27].  3.2.2 Ethics Ethical approval for this observational study was granted by The University of Melbourne (ID: 1136167) and The University of Queensland (ID: 2012000567 and ID: 2013001448) Human Research Ethics Committees. Participants provided written informed consent prior to participation. All procedures conformed to the Declaration of Helsinki principles.  3.2.3 Participants Through a medical chart review, consecutive patients were identified who had undergone arthroscopic single-bundle hamstring tendon autograft ACLR approximately 12 months previous and who were aged 18-50 years at the time of surgery. Exclusion criteria were: (i) previous injury/symptoms in the ACL-injured knee; (ii) ≤ five years between ACL injury and reconstruction; (iii) ≥15 months since ACLR at enrolment; (iv) inability to read or speak English; (v) subsequent injury or follow-up surgery to ACLR knee; (vi) presence of any other condition influencing daily function; and (vii) contra-indications for imaging.  3.2.4 Surgery – ACL reconstruction Arthroscopic single-bundle ACLR with hamstring tendon autograft procedures were performed between July 2010 and August 2011 by one of two orthopaedic surgeons in Melbourne, Australia (Mr. T. Whitehead, Mr. H. Morris). Surgical details were reported previously [27]. At the time of   83  surgery, the surgeon noted whether any patellofemoral articular cartilage lesions were present (defined as Outerbridge grade ≥2 [308]). Participants underwent ACLR at a mean 14 months (median 3 months) after injury. All patients received similar physiotherapy treatment following surgery.  3.2.5 Radiography: defining patellofemoral OA Posteroanterior, lateral and skyline radiographs were obtained for both knees in all participants. To assess the patellofemoral joint, lateral radiographs were acquired in full weight bearing with a SynaFlexer frame (Bioclinica/Synarc) holding knees flexed 30° and feet externally rotated 10°, and skyline radiographs were obtained in non-weight bearing with knees flexed 30°. For this study, I defined early patellofemoral OA as a definite osteophyte. This is equivalent to Kellgren Lawrence grade ≥ 2/osteophyte, a new paradigm to assess early OA proposed by Felson et al [88 114 133]. I chose to use radiography to define early OA in this study because criteria for MRI-defined early patellofemoral OA are not well established, and because MRI features of early OA such as patellofemoral bone marrow lesions and minor cartilage defects are prevalent even in young asymptomatic knees [159].  Two trained observers (K. Crossley, A. Culvenor) independently scored osteophytes for the medial and lateral patellofemoral compartments (both femoral trochlea and patella were read together, with the median ridge considered part of the medial compartment). Interrater reliability (kappa coefficient) was 0.78 (95% CI 0.67, 0.88) [27]. Consensus was used to resolve any discrepancies. Osteophytes were read prior to the design of this ancillary study, so both raters were unaware at the time of scoring that alignment and morphology outcomes would be read.   84  3.2.6 MRI: patellofemoral alignment and trochlear morphology MRI images were used to measure patellofemoral alignment and trochlear morphology. All MRI scans were acquired on a 3.0 Tesla scanner (Philips Achieva, The Netherlands) using a 16 channel knee coil (Invivo) with knees near full extension. To estimate alignment, a proton density-weighted 3D VISTA sequence was used (repetition time/echo time 1300 ms/27 ms; field of view 150 mm2; 0.35 mm isotropic; echo train length 64 ms; scan time 6 min 11 s).   I selected a random subset of 20 participant images in order to develop my assessment protocol. Image slices were selected using InteleViewer software version 4-3-4-P59 (Intelerad, Canada). I selected three slices for each participant and exported them into .jpg format: the median patellar slice in both the sagittal and axial planes, and the axial slice with the most prominent posterior condylar line [258].   I used Motion Analysis software version 9.9.0.1 (eHAB, Australia) to estimate alignment and morphology. In the sagittal plane, I obtained two measures, the Insall Salvati ratio [250] and patellotrochlear index [281] (Figure 3.1). In the axial plane, I assessed two measures of patellar displacement (bisect offset [309] and lateral displacement [146]); and two measures of tilt (lateral patellar tilt angle [146] and patellar tilt angle [309])(Figure 3.2). Trochlear morphology estimates in the axial plane included sulcus angle, lateral trochlear inclination, and trochlear angle [309] (Figure 3.3). I read all MR images independently and was blinded to radiograph findings. I calculated all measures twice in order to evaluate intra-rater reliability.      85          Figure 3.1 Sagittal plane alignment a. Insall-Salvati Ratio: ratio of (i) tendon length (TL) from patella tendon attachments (posteriorly) at tibia and patella to (ii) longest patella length (PL) from inferior to superior patella. Larger number indicates higher position of patella. b. Patellotrochlear index: ratio of (i) length of trochlear cartilage (TC) overlap to (ii) patellar cartilage (PC) length. Smaller number indicates less cartilage overlap. Reprinted with permission.14        Figure 3.2 Axial plane alignment a. Bisect offset: percentage of the widest patella length (PL) that is lateral to the line through deepest part of trochlea running perpendicular to posterior condylar line (PCL). A higher percentage is more laterally displaced. b. Lateral displacement: percentage of patella that lies lateral to a perpendicular line to the anterior condylar line (ACL) that runs through the highest point on the lateral femoral condyle. A higher percentage is more laterally displaced. c. Lateral patellar tilt: angle (θ) between PCL and the interior bony margin of the patella lateral facet (PF). A higher angle is less lateral tilt. d. Patellar tilt: angle (θ) between PCL and PL. A higher angle is more lateral tilt. Reprinted with permission.15                                                    14 Reprinted from Knee Surgery, Sports Traumatology, Arthroscopy, Macri EM. et al., “Lateral displacement, sulcus angle and trochlear angle are associated with early patellofemoral osteoarthritis following anterior cruciate ligament reconstruction”, Published Online First: doi: 10.1007/s00167-017-4571-1 with permission from Springer. 15 Reprinted from Knee Surgery, Sports Traumatology, Arthroscopy, Macri EM. et al., “Lateral displacement, sulcus angle and trochlear angle are associated with early patellofemoral osteoarthritis following anterior cruciate ligament reconstruction”, Published Online First: doi: 10.1007/s00167-017-4571-1 with permission from Springer.   86          Figure 3.3. Trochlear morphology a. Sulcus angle: angle (θ) joining the lateral (LF) and medial (MF) bony facet margins and the lowest point of the sulcus. Higher number indicates shallower sulcus. b. Lateral trochlear inclination: angle (θ) between posterior condylar line (PCL) and LF. A larger angle indicates a deeper sulcus laterally. c. Trochlear angle: angle (θ) between PCL and anterior condylar line (AL). Higher angle indicates deeper sulcus laterally. Reprinted with permission.16  3.2.7 Statistical analyses To evaluate intra-rater reliability, I obtained each alignment and morphology measure twice in the full sample. I created Bland-Altman plots to confirm homoscedasticity and to identify outliers [310]. I then performed repeated measures ANOVA to evaluate potential systematic error between measures, and estimated intra-rater reliability using intra-class correlation coefficients ICC(3,1), which I reported together with a standard error of measure (SEM) [311 312]. ICC values of ≥0.75 were defined as good, and ≥0.90 as excellent [311].  I performed logistic regression modelling [284] to evaluate the relationship of each alignment or morphology measure to prevalent early patellofemoral OA (medial, lateral, and any                                                  16 Reprinted from Knee Surgery, Sports Traumatology, Arthroscopy, Macri EM. et al., “Lateral displacement, sulcus angle and trochlear angle are associated with early patellofemoral osteoarthritis following anterior cruciate ligament reconstruction”, Published Online First: doi: 10.1007/s00167-017-4571-1 with permission from Springer.   87  compartment) in separate models. Age, sex, and BMI were considered for inclusion in the model. If any study participants with patellofemoral OA at one year had had cartilage damage noted at the time of surgery, sensitivity analyses were performed with those participants removed to account for the possibility of pre-existing OA. Sample size calculations were not done a priori because this was an ancillary analysis of an existing cohort. Statistical significance was defined as p<0.05. All statistical analyses were completed using Stata Intercooled 12.0 (StataCorp, Texas, USA).   3.3 Results 3.3.1 Participants Of 334 patients screened, 186 met eligibility criteria. Of those, eight were unreachable, 35 lived too far away to attend, and 31 declined invitation. One more participant was later excluded from analysis because their MRI revealed a full ACL graft tear (see Figure 3.4). This resulted in complete MR images for 111 participants (mean±SD age 30±8.5 years, 41 [37%] women, BMI 25.9±3.8 kg/m2) (see Table 3.2 for demographics by group). Median time between injury and surgery was 13 weeks (interquartile range 20; range 1, 1463).  The mean time between surgery and MRI acquisition was 14±2 months (median 13 months, range 11, 19). Demographics for those who were eligible but did not participate in the study (n=74) did not differ from the study sample (age, sex, pre-injury level of sports activity, time from injury to ACLR, or rate of concomitant injuries) (p>0.05).     88             Figure 3.4 Flow chart of participant recruitment. ^conditions included recent ankle fracture, recent Achilles tendon rupture, concurrent posterior cruciate ligament reconstruction, recent gastrocnemius tear, congenital talipes, large vision impairment, recent removal of thyroid, and current chemotherapy. *unable to undergo magnetic resonance imaging (MRI) because of distance. LARS= ligament augmentation and reconstruction system; BPTB = bone–patellar tendon–bone. Reprinted with permission17  Early patellofemoral OA was present in the ACLR knee of 19 participants (17%) compared to 8 participants (7%) in the contralateral knee (Table 3.2). Five of these participants had bilateral patellofemoral OA. Those with OA of the ACLR knee were older than those without, thus we accounted for age in all of our logistic regression models. Seven (37%) of the 19 with OA were aged over 40 years, one of whom had bilateral patellofemoral OA. Twelve participants had                                                  17 Reprinted from Arthritis and Rheumatology, Volume 67, Culvenor A et al., “Early Knee Osteoarthritis Is Evident One Year Following Anterior Cruciate Ligament Reconstruction: A Magnetic Resonance Imaging Evaluation”, pp 946-55, 2015, with permission from John Wiley and Sons   89  patellofemoral articular cartilage lesions assessed intra-operatively, two of whom had OA of the ACLR knee at the one year evaluation.   3.3.2 Alignment and morphology Intra-rater reliability was excellent, with coefficients ranging from 0.94 to 0.99 (Table 3.2).  Table 3.1 Reliability for alignment and morphology measures (N=111) Measure ICC(3,1) SEM Insall Salvati Ratio 0.99 0.02 Patellotrochlear Index (ratio) 0.98 0.02 Bisect Offset (%) 0.97 2.03 Lateral Displacement (%) 0.94 1.12 Lateral Patellar Tilt Angle ( ° ) 0.97 0.88 Patellar Tilt Angle ( ° ) 0.95 1.08 Sulcus Angle ( ° ) 0.95 2.74 Trochlear Angle ( ° ) 0.99 0.72 Lateral trochlear inclination ( ° ) 0.96 1.36   Table 3.2 provides summary statistics for all alignment and morphology measures. Logistic regression model results, including covariates from the final models, are reported in Table 3.3. Due to the relatively low number of osteophytes in medial and lateral compartments separately, compartment-specific evaluations were not performed.     90  Table 3.2 Demographics and summary of alignment and morphology measures (N=111)  Patellofemoral OA (n=19) Non-OA (n=92) p-value Age 36.3 (8.7) 29.2 (8.0) 0.00* Sex 6 (32%) 35 (38%) 0.60 BMI 26.7 (3.5) 25.7 (3.8) 0.26 Contralateral OA 5 (26%) 3 (3%) 0.00        History of trauma  2/5 3/3  Intraoperative cartilage lesions (Outerbride gr. ≥2) 2 (11%) 7 (8%) 0.67 Insall Salvati ratio 1.08 (0.19) 1.07(0.14) 0.82 Patellotrochlear index (ratio) 0.56 (0.10) 0.60 (0.14) 0.27 Bisect offset (%) 57.13 (8.01) 53.45 (5.19) 0.01 Lateral displacement (%) 3.38 (5.78) 4.05 (4.45) 0.57 Lateral patellar tilt angle ( ° ) 10.72 (5.51) 11.85 (4.26) 0.32 Patellar tilt angle ( ° ) 8.75 (4.59) 7.49 (4.55) 0.28 Sulcus angle ( ° ) 123.00 (10.45) 118.11 (7.50) 0.02 Trochlear angle ( ° ) 0.75 (3.15) -0.48 (2.44) 0.06 Lateral trochlear inclination ( ° ) 26.58 (5.36) 28.55 (4.64) 0.10 *bold indicates p-value ≤0.05  In the sagittal plane, there were no significant associations between alignment and patellofemoral OA (Table 3.3).   In the axial plane, bisect offset was larger (i.e. more laterally displaced) in those with early patellofemoral OA, with an odds ratio (OR) of 1.1 (95% CI 1.0, 1.2). The remaining measures were not significantly different.      91  Table 3.3  Logistic regression models for prevalent patellofemoral OA.   Covariates included in each model OR 95% CI p Insall Salvati ratio Age 1.35 (0.04, 44.74) 0.87 Patellotrochlear index Age 0.01 (0.00, 1.24) 0.06 Bisect offset Age 1.11 (1.02, 1.21) 0.02* Lateral displacement Age 0.98 (0.87, 1.10) 0.67 Lateral patellar tilt angle Age, LPTA2 0.96 (0.86, 1.07) 0.45 Patellar tilt angle Age 1.04 (0.92, 1.18) 0.49 Sulcus angle Age 1.09 (1.02, 1.17) 0.02 Lateral trochlear inclination Age 0.91 (0.81, 1.02) 0.10 Trochlear angle  Age 1.24 (1.00, 1.53) 0.05 *bold numbers indicate statistical significance at p ≤ 0.05  Two trochlear morphology measures were associated with early patellofemoral OA. Participants with a shallower sulcus angle had higher odds of prevalent patellofemoral OA (OR 1.1 [1.0, 1.2]); participants with a more anteriorly protruding lateral facet (i.e. higher trochlear angle) had higher odds of patellofemoral OA (OR 1.2 [1.0, 1.5]). Lateral trochlear inclination was not associated with patellofemoral OA.  3.3.3 Sensitivity analyses Two participants with patellofemoral OA had arthroscopically assessed patellofemoral chondral lesions. In sensitivity analyses with those two cases removed, bisect offset and sulcus angle model results did not change. The trochlear angle model OR did not change either but the association was no longer statistically significant (p=0.06).      92  3.4 Discussion The most important finding of the present study was that each of (i) patellar lateral displacement (measured by bisect offset), (ii) sulcus angle, or (iii) trochlear angle were associated with radiographic indices of early patellofemoral OA one year post-ACLR. In this sample of 111 participants, 17% had early patellofemoral OA. While data for comparison to asymptomatic controls are not available, this rate of early patellofemoral OA is substantially higher than the prevalence of symptomatic knee OA among non-obese men and women aged 25 – 44 living in the US, which ranges from 0.7% to 2.1% [15]. While longitudinal research is required to explore alignment and morphology as possible causative risk factors for patellofemoral OA, the findings of the present study may indicate that a malalignment phenotype of post-traumatic patellofemoral OA exists. Identifying and characterizing this phenotype may offer new insights for early intervention and secondary prevention strategies.   3.4.1 Alignment and patellofemoral OA: integration In the sagittal plane, observed Insall Salvati ratios were consistent with values previously reported one year after ACLR (1.1±0.2) [313], and with other musculoskeletal knee conditions including suspected ACL injury and patellofemoral OA [280 281 314]. For patellotrochlear index, results for this study’s OA group were the same as values reported for people under 40 years old with knee injury and mild cartilage defects (0.6±0.1), and slightly larger than those without cartilage defects (0.5 ±0.1) [281].  In this study, I did not observe a relationship between sagittal plane patellar alignment and early patellofemoral OA. In my previous systematic review (Chapter 2) [294], both cross-sectional and   93  longitudinal studies (over 2.5 years) found that patella alta is positively associated with patellofemoral OA features. This contrasts with findings in a study of post-traumatic OA, in which seven years after ACLR, those with moderate to severe patellofemoral OA had lower positioned patellae on radiographs (Insall Salvati ratio 0.86±0.1) than those without patellofemoral OA (0.95±0.1) [315]. Together, these data suggest that the relationship between sagittal plane patellar alignment and OA may differ when patellofemoral OA is associated with trauma as compared to non-traumatic patellofemoral OA. Patella alta and baja potentially alter patellofemoral joint stress in different ways [241]. Sagittal plane patellar alignment may also take longer than one year post-ACLR to cause detectable OA. At this time, the role of patella height in post-traumatic patellofemoral OA remains unclear.  Axial plane patellar alignment has not been reported in patients post-ACLR, with or without OA. Mean bisect offset for those with OA in our study (57.1±8.0%) was similar to values reported in other studies of knees with patellofemoral pain (59.6%) [316] but lower than other cohorts with patellofemoral pain (69±13%) [317] or patellofemoral OA (71.6±13.1%) [146]. Lateral patellar tilt angle was smaller (i.e. greater lateral patellar tilt, 10.7±5.5°) than knees with patellofemoral OA (16.1±9.5°) [146] yet similar to those with isolated patellofemoral OA (11.1±7°) [273]. Patellar tilt angle, in contrast, was slightly less laterally tilted in the present OA sample (8.8±4.6°) than in patellofemoral pain (12.5±7.6°) [317]. These contrasting findings could be explained by the different landmarks or methods used to determine alignment, and it is currently unknown which alignment measures (and methods) are best suited to an ACLR population.    94  3.4.2 Morphology and patellofemoral OA: integration While trochlear morphology post-ACLR has not been reported to our knowledge, previous studies have reported higher prevalence of trochlear dysplasia in individuals with ACL injuries [318 319]. In the present study, the trochlea measures were deeper on average than those reported in various knee conditions with or without patellofemoral OA [280 281]. It is unclear whether values measured in the present study are due to true differences from other study samples, or due to methodological differences. Regardless, the results of the present study indicate that a shallower trochlear groove may predispose the patient to post-traumatic patellofemoral OA. This is consistent with my systematic review that provides strong evidence that shallower trochleae are cross-sectionally associated with higher prevalence of non-traumatic patellofemoral OA [294].  3.4.3 Limitations There are several limitations to my study. First, the results of this study relate specifically to an early post-ACLR population, and thus generalizability is limited to this context and not to non-traumatic patellofemoral OA in the general population. Second, this is a cross-sectional study and associations do not infer causality. Pre-injury or pre-operative images of the reconstructed knee were not available, thus it is not possible to know whether, or to what extent, alignment changed as a result of the ACL injury or surgery. It is also possible that patellofemoral OA existed prior to, and independently of, ACL trauma. However, this is unlikely since there was a substantially higher proportion of ACLR knees with OA compared with the rate in contralateral knees. In addition, results did not change in sensitivity analyses to account for participants more likely to have had pre-existing patellofemoral OA (i.e. those with cartilage lesions assessed intra-  95  operatively). Older age in those with patellofemoral OA was also accounted for by using age as a covariate. As this is one of the first studies to explore patellofemoral alignment and trochlear morphology in early post-traumatic patellofemoral OA, I did not adjust for multiple testing. Due to a low number of osteophytes in medial and lateral compartments, I did not evaluate compartment-specific relationships.   The MR images in this study were taken statically and in a traditional closed-bore scanner and thus were non-weight bearing with quadriceps muscles relaxed and knees near full extension. Open-bore MR scanners would enable alignment to be assessed in full weightbearing, in different knee flexion angles, and with active and functional muscle recruitment patterns [83 283]. Since biomechanical abnormalities may be more pronounced with higher levels of loading or more challenging functional tasks [320-322], it is possible that weightbearing MRI may capture mechanical changes that more closely relate to OA onset.   3.4.4 Clinical implications In this study, patellar lateral displacement (bisect offset) and trochlear morphology (sulcus angle, trochlear angle) were associated with radiographic indices of early patellofemoral OA one year after ACLR. Further longitudinal studies are required to evaluate the role of alignment and morphology as risk factors for post-traumatic patellofemoral OA.   Clarifying the role of alignment and morphology in the onset of post-traumatic patellofemoral OA may inform intervention strategies in a subset of patients who may be at increased risk for developing post-traumatic patellofemoral OA. In particular, tailored interventions designed to   96  address malalignment (e.g. patellar taping, bracing, targeted neuromuscular control exercises) [146 147 293 295 323] are examples of approaches that could be tested in futures studies to prevent or delay post-traumatic patellofemoral OA.    97  Chapter 4 The Framingham Community Cohort: a population-based study of patellofemoral alignment and morphology  4.1 Background Despite the clinical importance of patellofemoral OA, risk factors and features associated with this important subtype of knee OA are not well understood [142 143 174]. It is unknown to what extent patellofemoral alignment and trochlear morphology contribute - either independently or together - to the onset or progression of patellofemoral OA [257 260]. Given that patellofemoral alignment may be modifiable with conservative treatments (e.g. bracing, taping, exercise) [146 293 295], a comprehensive evaluation of alignment and morphology as potential risk factors for patellofemoral OA could inform clinical management. In order to determine who may be at higher risk of onset or worsening of patellofemoral OA, it is necessary to first understand what typical values of alignment and morphology are and how they relate to patellofemoral OA cross-sectionally.    A critical gap in the literature is the absence of robust reference values for common measures of patellofemoral alignment and trochlear morphology in patellofemoral joints without OA or knee pain, as well as an understanding of their dose-response patterns (i.e. odds of having patellofemoral OA features across the range of alignment or morphology values) [285]. My systematic review (see Chapter 2) revealed that only one of 16 studies investigating patellofemoral alignment or morphology recruited an asymptomatic comparison group [146 294]. This is in contrast to other important medical conditions (e.g. femoroacetabular   98  impingement, osteoporosis, cardiovascular disease) where reference values have been defined for healthy or asymptomatic populations, and cut-points established, to guide interventions or identify those at highest risk of an outcome [289 324-331].   4.1.1 Study aims In this study, I aimed to: (i) determine reference values for patellofemoral alignment and trochlear morphology in individuals without MRI-defined features of patellofemoral OA or knee pain; and (ii) evaluate dose-response patterns of each of these alignment and morphology measures with prevalent MRI-defined features of patellofemoral OA and/or knee pain.   4.2 Methods 4.2.1 Study Design This cross-sectional study was an ancillary analysis of an existing cohort [162 332].  4.2.2 Ethics Ethical approval for this study was provided by the institutional review board of Boston University Medical Center (ID # H-22674). Participants provided written informed consent prior to participation.   4.2.3 Participants  The Framingham Community Cohort is a population-based sample of ambulatory (with or without walking aids), community-dwelling adults aged 50 years or older living in Framingham, Massachusetts. Details of the recruitment and enrollment process have been published previously   99  [332]. Briefly, participants were selected irrespective of knee pain or OA status, and were unaware of the purpose of the study. Individuals were excluded if they planned to move out of the area within the next five years, had bilateral total knee replacements, rheumatoid arthritis, dementia, terminal cancer, or had contraindications to MRI [162 332]. Within the cohort, 996 participants had MR images of at least one knee.  4.2.4 Image acquisition While lateral view radiographs were taken in addition to MRI scans, I used MR images for this study to quantify patellofemoral OA features. As described in Section 1.5.4, MRI is more sensitive in visualizing tissues related to OA (i.e. can directly image cartilage and BMLs) [160], and is more reliable to quantify patellofemoral OA features than radiographs [157 333]. Using MRI, I was able to use direct measures of cartilage morphology as an OA outcome measure (i.e. rather than using radiographic JSN as a proxy for cartilage damage). In addition, because this was a population-based study (i.e. included people with and without OA), the use of MRI for this study was appropriate in helping to address a knowledge gap regarding semiquantitative OA features and their relationship to pain.  MRI scans were acquired on a 1.5-tesla scanner (Siemens Healthineers, Erlangen, Germany) using an eight-channel phased array knee coil [162 332]. To assess axial alignment and morphology, axial reformatted 3D fast low angle shot sequences were acquired (see Table 4.1 for sequence parameters) [162 332]. To assess OA features, axial, sagittal and coronal fat suppressed, proton density weighted, turbo spin echo images were acquired (see Table 4.1) and sagittal T1 weighted spin echo images without fat suppression (see Table 4.1) [162 332] .   100  Table 4.1 MRI sequence parameters.          *FLASH = fast low angle shot; PDwTSE = proton-density weighted turbo spin echo; T1w SE = T1 weighted spin echo.   4.2.5 Patellar alignment and trochlear morphology I analyzed all measures of alignment and morphology based on previously established methods [258 275]. Prior to analysis, and in consultation with a second assessor (J Stefanik), I used a random sample of 30 left knee MR images (i.e. contralateral MR images, not read for semiquantitative scoring) from the Framingham data set to standardize assessment methods.   Using the established standardized protocol, I first selected the axial MRI slice with the largest (i.e. most posterior) posterior femoral condyles (i.e. adjacent slices demonstrated a smaller amount of posterior femoral condyle) (see Figure 4.1) [258]. Using this slice I evaluated trochlear morphology with four different measures: sulcus angle, lateral trochlear inclination, medial trochlear inclination, and trochlear angle [258] (see descriptions, Figure 4.1). I also selected the axial slice with the maximum mediolateral patellar diameter and used both slices to evaluate two patellar alignment measures: bisect offset and patellar tilt angle [309] (Figure 4.1). All measurements were made in OsiriX Lite version 8.0 (Pixmeo SARL, Switzerland).   3D FLASH* PDw TSE T1w SE Repetition time (ms) 16.8 3610  475 Echo time (ms) 7.64  40  24 Flip angle 15 - - Slice thickness (mm) 1.5  3.5 3.5 Gap thickness (mm) 0  0 0 Echo train length - 7 - Matrix 512 x 512 256 x 256 256 x 256   101   Figure 4.1 Alignment and morphology measures (nb some measures were previously reported in Figure 3.2 (a, d) and Figure 3.3 (a, b, c). A. Patellar tilt angle (PTA) is the angle between the PCL (transposed from the slice with the largest PCL) and patellar width line. Larger positive value means more lateral tilt.  Bisect offset (BO) is the % of the patella lateral to midline of the trochlea (the midline is transposed from the slice with the largest PCL to the slice with the widest patella). Larger value means more lateral displacement. B. Sulcus angle (SA) is the angle between the medial and lateral trochlear facets. Larger value means wider sulcus.  Lateral trochlear inclination (LTI) is the angle between the posterior condylar line (PCL) and the lateral trochlear facet margin. Larger value means more prominent lateral facet.  Medial trochlear inclination (MTI) is the angle between the PCL and the medial trochlear facet margin. Larger value means more prominent medial facet.   C. Trochlear angle (TA) is the angle between the PCL and the anterior condylar line. Larger positive value means lateral facet protrudes more anteriorly than the medial facet. C. Reprinted with permission.18     Using this standardized protocol, I determined alignment and morphology values in a random sample of 20 left knees. I repeated these measures after a 48 hour time period elapsed for the purposes of estimating intra-rater reliability. A second assessor (J Stefanik) repeated these measures one time for the purposes of evaluating inter-rater reliability. Once reliability was established, I evaluated alignment and morphology in the study knees.                                                   18 Reprinted from Osteoarthritis and Cartilage, Macri EM, et al “Patellofemoral morphology and alignment: reference values and dose-response patterns for the relation to MRI features of patellofemoral osteoarthritis” Published Online First: doi: 10.1016/j.joca.2017.06.005, 2017, with permission from Elsevier    102  4.2.6 MRI-defined patellofemoral OA grading Patellofemoral structural damage was defined and graded using MRI in this study. MRI features were scored by two experienced musculoskeletal radiologists (A Guermazi, F Roemer) using the ‘Whole Organ MRI Score’ (WORMS) [90 334]. Images were read for the right knee only when possible (n=983); the left knee (n=13) was read when the right knee images were either not acquired or were not readable. Inter-rater agreement (weighted kappa statistic, κ) for cartilage damage was κ =0.89; and for BMLs was κ= 0.85 [332]. For this study I evaluated four patellofemoral joint sub-regions defined in WORMS: medial and lateral patella, and medial and lateral trochlea (see Figure 4.2). The four sub-regions were combined to look at the entire patellofemoral joint as an outcome, and also by medial and lateral compartments separately. This decision was informed by knowledge that medial and lateral compartmental prevalence of OA features differs, and associations with alignment or morphology also differ by compartment [258 334].      103        Figure 4.2 Using the WORMS scoring method, the knee is subdivided into regions. Relevant to this study, the patella (left) is divided into medial (M) and lateral (L) regions, with the central median ridge included with M. The anterior femur (A, middle image) is scored as the femoral portion of the patellofemoral joint. The anterior femur is also divided into M and L regions (right), with the trochlear groove included with M. C=central, P=posterior, S=tibial spine. Reprinted with permission19.  The primary outcome for this study, and operational definition of patellofemoral OA, was full thickness patellofemoral cartilage damage (WORMS scores of either 2.5, 5, or 6) irrespective of tibiofemoral joint involvement (see Figure 4.3). Informed by previous work [64], I conducted secondary analyses using two additional patellofemoral OA features: (i) ‘any’ cartilage damage (WORMS scores ≥2); and (ii) ‘any’ bone marrow lesion (BMLs, WORMS score ≥1).      Figure 4.3 Using the WORMS scoring method, cartilage signal and morphology is divided into eight categories. Each knee region is scored independently. Reprinted with permission20.                                                  19 Reprinted from Osteoarthritis and Cartilage, Volume 12, Peterfy CG et al., “Whole-Organ Magnetic Resonance Imaging Score (WORMS) of the knee in osteoarthritis”, pp 177-90, 2004, with permission from Elsevier. 20 Reprinted from Osteoarthritis and Cartilage, Volume 12, Peterfy CG et al., “Whole-Organ Magnetic Resonance Imaging Score (WORMS) of the knee in osteoarthritis”, pp 177-90, 2004, with permission from Elsevier.   104  4.2.7 Knee pain outcomes  Symptomatic OA features may provide different and clinically relevant information and thus, were also evaluated. Study participants were asked “In the past month, have you had any pain, aching, or stiffness in your knees?” Additional questions were asked to identify those who reported at least ‘mild’ pain on most days of the previous month. These individuals were deemed to have ‘knee pain’. Since this definition of ‘knee pain’ could not necessarily be localized to the patellofemoral joint, I identified an additional subgroup of participants who had ‘knee pain’ but also reported at least mild pain when climbing stairs [144], an item from the WOMAC scale and a task known to increase the load at the patellofemoral joint. I confirmed allocation to this second group, ‘pain with stairs’ if pain was unilateral and affected only the scored knee, or if pain was bilateral and the pain severity reported for the scored knee was at least as severe as the pain in the contralateral unread knee.   In total, I evaluated three regional outcomes (patellofemoral joint as a whole, and medial and lateral patellofemoral compartments separately); three structural features (full thickness cartilage damage, the primary outcome measure; and also any cartilage damage and any BMLs), and two pain outcomes (knee pain; pain with stairs) that I used as outcome measures alone and also in combination with the structural features.  4.2.8 Statistical analyses  4.2.8.1 Reliability I evaluated intra-rater and inter-rater reliability by performing repeated measures analysis of variance (ANOVA) and then calculating intraclass correlation coefficient, ICC(2,1), with   105  absolute agreement using a two-way random effects model [335]. In addition, I calculated standard error of measure (SEM) based on the root mean square error of the ANOVA and from this determined the smallest detectable change at 95% confidence (SDC95). I defined ICC values of ≥0.75 as good, and ≥0.90 as excellent [311].  4.2.8.2 Reference intervals To generate reference values, I identified a subsample of knees with no full thickness cartilage damage on MRI in the patellofemoral joint as well as no knee pain (i.e. no symptomatic patellofemoral OA). I then calculated mean and standard deviation (SD) for all alignment and morphology measures, and provided reference values (mean ±two SD) for each measure [327 328]. If a measure differed by sex, I reported values separately for women and men.  4.2.8.3 Dose-response patterns To examine the dose-response relationship between each alignment or morphology measure and MRI-defined structural features of patellofemoral OA, I performed multivariable restricted cubic spline logistic regression [336]. The justification for using this method instead of examining the ranges of exposure in categories (e.g. quintiles) is because it was expected that only individuals with the most extreme values of alignment or morphology would have higher odds for prevalent OA. Categorizing can result in collapsing those at the extreme ends (who are at true higher odds of an outcome) within a broader category (at relatively lower odds), which can mask a true significant relationship [285]. Restricted cubic spline analysis is thus a more robust method for modeling dose-response patterns [337]. I used three knots for each alignment or morphology variable (10th, 50th and 90th percentile) [336] because I expected, based on biology, that odds   106  would be higher at one or both extreme ends of alignment or morphology values, but likely would not increase within mid-range values. In each model, I adjusted for age, sex and BMI. For all models, I compared those with full thickness cartilage damage to those remaining from the sample who had no patellofemoral joint full thickness cartilage and no pain.   After performing spline logistic regression, I exported the predicted odds ratios for full thickness cartilage damage for the entire range of the independent variable (alignment or morphology measure), using the median value of each variable for the sample with no patellofemoral OA and no pain as the referent point [330]. Dose-response patterns were presented graphically using line graphs of the predicted odds ratios for each alignment or morphology measure. Finally, I reported values for each variable where predicted odds ratios were significantly higher. This included the first value at which the odds ratio was significantly higher than unity, as well as values at which odds ratios achieved 1.5, 2.0 and 3.0. Multiple values were reported in this fashion because their clinical relevance is subjective and may vary with their intended use [338].    Using the same approach as above, I examined each alignment and morphology measure with other patellofemoral OA measures (i.e. any cartilage damage, any BMLs, knee pain alone, pain with stairs alone, plus all three structural outcomes combined with the two definitions of pain). Finally, I repeated these methods by medial and lateral compartments separately. No adjustments were made to statistical significance levels to account for multiple testing. All statistical analyses were done using Stata Intercooled 12.1 (StataCorp, TX, USA).    107  4.3 Results 4.3.1 Participants Of the 996 participants, one was excluded from this study due to having no patella in the imaged knee; one due to having no axial plane MR image; and nine due to inadequate coverage of the knee joint or poor image quality. This resulted in 985 participants for analysis: mean age was 64 ± 9 years old; 562 (57%) were women; and mean BMI was 28.5 ± 5.6 kg/m2 (see Table 4.2 for additional demographics). The sample was primarily Caucasian. Compared to those with no patellofemoral OA and no pain, the subsample with patellofemoral full thickness cartilage damage was about 3 years older (p<0.001), had 1.7 kg/m2 higher BMI (p<0.001), and had 10% more women (p=0.01).  Table 4.2 Participant characteristics. All values are N(%) unless otherwise noted.  Full sample (n=985) No patellofemoral OA, no pain* (n=563) Full thickness cartilage damage (n=255) Women  562 (57%) 305 (54.2%) 164 (64.3%) Age, mean(SD) 63.5 (8.8) 62.4 (8.2) 65.3 (9.1) BMI mean(SD) 28.5 (5.6) 27.7 (5.2) 29.4 (5.6) <25 N 267 (27.5%) 178 (31.8%) 55 (22.0%) 25-30  390 (40.16%) 230 (41.1%) 98 (39.2%) >30  314 (32.34%) 151 (27.0%) 97 (38.8%) Knee pain  222 (23.2%) 0 (0.0%) 83 (33.9%) Pain with stairs^ 164 (17.4%) 0 (0.0%) 64 (26.9%) Any cartilage damage 627 (64.6%) 274 (48.7%) 255 (100%) Full thickness cartilage damage 255 (26.3%) 563 (0%) 255 (100%) Any BMLs 385 (39.3%) 124 (22.0%) 200 (78.7%) *Neither patellofemoral joint full thickness cartilage lesions nor knee pain ^Knee pain, plus in WOMAC item #2 regarding pain during stairs, pain rated at least ‘mild’   Reliability was good to excellent for all measures. Inter-rater reliability ranged from 0.85 to 0.98, and intra-rater reliability ranged from 0.90 to 0.98 (see Table 4.3, Table 4.4).    108  Table 4.3 Inter-rater reliability (n=20) Measure Mean Difference  ICC(2,1)  [95%CI] SEM SDC95 Patellar tilt angle 0.42° 0.93 (0.84,0.97) 1.03° 2.86° Bisect offset ratio 0.02 0.93 (0.12, 0.99) 0.01 0.03 Trochlear angle 0.31° 0.98 (0.94,0.99) 0.37° 1.03° Sulcus angle 0.08° 0.96 (0.90,0.98) 1.52° 4.21° Lateral trochlear inclination 0.02° 0.85 (0.67,0.94) 1.45° 4.02° Medial trochlear inclination 0.90° 0.93 (0.83,0.97) 1.59° 4.41° ICC = intraclass correlation coefficient, SEM = stand error or measure, SDC95=smallest detectable difference    Table 4.4 Intra-rater reliability (n=20) Measure Mean Difference  ICC(3,1)  [95%CI] SEM SDC95 Patellar tilt angle 0.27° 0.92 (0.82,0.97) 1.16° 3.22° Bisect offset ratio 0.02 0.98 (0.96, 0.99) 0.01 0.03 Trochlear angle 0.24° 0.98 (0.95,0.99) 0.46° 1.28° Sulcus angle 0.59° 0.96 (0.91,0.99) 1.42° 3.94° Lateral trochlear inclination 0.73° 0.90 (0.77,0.96) 1.26°  3.49° Medial trochlear inclination 0.81° 0.95 (0.89,0.98) 1.40° 3.88° ICC = intraclass correlation coefficient, SEM = stand error or measure, SDC95=smallest detectable difference   4.3.2 Reference values Reference values are reported along with mean ± SD for all alignment and morphology measures (Table 4.5 for reference sample, i.e. no patellofemoral OA or pain; Table 4.6 for patellofemoral OA). Only two of the measures, medial trochlear inclination and patellar tilt angle, differed by sex and are reported separately.      109  Table 4.5 Reference values among those with no patellofemoral joint full-thickness cartilage damage or knee pain (n=563).   Mean (SD) 2.5 Centile 97.5 Centile Patellar tilt angle (°) 8.6(4.4) 0.0 17.2 Women (n=305)* 9.1(4.6) 0.1 18.1 Men (n=258) 8.1(4.1) 0.1 16.0 Bisect offset (%) 52.2(4.8) 42.8 61.6 Sulcus angle (°) 128.8(6.5) 116.0 141.5 Lateral trochlear inclination (°) 27.1(4.4) 18.6 35.7 Medial trochlear inclination (°) 31.5(5.1) 21.5 41.4 Women  32.2(5.1) 22.1 42.3 Men  30.6(4.8) 21.1 40.1 Trochlear angle (°) -0.6(2.7) -6.0 4.7 *Only medial trochlear inclination and patellar tilt angle differed by sex, thus are reported separately here   4.3.3 Dose-response patterns Dose-response patterns were curvilinear for all six predictors (two alignment and four morphology measures) (see Figure 4.4, Figure 4.5). Patterns were generally similar for the full patellofemoral joint and also by medial and lateral compartments separately. However, odds of having MRI-defined structural features of patellofemoral OA (e.g. full thickness cartilage damage) as a function of alignment or morphology were generally highest in the lateral compartment as compared with the medial compartment or the whole patellofemoral joint. Values below are therefore reported for the lateral compartment unless otherwise specified. Superimposing the reference values from the subsample with no full thickness cartilage damage or pain onto the graphs (Figure 4.4, Figure 4.5) illustrated the highest odds for having MRI-defined structural features of patellofemoral OA generally occur beyond those reference values in the extreme ranges of alignment or morphology.      110  Table 4.6 Reference values for sample with full thickness patellofemoral cartilage damage regardless of symptoms (‘Cartilage’, n=255); full-thickness cartilage damage plus knee pain (‘+Pain’, n=83); and full-thickness cartilage damage plus pain with stairs (‘+Stairs’, n=64).  Sample Mean (SD) 2.5th centile  97.5th centile Patellar tilt angle (°) Cartilage 8.1 (5.8) -3.2  19.4  Women (n=164)*  8.8 (5.8) -2.7  20.2  Men (n=91)  6.9 (5.4) -3.7  17.6   +Pain 8.7 (6.3) -3.6  20.9  Women (n=62)  9.3 (6.1) -2.7  21.2  Men (n=21)  6.9 (6.6) -6.0  19.8   +Stairs 9.3 (6.4) -3.2  21.8  Women (n=50)  9.9 (5.9) -1.6  21.3  Men (n=14)  7.4 (8.0) -8.2  23.0  Bisect offset (%) Cartilage  55.6 (8.4) 39.1  72.1   +Pain 56.0 (9.2) 37.9  74.1   +Stairs 56.6 (10.0) 37.0  76.2  Sulcus angle (°) Cartilage  130.6 (7.9) 115.1  146.0   +Pain 130.0 (7.0) 116.3  143.8   +Stairs 130.0 (7.4) 115.5  144.5  Lateral trochlear inclination (°) Cartilage  26.3 (4.7) 17.0  35.6   +Pain 26.3 (4.9) 16.8  35.9   +Stairs 26.0 (4.9) 16.4  35.7  Medial trochlear inclination (°) Cartilage 30.1 (5.9) 18.6  41.6  Women   30.6 (6.0) 18.8  42.3  Men   29.2 (5.6) 18.3  40.1   +Pain 30.3 (5.5) 19.5  41.1  Women   31.0 (5.6) 19.9  42.0  Men   28.4 (4.7) 19.2  37.6   +Stairs 30.3 (5.3) 19.9  40.7  Women   30.8 (5.6) 19.9  41.7  Men   28.4 (3.9) 20.8  36.1  Trochlear angle (°) Cartilage  -0.3 (2.9) -6.0  5.5   +Pain -0.2 (3.0) -6.0  5.7   +Stairs -0.2 (3.0) -6.0  5.6  *Only medial trochlear inclination and patellar tilt angle differed by sex, thus are reported separately here      111  Patellar tilt angle was associated with MRI-defined features of patellofemoral OA both with and without pain (Figure 4.6). The dose-response pattern was generally ‘U’ shaped, thus for most outcomes the odds rose with both lateral and medial tilt. Predicted odds of full thickness cartilage damage for the full patellofemoral joint were higher at or below 7° (i.e. less lateral tilt), but when the outcome was defined as full thickness cartilage damage combined with knee pain, the threshold was ≤6°, and when combined with pain with stairs it was even lower at ≤2° (Table 4.7). Upper threshold values for the same outcomes were ≥16°, ≥13° and ≥12°, respectively. Clinically relevant values (defined as odds ratio ≥ 1.5) for full thickness cartilage damage were achieved at or below 4° (medial compartment), and at or above 13° (lateral compartment) (Table 4.8). As with other measures, the dose-response shape persisted for secondary outcomes, with slightly differing odds ratios.          Figure 4.4 Dose-response patterns for alignment. Line graphs show predicted odds ratios for having patellofemoral OA (full thickness cartilage damage) against each alignment predictor, compared to the median predictor value of the sample with no patellofemoral OA and/or pain. Two vertical lines indicate   112  reference values for subgroup with no patellofemoral OA and/or pain (reference values reported in Table 4.5). Left to right: patellar tilt angle, median 8°; bisect offset, median 51%. Modified with permission21  Bisect offset was associated with all patellofemoral OA features, achieving significance at or above 52% for the lateral compartment, and achieving clinical relevance at or above 55%. Patterns and thresholds were similar for secondary outcomes.             Figure 4.5 Dose-response patterns for morphology measures. Line graphs show predicted odds ratios for having full thickness cartilage damage against each morphology predictor, compared to the median value of the sample with no patellofemoral OA and/or pain. The two vertical lines indicate reference values for subgroup with no patellofemoral OA and/ or pain (reference values reported in Table 4.5). From top, left to right: medial trochlear inclination, median 31°; lateral trochlear inclination, median 27°; sulcus angle, median 128°; trochlear angle, median = -1°. Modified with permission22                                                   21 Modified from Osteoarthritis and Cartilage, Macri EM, et al “Patellofemoral morphology and alignment: reference values and dose-response patterns for the relation to MRI features of patellofemoral osteoarthritis” Published Online First: doi: 10.1016/j.joca.2017.06.005, 2017, with permission from Elsevier 22 Modified from Osteoarthritis and Cartilage, Macri EM, et al “Patellofemoral morphology and alignment: reference values and dose-response patterns for the relation to MRI features of patellofemoral osteoarthritis” Published Online First: doi: 10.1016/j.joca.2017.06.005, 2017, with permission from Elsevier   113  Medial trochlear inclination was consistently associated with all patellofemoral OA features (i.e. full thickness cartilage damage, any cartilage damage, BMLs, pain alone or combined with structural features), with the exception of symptomatic lateral full thickness cartilage damage (i.e. when combined with either of the two pain definitions) (see Figure 4.5; Figure 4.6, Table 4.7, Table 4.8). The dose-response pattern was ‘reverse J’ shaped, indicating odds were higher with a less inclined (i.e. flattened) medial trochlear facet. Predicted odds became significantly higher at or below 30.5°, and first reached clinical relevance at or below 27.0° (see Table 4.7, Table 4.8).   Lateral trochlear inclination was associated with full thickness cartilage damage in a ‘reverse J’ shape, with higher odds at or below 26.5° (26.0° if combined with either pain definition), and achieving clinical relevance at or below 22.5°. The shape of the dose-response pattern generally persisted for secondary outcomes, with slightly differing odds ratios. One exception was that for knee pain alone, odds were higher at or above 37.0°. The association with full thickness cartilage damage was highest when pain with stairs was included in the outcome (see Figure 4.6).  Sulcus angle was associated with full thickness cartilage damage with a sulcus angle at or above 129.0°, and achieved clinical relevance at or above 133°. Findings were similar for secondary outcomes.    Trochlear angle was not associated with full thickness cartilage damage. Odds were higher only for medial compartment full thickness cartilage damage combined with knee pain, with a trochlear angle at or above 0.5°.    114  Table 4.7 Thresholds above or below which predicted odds ratios are significantly higher than unity for having patellofemoral OA features and/or pain.   Knee  pain Pain with stairs Full thickness cartilage Full thickness cartilage & knee pain Full thickness cartilage & pain with stairs Any  cartilage Any cartilage & knee pain Any cartilage & pain with stairs Any  BMLs Any BMLs & knee pain Any BMLs  & pain with stairs Patellar tilt angle PFJ ≤5.0°/ ≥12.0° ≥11.0° ≤7.0°/ ≥16.0° ≤6.0°/  ≥13.0° ≤2.0°/ ≥12.0° ≤7.0° ≤6.0°/  ≥13.0° ≥11.0° ≤7.0° ≤5.0°/ ≥12.0° ≤4.0°/ ≥11.0°  LAT    ≤6.0°/ ≥11.0° ≥10.0° ≤4.0° ≤7.0° ≤6.0°/  ≥12.0° ≤4.0°/  ≥11.0° ≤7.0°/ ≥18.0° ≤7.0°/ ≥12.0° ≤6.0°/ ≥11.0°  MED   ≤7.0°  ≤6.0°/  ≥16.0° ≤1.0°/ ≥13.0° ≤7.0°  ≤6.0°/  ≥15.0° ≥13.0° - ≤1.0°/ ≥12.0° ≤-2.0°/ ≥11.0° Bisect  PFJ ≥53.0% ≥52.0% ≥53.0%   ≥52.0% ≥53.0%  ≥53.0%  ≥54.0%  ≥54.0%  ≥54.0%  offset LAT    ≥52.0%   ≥52.0% ≥53.0%  ≥53.0%  ≥52.0%  ≥54.0%  ≥54.0%   MED   ≥54.0%   ≥52.0% ≥53.0%  ≥53.0%  ≥55.0%  ≥54.0%  ≥54.0%  Medial  PFJ ≤30.5° ≤30.5° ≤30.5° ≤30.5° ≤30.5° ≤30.5° ≤30.5° ≤30.5° ≤30.5° ≤30.5° ≤30.5° trochlear  LAT    ≤30.5° - - ≤30.5° ≤30.5° ≤30.5° ≤30.5° ≤30.5° ≤30.5° inclination MED   ≤30.5° ≤30.5° ≤30.5° ≤30.5° ≤30.5° ≤30.5° ≤30.5° ≤30.5° ≤30.5° Lateral  PFJ ≥37.0° ≤21.5° ≤25.0° - - - - - - - - trochlear  LAT    ≤26.5° ≤26.0° ≤26.0° - - ≤26. 0° - - ≤25.5° inclination MED   - - - - - - - - - Sulcus  PFJ ≥132.0° ≥131.0° ≥129.0°  ≥130.0° ≥131.0° ≥129.0°   ≥131.0° - ≥130.0°  - - angle LAT    ≥129.0° ≥131.0° - ≥129.0° - ≥130.0° ≥129.0° ≥129.0° ≥129.0°  MED   ≥130.0° ≥130.0° ≥130.0° ≥131.0° - - ≥135.0° - - Trochlear  PFJ ≥0.0° ≥1.5°  - - - ≥0.5°  ≥0.0° ≥3.0° - ≥0.5° ≥1.5° angle LAT    - - - ≥0.5°  ≥0.0°  - - ≥0.5°  ≥0.5°   MED   - ≥0.5°  - ≥0.5°  ≥0.0°  - - ≥0.5°  ≥1.5° PFJ = patellofemoral joint; LAT = lateral compartment; MED = medial compartment    115  Table 4.8 Thresholds for predicting clinically relevant odds ratios (OR) of having full thickness cartilage damage.    OR ≥1.5 OR ≥2.0 OR ≥3.0 Patellar tilt angle PFJ ≤3° /  ≥16° ≤1° / ≥20° ≤-1° / ≥27°  Lateral  ≤3° / ≥13° ≤2° / ≥15° ≤-1° / ≥18°  Medial ≤4° ≤2° ≤-1° Bisect offset PFJ ≥57.0% ≥59.0% ≥61.0%  Lateral  ≥55.0% ≥57.0% ≥59.0%  Medial ≥57.0% ≥60.0% ≥62.0%      Medial trochlear  PFJ ≤26.0° ≤23.5° ≤20.5° inclination Lateral  ≤27.0° ≤25.0° ≤22.0°  Medial ≤26.0° ≤23.0° ≤19.5° Lateral trochlear  PFJ ≤19.5° ≤15.5° - inclination Lateral  ≤22.5° ≤20.0° ≤17.0°  Medial - - - Sulcus angle PFJ ≥134° ≥137° ≥140°  Lateral  ≥133° ≥135° ≥138°  Medial ≥135° ≥137° ≥141° Trochlear angle PFJ - - -  Lateral  - - -  Medial - - -  PFJ = full patellofemoral joint; - indicates OR did not reach a priori selected threshold       116   Figure 4.6 Symptomatic patellofemoral OA dose-response patterns: all six measures of alignment and morphology, line graphs show predicted odds ratios for having full thickness cartilage damage of the lateral compartment only, alone and in combination with both pain definitions (compared to the median predictor value of the subsample with no patellofemoral OA and/or pain). Modified with permission23  4.4 Discussion This study determined reference values and dose-response patterns for alignment and morphology in knees with and without patellofemoral OA and pain, in a population-based sample of ambulatory adults aged ≥50 years. In comparing the reference values of those with                                                  23 Reprinted from Osteoarthritis and Cartilage, Macri EM, et al “Patellofemoral morphology and alignment: reference values and dose-response patterns for the relation to MRI features of patellofemoral osteoarthritis” Published Online First: doi: 10.1016/j.joca.2017.06.005, 2017, with permission from Elsevier   117  and without patellofemoral OA, the results of this study indicate large within-group variability and substantial overlap between groups. This precludes alignment and morphology from being used as diagnostic indicators for patellofemoral OA. In other conditions such as recurrent patellar dislocation [339] or femoroacetabular impingement [340], morphology cut-points have been identified that distinguish these conditions from their asymptomatic counterparts with reasonable sensitivity and specificity. This was not possible in the present study. It may be that alignment and morphology are not strong independent predictors of patellofemoral OA. Rather, they may become important under certain conditions, such as in occupations that substantially increase cumulative joint load, with resultant joint stress exacerbated by malalignment or instability. It may also be that alignment and morphology predict patellofemoral OA longitudinally more powerfully than they do cross-sectionally.  This study demonstrated a dose-response pattern for alignment and morphology with prevalent patellofemoral OA and/or knee pain. Odds increased monotonically for all measures except patellar tilt, which rose with both lateral and medial tilt. Associations were generally stronger when the lateral compartment was evaluated alone, as compared with the medial compartment or full patellofemoral joint. This is also the case when evaluating the association of pain with patellofemoral structural damage [334]. This study adds to the literature that highlights the importance of evaluating the lateral compartment separately, as distinct from the full patellofemoral joint. Evaluating the full patellofemoral joint may statistically mask actual biological effects in the lateral compartment. This study suggests that poor alignment and abnormal morphology may have a more important role in lateral patellofemoral OA than medial.    118  Medial trochlear inclination was the only morphology measure that was significantly associated with every pain and OA definition used in this study. However, sulcus angle had a stronger dose-response relationship with MRI-defined features of patellofemoral OA than the other three morphology measures. In addition, when the outcome included pain rather than structure alone, lateral trochlear inclination also demonstrated a stronger association. Thus, it is recommended that both sulcus angle as well as lateral trochlear inclination be used for future studies investigating morphology and patellofemoral OA, as they may be measuring different aspects of morphology that relate to structural features differently. Both alignment measures (patellar tilt, bisect offset), on the other hand, showed strong associations and should continue to be used in future studies.   Several epidemiology studies investigating alignment or morphology have included knee pain in their inclusion criteria [257 259], but they have not specified anterior knee pain in defining symptomatic patellofemoral OA. This may explain why a relationship between alignment or morphology and pain has not previously been demonstrated in patellofemoral OA [257 341]. While the Framingham Community Cohort did not assess the location of knee pain either, I used two definitions of pain to see if pain with stairs (a task that increases patellofemoral joint load) would show different results than simply generalized knee pain. Results showed that alignment and morphology are cross-sectionally associated with symptomatic patellofemoral OA using both pain definitions. Mean alignment and morphology values did not differ with pain definition (Table 4.6), however dose-response patterns (Figure 4.6) show possible stronger associations with lateral trochlear inclination, patellar tilt angle, and bisect offset, when pain with stairs is included in the outcome over generalized knee pain. This may indicate that these measures are   119  implicated in symptomatic patellofemoral OA. Future studies should define and localize patellofemoral pain as distinct from general knee pain.   4.4.1 Limitations Given that this is a cross-sectional study, it is not possible to determine if alignment or morphology are etiological risk factors for patellofemoral OA. Second, this study was an ancillary analysis of an existing cohort, and as a result, I was unable to accurately define and identify study participants with patellofemoral pain. Another limitation is that there is not yet a gold standard for defining patellofemoral OA using MRI, which leaves a possibility of group misclassification in our comparisons. We operationally defined patellofemoral OA as full thickness cartilage damage on account of the fact that partial thickness cartilage damage was so highly prevalent in this cohort (65%) so may not indicate clinically important structural change. BMLs are common even in young asymptomatic knees [159] and can fluctuate over relatively short time periods [342] and thus may not represent a stable indicator of disease state. While we did not report reference values using these two structural features, we did evaluate them in sensitivity analyses, and found similar dose-response patterns and thresholds.   As in the previous chapters of this dissertation, the MR images from this study were taken statically and in a traditional closed-bore scanner, and, as in Chapter 3, participants were non-weight bearing with quadriceps muscles relaxed and knees near full extension. Open-bore MR scanners would enable alignment to be assessed in full weightbearing, in different knee flexion angles, and with active and functional muscle recruitment patterns [83 283]. Given that patellofemoral OA is not typically pain-provoking in an unloaded supine position, it is possible   120  that weightbearing MRI may capture mechanical features that more closely relate to OA prevalence.   This study determined reference values and dose-response patterns for alignment and morphology in knees with and without patellofemoral OA, in older adults from Framingham MA who were primarily Caucasian. Reference values may not, therefore, be generalizable to other populations such as different age groups or ethnicities.   I did not evaluate the association with tibiofemoral OA. Importantly, the patellofemoral joint is in close proximity, and shares its articular capsule with, the medial and lateral tibiofemoral joint. Thus, the joints can influence each other both mechanically and through intra-articular communication. Therefore, alignment and morphology might also affect the tibiofemoral joint. Given their mutual influence, accounting for the patellofemoral joint itself while evaluating the tibiofemoral joint would be statistically complicated, and I therefore chose to limit my focus to the patellofemoral joint.  4.4.2 Clinical implications Reference values indicate large within-group variability and substantial overlap between those with and without patellofemoral OA. Dose-response patterns for both alignment and morphology are curvilinear and odds of having patellofemoral OA are higher in a single direction (i.e. J-shaped or reverse J-shaped pattern) with the exception of patellar tilt, which has higher odds whether malaligned towards lateral or medial tilt (U-shaped pattern). Of the morphology measures, sulcus angle and lateral trochlear inclination are most strongly associated with MRI-  121  defined features of patellofemoral OA. Both alignment measures had consistent strong associations with patellofemoral OA features.   The definition of pain for describing symptomatic patellofemoral OA is an important consideration that may help better elucidate the clinical relevance of alignment and morphology in the onset and progression of patellofemoral OA. The association between alignment or morphology and both structural features and knee pain justifies further investigation into conservative treatments that address malalignment [146 295] to potentially mitigate pain or reduce / delay structural progression. Specifically, these studies should be targeted towards a subset of patellofemoral OA that may constitute a ‘malalignment phenotype’. Based on the results of the present study, these individuals could justifiably be identified either by using the reference values listed in Table 4.5 or by using the thresholds listed in Table 4.8 to define ‘malalignment’.    122  Chapter 5 The ‘MRI Open-bore Osteoarthritis Vancouver’ (MOOV) study part I: 3D knee alignment in a vertically oriented open-bore scanner differs in standing compared to supine.  5.1 Background The evaluation of knee alignment using MRI has commonly been performed using traditional closed-bore scanners, where participants are positioned supine with the knee non-weightbearing and in or near full extension (Chapter 2, Chapter 3, Chapter 4) [249 294]. However, individuals with patellofemoral OA tend to report pain during weightbearing tasks that increase load at the patellofemoral joint (e.g. stairs, squatting)[41], and they are usually pain free in supine. It may be more relevant to investigate knee alignment during tasks that load the patellofemoral joint, because knee alignment may be influenced by weightbearing movements, pain avoidance, or impairments (e.g. muscle weakness) [343].  More functional environments have been previously used in MRI to evaluate knee alignment. Equipment such as rigs have been used to mimic weightbearing [256], though these typically are limited to 30% of bodyweight [255]. In vertically oriented open-bore scanners, participants have been scanned in standing statically [295 344 345] as well as dynamically [79 283 346]. However, such scanners typically require the individual to lean against a backrest, which alters the relative movements of the joints (see Figure 5.1). Using both static and dynamic (kinematic [347]) MRI protocols, knee alignment differed between supine and standing positions in individuals with and without patellofemoral pain or instability [345 346 348]. This highlights the importance of   123  developing MRI technology capable of measuring knee alignment during upright functional tasks.         Figure 5.1 Illustration of side view of an open-bore MRI scanner and backrest used in previous research. Reprinted with permission.24  Most studies to date have measured two-dimensional (2D) alignment from a single MRI slice (Chapter 2, Chapter 3, Chapter 4) [275 348]. This methodology may oversimplify the complexity of the three-dimensional (3D) knee, and is more vulnerable to measurement error than 3D depending on the variation in scan plane during image acquisition, and slice chosen for analysis. Studies evaluating 3D alignment in patellofemoral disease have been limited to the supine position [349], with exception of one study of 10 healthy knees that were evaluated in 3D during stance [344]. In the latter study [344], participants squatted to 30° in a neutral position, as well as in a position with the femur maximally adducted and internally rotated. This study revealed correlations between both tibial external rotation or abduction and patellar lateral displacement.                                                  24 Reprinted from Journal of Orthopaedic Research, Volume 27, Draper CE et al., “Using real-time MRI to quantify altered joint kinematics in subjects with patellofemoral pain and to evaluate the effects of a patellar brace or sleeve on joint motion”, pp 571-7, 2009, with permission from John Wiley and Sons   124  In addition, 2D alignment was only moderately correlated to 3D alignment and this was not significant. However, standing was not compared to supine. These findings suggest that further evaluation of standing in a diseased population is warranted, because quality of stance (e.g. tibia or femur positioning) may affect patellar positioning, and the evaluation of alignment in 3D may capture more complete information than 2D alignment.   5.1.1 Study aims The purpose of part I of this study was to: i) develop methods for evaluating 3D knee alignment in individuals positioned upright and fully weightbearing in a vertical open-bore MRI scanner; ii) evaluate the repeatability of alignment results obtained using the open-bore scanner in individuals with patellofemoral OA and in healthy controls; and (iii) explore how 3D knee alignment differs in standing compared to supine in individuals with and without patellofemoral OA. My primary hypothesis for this chapter was that the patella would be more laterally displaced and tilted, and more proximally translated, when standing compared to supine in individuals with and without patellofemoral OA [345].   5.2 Methods 5.2.1 Study design For this cross-sectional study, I recruited a convenience sample of 30 individuals: 15 with patellofemoral OA and 15 asymptomatic controls. In order to address the aims of Part II of this study (see Section Chapter 6), this was a cross-sectional study with matched controls.    125  5.2.2 Ethics Prior to commencing study recruitment, I obtained full ethical approval from the UBC Clinical Research Ethics Board (certificate number H13 – 01993) as well as Vancouver Coastal Health Research Institute Operational Review (number V13-01993).  5.2.3 Recruitment Recruitment strategies were targeted towards local physiotherapy clinics, sports medicine clinics, Vancouver Coastal Health staff members, fitness and community centres, and other general community sites. I posted notifications on bulletin boards where available, advertised in local newsletters (both print and online) and on available study recruitment websites (UBC’s Centre for Hip Health and Mobility, UBC’s MRI Research Centre, and Vancouver Coastal Health Research Institute). In addition, I contacted participants from another UBC laboratory (M Hunt) who had volunteered for previous studies (or at least had been screened for eligibility) and had provided consent to be contacted. Finally I advertised through social media platforms, including Twitter and Facebook.    Interested individuals were screened by telephone and, if eligible (see eligibility criteria below, Section 5.2.4), they provided fully informed written consent. Subsequently, eligibility of participants with patellofemoral OA was confirmed by having radiographs reviewed by a musculoskeletal radiologist (B. Forster) (see Figure 5.2 for participant screening flow chart). I obtained copies of any radiographs and associated reports that had been taken within the year prior to screening. Those who did not have a recent skyline view radiograph were asked to obtain one through their family doctor if the doctor believed it was clinically indicated, in which case   126  the radiograph would also become part of the individual’s clinical record. If not, individuals were invited to attend a private radiology clinic (False Creek Healthcare Centre) to have the skyline view obtained as part of the study (or all three views if no radiographs from the past year were available). The effective radiation dose associated with any radiographs taken as part of this study was less than 0.02 mSv. For comparison, the annual effective natural background dose of radiation due to daily living is about 2 mSv [350].    Figure 5.2 Flow chart of participant screening procedures for those with patellofemoral OA. Controls followed a similar flow chart, though after agreeing to participate they were added to a pool of eligible controls to await a match, and only completed the study if they were successfully matched.    127  5.2.4 Participants Inclusion criteria for early patellofemoral OA were:   age 40 years or older   peri- or retro-patellar knee pain   pain aggravated by at least one activity that loads the patellofemoral joint (e.g. squatting, climbing stairs)   pain severity of at least three on a numeric pain rating scale (0-10, 10 is the worst pain imaginable) during aggravating activities   pain during aggravating activities present most days of the past month  radiographic patellofemoral OA severity of at least Kellgren and Lawrence (KL) grade 1[114] on skyline and/or lateral view radiograph (see Table 1.2).  While OA is typically diagnosed on the basis of KL grade ≥ 2, I chose to include grade 1 to also capture early patellofemoral OA. This approach is supported by a prospective longitudinal study by Hart and Spector [151] that found more than 60% of those with “doubtful” osteophytes (i.e. KL grade 1) at baseline progressed to “definite” osteophytes at a 10 year follow up. It is also supported by the literature suggesting a link between patellofemoral pain and patellofemoral OA indicating similar impairments in these two populations [82] (see Section 1.7.3 for details and related references).  Controls were included if they were age 40 years or older and had no knee or other lower-limb symptoms in the past year. They were matched on a 1:1 ratio with patellofemoral OA cases on four known confounders for OA: age (± 5 years) [142 190], sex [142 171], ethnicity [351] and   128  BMI (± 5 kg/m2) [89 142 171 190]. In addition, they were matched on current physical activity level using the International Physical Activity Questionnaire-Short (IPAQ-S) (same category - low, moderate, or high – based on the number of reported hours spent walking, and doing moderate and/or vigorous activities in the previous week) [352] so that tests of strength and function were more likely to be direct reflections of disease state rather than secondary deconditioning (see Appendix D).   Exclusion criteria for patellofemoral OA cases and controls were:   concomitant pain from hip, ankles, feet or lumbar spine   recent knee injections (past 3 months)  planned lower-limb surgery in the following 6 months   BMI of ≥35 (due to the limited scanner bore size)  knee or hip arthroplasty, osteotomy, reconstruction, meniscectomy  hip fracture  history of major traumatic knee injury requiring non-weightbearing for ≥24 hours (e.g. fracture, dislocation, complete ligament rupture)  physical inability to participant in testing  contraindications to MRI  inability to understand written and spoken English.       129  Potential patellofemoral OA participants were also excluded if they had:   contraindications to radiation and did not already have three views of radiographs (PA, lateral, skyline) taken within the past year  tibiofemoral joint OA of KL severity grade 3 or 4  worse radiographic OA severity at the tibiofemoral joint than the patellofemoral joint  generalized knee pain  5.2.5 Demographic data collection Demographic data collected included age, sex, height and weight. I measured height with a wall-mounted digital stadiometer (Rosscraft Inc, White Rock, British Columbia, CA) using a standardized protocol. I measured weight, with shoes removed, using a digital scale (Seca Model 242, Hanover, Maryland, US). I also inquired about leg dominance (self-reported kicking leg), and smoking status (current, previous [pack-year history], never). Pain duration was categorized as 3-6 months, 6-12 months, 1-2 years, or >2 years. Current crepitus was defined as a ‘yes’ response to the question “Do you have any grinding in your knee?” History of pain was defined as a ‘yes’ response to the question, “Have you had pain of the knee cap in the past, either alone or in combination with locking and/or grinding, especially during walking stairs, headwind cycling, squatting and/or sitting” [143]. History of swelling was defined as a ‘yes’ response to the question, “Have you ever had dramatic swelling around the knee cap?” [142]. Finally, I took clinical measurements in order to acquire a Beighton score of general hypermobility [353].    130  All participants completed the EuroQol 5 Dimension, 5 Level (EQ-5D-5L) health status measure [354 355]. In addition, participants with patellofemoral OA completed the following patient-reported outcome measures (see Appendix D):  Knee injury and Osteoarthritis Outcome Score (KOOS) [356], including a new sixth subscale, the Patellofemoral subscale (KOOS-PF) [357]  Anterior Knee Pain Scale (AKPS) [358]  Tampa Scale of Kinesiophobia [359]  Knee Self-Efficacy Scale (K-SES) [360]  5.2.6 Imaging In participants with early patellofemoral OA, the painful knee was scanned (or most painful knee if there was bilateral pain). For the controls, I selected the knee to scan based on leg dominance of each control’s matched case (i.e. dominant or non-dominant knee). Images were acquired using two different MRI scanners on separate days within a two week window for each participant. Prior to MR scans, all participants were screened by an MRI technologist to ensure participants had no contraindications for scanning.  5.2.6.1 Conventional scanner: high-resolution images In a closed-bore 3T MRI scanner (Philips Achieva Best, NL), high-resolution sagittal images with a T1-weighted turbo spin echo sequence were acquired using a dual SENSE Flex-M coil set-up with participants in the supine position (see Figure 5.3, Table 5.1) [361]. This sequence and set-up was selected to obtain detailed participant-specific anatomical information, with a large field of view for improved accuracy in the subsequent registration process. A second   131  image, using a proton density weighted 3D VISTA sequence using an 8-channel SENSE knee coil (Invivo, Gainesville, FL, USA), was acquired for assessing patellofemoral alignment in 2D.     Figure 5.3 a. Dual SENSE Flex-M coil set-up for high-resolution scanner, large field of view image of index knee. b. Sample image taken from 3T scanner with dual coil set-up. Reprinted with permission25                                                    25 Reprinted from Journal of Magnetic Resonance Imaging, Macri EM et al., “Patellofemoral and tibiofemoral alignment in a fully weight-bearing upright MR: Implementation and repeatability” Published Online First: doi:10.1002/jmri.25823, 2017, with permission from John Wiley and Sons   132    Table 5.1 MRI sequence parameters.           *T1w TSE: T1-weighted Turbo Spin Echo; 3D PDw VISTA: 3D proton density weighted volume isotropic turbo spin echo acquisition; 3D GE: 3D gradient echo ^Scan times are approximate on account of the size of the knee, with larger knees requiring slightly longer scan times.   5.2.6.2 Upright scanner: low resolution images In the vertically oriented open-bore 0.5T MRI scanner (ParaMed MROpen Genoa, Italy) sagittal images were acquired using a gradient echo sequence with a commercial knee surface coil (see Table 5.1). This sequence was selected to obtain adequate image quality and sufficient number of slices to appropriately cover the knee anatomy (similar to the quality and number of slices used previously [362]). It was also selected to minimize stand time (hence limiting potential motion artifact due to fatigue, pain, or discomfort). Participants were scanned in two conditions (supine and standing two-legged), and at four sequential knee flexion angles (0○, 15○, 30○, 45○) (Chapter 6). One additional scan was obtained in standing, one-legged at 30○ flexion, for a total of nine different positions (Chapter 6).  In supine, I developed an MR-safe rig to permit partial weight bearing of the imaged limb (see Figure 5.4). This has been done previously where upright imaging was not possible [210 241 254  3.0T Philips Achieva: T1w TSE* 3.0T Philips Achieva:  3D PDw VISTA 0.5T Paramed Open GE Repetition time (ms) 700  1300 415  Echo time (ms) 10  35.1 10  Field of view (mm) 320  150 300 Acquisition matrix size 512x460 376x310 160x128 Slice thickness (mm) 2.0  (0.35 isotropic) 4.0 Gap thickness (mm) 0.0  0.0 0.4 Total scan time (min) ~16^ ~6  ~1   133  256 362 363]. I chose to load the limb during supine imaging to control for weightbearing status, because I was primarily interested in how alignment differs between standing and supine, more so than how weightbearing affects alignment [255 364]. The limb was loaded to 15% of body weight by placing weight in a bucket attached to a calibrated pulley system. This weight was selected based on a previous study where healthy participants had difficulty maintaining position with the limb loaded to 30% body weight [255]. A foot map was mounted to the foot plate to position the foot at 10° external rotation. The participant was instructed to press against the foot plate until the plate was vertically oriented, but no more (the plate would butt against a stopper if pushed too hard). For knee flexion angles, hard density foam wedges were built with 15°, 30° and 45° degree angles with space removed at the peak to accommodate the knee coil (see Figure 5.4).            Figure 5.4 Upright scanner positioning in supine. a. participant in supine set-up with loading rig; b. design of hard density foam wedges for knee positioning.     134  In standing, participants stood two-legged on a standardized foot map. The knee coil was attached to a lumbo-pelvic support brace that had been modified with a series of webbing loops added along the full length of the brace. The coil was attached using four pieces of 20 mm flat webbing with adjustable buckles (see Figure 5.5). A 12-inch goniometer was used to estimate knee flexion angle. A non-magnetic plumb bob was used to cue participants to position their knees behind the anterior border of the great toe. Three support bars (shins, buttocks, hands) helped participants to remain still during image acquisition (participants were instructed not to bear weight through the bars) (see Figure 5.6).          Figure 5.5 Donning knee coil in preparation for standing scans.     135            Figure 5.6 Upright scanner positioning in standing. a. positioned for standing scans with pressure bars in place (foot map and goniometer not shown). b. Sample image taken from open bore scanner. Reprinted with permission26  In a subset of 20 participants (10 patellofemoral OA, 10 controls), I repeated image acquisition at the 30° angle in standing two-legged, three times, to evaluate repeatability. This subset did not adhere to the case-control matching, and was instead selected as the first 10 consecutively from each group with adequate time in the scanner to complete repeated scans. I selected this position for evaluating repeatability because this is approximately where the patella engages with the trochlear groove, and can be pain provoking and thus challenging for individuals with patellofemoral OA [77]. Between each scan, participants were given a recovery period of approximately two minutes to stand and stretch while remaining in the scanner.                                                  26 Reprinted from Journal of Magnetic Resonance Imaging, Macri EM et al., “Patellofemoral and tibiofemoral alignment in a fully weight-bearing upright MR: Implementation and repeatability” Published Online First: doi:10.1002/jmri.25823, 2017, with permission from John Wiley and Sons   136   5.2.6.3 Image processing  Using Analyze 10.0 (Analyze Direct, KS, USA), I manually segmented the bony outlines of the femur, tibia and patella on all slices of the high-resolution images (i.e. 3T conventional MRI) to create participant-specific 3D anatomical surface models (see Figure 5.7 for complete image processing methods). I then manually assigned a joint coordinate system [365] to the models using anatomic landmarks [362 366]. These procedures were previously validated [362 366].  Figure 5.7 Image processing methods for one participant. This diagram shows the low-resolution images used for repeatability only (i.e. to determine alignment in three trials of standing, 30° knee flexion). Reprinted with permission27   Low-resolution images (0.5T open-bore MRI) were segmented in a similar manner. Registration (shape-matching) of the high-resolution model to the low-resolution outlines was done using an iterative closest points algorithm [367], then 3D knee alignment was calculated using a custom program in MATLAB (R2011a) (MathWorks, MA, USA) [256 361 362 366]. Patellofemoral                                                  27 Modified from Magnetic Resonance in Medicine, Volume 69, d’Entremont AG et al., “Do dynamic-based MR knee kinematics methods produce the same results as static methods?” pp 1634-44, 2013, with permission from John Wiley and Sons.   137  alignment was determined in all six degrees of freedom (i.e. three rotations, three translations) for the patella relative to the femur (see Figure 5.8 for specific alignment labeling conventions). While patellofemoral alignment was of primary interest, tibiofemoral alignment was also calculated (tibia relative to the femur -Figure 5.9).  Figure 5.8 Patella alignment relative to femur: a. flexion (rotation, sagittal plane); b. medial spin (rotation, frontal plane); c. medial tilt (rotation, axial plane); d. proximal translation; e. lateral translation; f. anterior translation. Positive (+) values indicate alignment relative to reported movement (e.g. in flexion, positive alignment is greater flexion). Illustrations by Vicky Earle. Reprinted with permission.28  Figure 5.9 Tibia alignment relative to femur: a. flexion; b. adduction; c. internal rotation; d. proximal translation; e. lateral translation; f. anterior translation. Illustrations by Vicky Earle. Reprinted with permission29                                                  28 Reprinted from Journal of Magnetic Resonance Imaging, Macri EM et al., “Patellofemoral and tibiofemoral alignment in a fully weight-bearing upright MR: Implementation and repeatability” Published Online First: doi:10.1002/jmri.25823, 2017, with permission from John Wiley and Sons 29 Reprinted from Journal of Magnetic Resonance Imaging, Macri EM et al., “Patellofemoral and tibiofemoral alignment in a fully weight-bearing upright MR: Implementation and repeatability” Published Online First: doi:10.1002/jmri.25823, 2017, with permission from John Wiley and Sons   138   5.2.7 Statistical analyses All statistical analyses were completed using Stata 13 (StataCorp, TX, USA). Statistical significance was defined at p≤0.05, with no adjustments to p-values for multiple testing.   5.2.7.1 Repeatability I evaluated repeatability for standing two-legged at 30° knee flexion in the 0.5T open-bore scanner only; inter-rater reliability for similar methods using a conventional scanner have been published previously [362]. Therefore, I completed segmentation of the high-resolution images and assigning of the joint coordinate system one time only.  I performed repeated measures ANOVA and then calculated ICC(2,1) with absolute agreement [312 335]. This coefficient accounted for re-positioning between scans, gross participant movement (e.g. altered motor control strategies between scan trials), rater bias during segmentation of the images acquired in the 0.5T scanner, and software registration/shape matching. I calculated standard error of measure (SEM) based on the root mean square error of the ANOVA and from this determined the smallest detectable change at 95% confidence (SDC95).   Repeatability was assessed for each group separately to evaluate potential differences between those with and without symptomatic patellofemoral OA. If values did not differ by group, I estimated and reported repeatability with the full study sample to optimize precision of the   139  estimates. Acknowledging that the acceptability of reliability coefficients is subjective, I defined ICC values of ≥0.75 as good, and ≥0.90 as excellent [311].   5.2.7.2 Descriptive statistics Demographics of both groups were summarized in tables and graphs. Categorical variables were summarized by frequencies and percentages, while continuous variables were summarized using mean and standard deviation. Univariate analysis was explored visually with dot plots, histograms and box plots. Alignment outcomes were explored as a function of knee flexion angle using scatter plots with overlying lowess curves (locally weighted scatterplot smoothing) [368] to guide subsequent analysis.   5.2.7.3 Inferential statistics: unadjusted within-group differences Within-group differences between supine and standing in full knee extension (i.e. alignment in standing minus that in supine) was reported along with Cohen’s d effect sizes [369 370]. A small effect size was defined as Cohen’s d ≥ 0.2, medium at d ≥ 0.5, and large as d ≥ 0.8 [369].   5.2.7.4 Inferential statistics: mixed effects models To evaluate differences between cases and controls across all angles and positions in a single model, I constructed mixed effects models, which incorporate both fixed and random effects into a model [371]. I used this approach for several reasons. First, the data in the present study were correlated on account of the repeated measures (i.e. different positions and different knee angles for each participant) as well as the case-control matched design (i.e. individuals within each pair were matched and thus correlated). Mixed effects modeling accounts for the lack of independent   140  observations (i.e. correlation) within a data set, in contrast to ordinary least squares methods in which all observations are treated on the assumption that they are independent and identically distributed [371]. A second benefit of mixed effects modeling over simpler general linear models is that these models are better at managing unbalanced data [371]. A third benefit is that the modeling of random effects enables more generalized inferences about the population of interest [371]. Finally, I fit each model using maximum likelihood, thus I could use likelihood ratio tests to help determine the best model (i.e. which covariates to include) for each outcome [371].  For the fixed effects portion, I modeled each alignment measure as a function of group assignment (patellofemoral OA or control), position (supine or standing), and knee flexion angle. I considered whether knee flexion angle was better modeled as an indicator variable (0°, 15°, 30°, 45°) [284] based on the goniometer measures, or as a continuous variable based on the tibiofemoral flexion angle calculated after image processing. I theorized that the continuous variable might more accurately estimate knee flexion angle since the use of goniometry (in particular with the participant inside the MR scanner) might have more error than the image processing method. If a continuous variable was better, I also considered a square term in the model (centered on the grand mean) to measure curvature in the plots. All possible interaction terms were also considered for inclusion into the model: Angle*Position, Angle*OA, Position*OA, and Angle*Position*OA. For each possible covariate, I used a likelihood ratio test to determine if the new variable significantly improved the overall model. I also observed how the model changed with the new variable under consideration: (i) how much did the estimated beta coefficient (β) change for the primary predictor variable (group assignment); (ii) did the   141  standard error of the primary predictor variable decrease; and (iii) did Akaike’s information criteria decrease [372]?   For the random effects portion of the model, I fit a random intercept at the matched pair level if angle was entered as an indicator variable in the fixed portion; otherwise I fit a random intercept and random slope for knee flexion angle (as a continuous variable) at the pair level. Finally, I considered the covariance structure of the repeated measures. An independent covariance (aka diagonal) structure allows for a distinct variance for each random effect but assumes that all covariances are zero. This is the default for repeated measures models, which I kept unless a likelihood ratio test revealed that an unstructured covariance (which makes no assumptions about covariance) significantly improved the model.   Following selection of the final model, the distribution of the residuals against the fitted values were evaluated to confirm the model did not violate the assumption of normally distributed residuals. To do this, I evaluated the data visually (histogram of residuals, scatter plot of residuals vs. fitted values) and performed a Shapiro Wilks statistical test. I also checked for influential observations by calculating dfbeta and Cook’s D values as well as by visually inspecting scatter plots (each alignment measure vs. knee flexion angle) and box plots of each alignment measure at each knee flexion angle.   After confirming the final model for each alignment outcome, I reported all coefficients and also presented the models graphically.     142  5.2.7.5 Two-dimensional alignment: malalignment subgrouping Because 3D upright alignment has not been done previously in a sample with patellofemoral OA, I also measured unweighted supine 2D alignment using the 3D PDw VISTA images taken in the conventional scanner. Three measures were calculated (bisect offset, patellar tilt angle, Insall-Salvati ratio). Participants whose 2D alignment measures fell beyond normative reference values were identified. Cut-points were defined as ≥61.6% for bisect offset (Table 4.5); ≥18.1° and 16.0° for patellar tilt angle for men and women, respectively (and ≤0.1° for both sexes) (Table 4.5); and ≥1.2 for Insall-Salvati ratio [250]. To see if those with malalignment differed by position (standing vs. supine) in a different way than those who are not malaligned, a ‘malalignment’ subset was created of participants with at least two of the three measures lying beyond the reference values. I then summarized the change in position (standing value minus supine value) by those with and without malalignment.   5.3 Results 5.3.1 Participants Participant recruitment took place from August 2014 until August 2016. Two hundred and twenty two (222) individuals were screened to complete recruitment of 30 people (see Figure 5.10 for screening flow chart). The full sample had a mean age of 56 ± 7 years, mean BMI 23.5 ± 3.7 kg/m2, and comprised 24 women (80%) and six men (see Table 5.2). Twelve pairs were of European ethnicity, three pairs were of Chinese ethnicity, and one pair was of Indian ethnicity. Ten of the OA participants reported OA in other joints (contralateral knee n=8, hands n=3) while no controls reported OA in any joint. All participants were non-smokers, though two controls reported a smoking history (one with a 15 pack-year history, one with 30). Among those with   143  patellofemoral OA, all 15 reported a pain duration lasting greater than 2 years (see Table 5.2 for self-reported pain and function). No participants had general hypermobility as indicated by the Beighton score.              Figure 5.10 Flow chart for participant screening.      144  Table 5.2 Participant characteristics (n=30). All values are mean (SD) unless otherwise noted.  Patellofemoral OA (n=15) Controls (n=15) Age, mean(SD) 56 (8) 56 (7) BMI mean(SD) 23.5 (4.1) 23.5 (3.4) Women n(%) 12 (80%) 12 (80%) History of crepitus n(%) 9 (60%) 0 (0%) History of swelling n(%)  11 (73%) 0 (0%) EQ-5D-5L n(%)   Mobility 12 (80%) 15 (100%) Self-care 15 (100%) 15 (100%) Usual activities 11 (73%) 14 (93%) Pain/discomfort 1 (7%) 13 (87%) Anxiety/depression 12 (80%) 15 (100%) VAS 85.0 (8.7) [70,100] 92.7 (6.4) [80,100] KOOS   Symptoms 68.8 (16.7) [32.1, 89.3] - Pain 71.5 (11.6) [50.0, 86.1] - ADLs 82.9 (12.6) [54.2, 100.0] - Sport/Rec 46.0 (28.1) [10.0, 95.0] - Quality of Life 48.3 (17.9) [18.8, 68.8] - Patellofemoral 56.8 (17.9) [18.2, 77.3] - Kujala AKPS 70.7 (14.1) [49.0, 98.0] - K-SES TOTAL: 6.6 (1.8) [3.2, 9.6] - Daily activities 7.6 (2.0) [3.1, 9.9] - Sport/Leisure 7.0 (2.2) [1.0, 10.0] - Physical activity 5.8 (2.3) [2.5, 10.0] - Knee function 5.1 (1.8) [3.2, 9.6] - Tampa score ‘high’ n(%) 7 (47%) - BMI: body mass index; EQ-5D-5L: EuroQol Health Status Measure (reported number of participants who reported the best possible quality of life score in each domain); KOOS: Knee injury and Osteoarthritis Outcome Score (ranges from zero, maximum problems, to 100, no problems); ADLs: activities of daily living; Rec: recreation; AKPS: Anterior Knee Pain Scale (ranges from zero, maximum problems, to 100, no problems); K-SES: Knee Self-Efficacy Scale (ranges from zero, maximum problems, to 10, no problems); VAS visual analogue scale (ranges from zero, no quality of life, to 100, best quality of life)   5.3.2 Repeatability Repeatability for the combined sample (n=20) was good to excellent, with ICC values of ≥ 0.94 with the exception of patellar spin, which had ICC 0.79 (see Table 5.3). All SEM values for the combined sample were < 2° rotation and < 1 mm translation (see Table 5.3 for individual SEMs and their associated SDC95 values). Repeatability was similar when evaluated separately by group, with the exception of patellar spin, with a lower ICC in the control group of 0.57 (SEM 2°) compared to 0.91 in the patellofemoral OA group (SEM 1°).    145   Table 5.3 Repeatability reported separately by group, and combined (standing, 30° knee flexion angle).  Patellofemoral OA (n=10)  Controls (n=10) Combined (n=20) Patellofemoral joint ICC (2,1)  SEM SDC95 ICC (2,1)  SEM SDC95 ICC (2,1)  SEM SDC95 Flexion  0.95 (0.87, 0.99) 1.1° 2.9 0.94 (0.83, 0.98) 1.3° 3.6 0.94 (0.89, 0.98) 1.2° 3.2 Medial spin 0.91 (0.76, 0.98) 1.3° 3.7 0.57 (0.19, 0.85) 2.3° 6.3 0.79 (0.62, 0.90) 1.9° 5.3 Medial tilt 0.94 (0.84, 0.98) 1.5° 4.0 0.97 (0.91, 0.99) 1.3° 3.7 0.97 (0.93, 0.99) 1.6° 4.4 Proximal translation 0.99 (0.97, 1.00) 0.6mm 1.8 0.98 (0.96, 1.00) 1.1 mm 2.9 0.99 (0.98, 1.00) 0.9 mm 2.4 Lateral translation 0.99 (0.98, 1.00) 0.4mm 1.2 0.95 (0.88, 0.99) 0.5 mm 1.4 0.99 (0.97, 0.99) 0.5 mm 1.3 Anterior translation 0.97 (0.91, 0.99) 0.5mm 1.5 0.99 (0.98, 1.00) 0.4 mm 1.2 0.98 (0.97, 1.00) 0.5 mm 1.4 Tibiofemoral joint          Flexion 0.99 (0.97, 1.00) 0.9° 2.5 0.98 (0.96, 1.00) 1.4° 3.9 0.99 (0.97, 0.99) 1.17° 3.2 Adduction (varus) 0.99 (0.96, 1.00) 0.5° 1.3 0.99 (0.96, 1.00) 0.6° 1.6 0.99 (0.97, 0.99) 0.53° 1.5 Internal rotation 0.95 (0.86, 0.99) 1.2° 3.2 0.95 (0.86, 0.99) 1.1° 2.9 0.95 (0.90, 0.98)     1.23° 3.4 Proximal translation 0.98 (0.96, 1.00) 0.3mm 0.9 0.94 (0.84, 0.98) 0.2 mm 0.7 0.98 (0.95, 0.99) 0.29 mm 0.8 Lateral translation 0.98 (0.93, 0.99) 0.4mm 1.0 0.99 (0.98, 1.00) 0.3 mm 0.8 0.99 (0.97, 0.99) 0.33 mm 0.9  Anterior translation 0.99 (0.96, 1.00) 0.5mm 1.4 0.96 (0.90, 0.99) 0.5 mm 1.3 0.98 (0.96, 0.99) 0.47 mm 1.3  ICC: intraclass correlation coefficient; SEM: standard error of measurement; SDC95: smallest detectable change at 95% confidence  Table 5.4 Mean values of patellofemoral alignment at full extension, supine and standing, in both groups; difference of standing minus supine, and Cohen’s d for the within-group effect size (standing minus supine).  Patellofemoral OA (n=15)   Control (n=15)  Patellofemoral joint Supine (Su) Standing (St) St - Su Cohen’s d Supine Standing St - Su Cohen’s d Flexion  -8.9 (4.1) -6.6 (3.6) 2.3 (3.4) 0.7* -9.3 (3.9) -4.9 (5.4) 4.4 (4.8) 0.9 Medial spin 0.5 (4.9) 1.6 (5.3) 1.1 (2.8) 0.4 -1.0 (3.5) 0.9 (4.2) 1.9 (3.2) 0.6 Medial tilt 11.4 (6.8) 9.1 (6.7) -2.3 (3.4) -0.7 16.2 (5.5) 15.0 (6.3) -1.2 (3.2) -0.4 Proximal translation 30.1 (6.8) 31.5 (7.2) 1.4 (4.4) 0.3 25.2 (6.1) 25.7 (8.0) 0.5 (4.0) 0.1 Lateral translation 3.0 (4.1) 4.6 (4.4) 1.6 (3.2) 0.5 0.0 (2.9) 2.2 (2.8) 2.1 (3.1) 0.7 Anterior translation 25.5 (3.0) 26.6 (3.9) 1.1 (1.7) 0.7 25.3 (2.7) 26.7 (2.4) 1.4 (1.9) 0.7 St = standing; Su = supine *bold indicates Cohen’s d ≥0.5 (i.e. moderate or large effect size)    146  5.3.3 Patellofemoral alignment  5.3.3.1 Unadjusted within-group differences Table 5.4 reports mean ± SD patellofemoral alignment for each group, in supine and in standing, with knees at full extension. Cohen’s d is reported as a standardized effect size of within-group differences by position (standing compared to supine). The only large effect size when changing from supine to standing (without introducing knee flexion) was in patellar flexion for the control group, with the patella approximately 4 ±5° more flexed in standing compared to supine. Effect sizes were moderate in the controls with approximately 2 ± 3 mm more patellar lateral translation and 1 ± 2 mm more anterior translation. Effect sizes were moderate in the OA group with approximately 2 ± 3° more patellar flexion, 2 ± 3° more lateral tilt, 2 ± 3 mm more lateral translation and 1 ± 2 mm more anterior translation in standing compared to supine. While spin also had a moderate effect size for the control group, the reliability of spin was not adequate (see Section 5.2.7.1) and therefore results for spin for have been presented in tables but not reported in text for the remainder of Chapter 5 and Chapter 6.  Systematic error was introduced during data collection with knees bent in supine. Specifically, part way through data collection, goniometry was added to the methods for confirming supine knee flexion angle, when it was determined that the hard density foam wedges were underestimating this angle for many participants. Within-group comparisons from supine to standing with flexed knees were therefore not evaluated using the unadjusted data.     147  5.3.3.2 Mixed effects models Knee flexion was modelled as a continuous variable based on the 3D alignment results calculated during image processing. This enabled a less biased estimate of alignment in supine flexion (to account for the systematic error described above). It also allowed knee flexion to be treated as a random effect in the model.    The following graphs (Figure 5.11 through to Figure 5.16) show the linear prediction from the fitted model of all patellofemoral alignment outcomes, combined with the upper and lower bounds of the standard error of the linear prediction (i.e. variability about the mean). The left graph shows within-OA group differences between standing and supine; and right shows within-control group differences between standing and supine. Along with each graph is a table (Table 5.5 through to Table 5.10) providing the coefficients for all variables included in the model, as well as their respective p-values.   During data collection, it was revealed that two participants had conditions with potential to influence study outcomes. One of the patellofemoral OA participants reported having Marie Charcot Tooth disease. Since this condition can lead to both motor and/or sensory neuropathy, it had the potential to influence strength and function, despite the participant reporting that she was ‘asymptomatic’. In addition, one of the controls reported spinal scoliosis. This condition had the potential to influence functional alignment in the lower limbs to compensate for more proximal misalignment, particularly during one-legged tasks but likely less so when standing bilaterally.     148  To address these concerns, I located these two participants on all exploratory scatter plots, and in any measures where their values laid on the periphery of overall sample values (even if they were not frank outliers) I conducted sensitivity analyses with those participants removed. If this occurred, I reported coefficients both with and without those participants included in the analysis. If not, I kept them in the sample and reported results for the full sample.  Only one patellar alignment measure (patellar flexion) was influenced by these two participants with medical conditions. Therefore, coefficients are reported in Table 5.5 as a full sample and with the two participants removed.   Table 5.5 Patellar flexion results, mixed effects: model coefficients (β) with 95% confidence intervals (CI) and p-values. ‘Position’ represents standing compared to supine TFJ = tibiofemoral joint, PFOA = patellofemoral OA *one participant with patellofemoral OA who had Marie Charcot Tooth, and one control who had scoliosis were removed because their alignment values were influential         Figure 5.11 Patellar flexion, standing (solid) and supine (dash), linear prediction from the fitted model (black) plus standard error (grey), by group:  a. patellofemoral OA; b. controls. PFOA = patellofemoral OA; CON = control; UCI = upper bound of standard error; LCI = lower bound of standard error.    Full sample Two participants removed*  β 95% CI p-value β 95% CI p-value Position 4.07 2.44, 5.71 <0.001 4.40 2.79, 6.02 <0.001 Position*TFJ flexion -0.15 -0.21, -0.08 <0.001 -0.16 -0.23, -0.10 <0.001 Tibiofemoral flexion 0.68 0.62, 0.75 <0.001 0.70  0.63, 0.77 <0.001 Tibiofemoral flexion2 0.006 0.004, 0.008 <0.001 0.006 0.004, 0.008 <0.001 Patellofemoral OA -1.30  -2.95, 0.35 0.12 -1.93  -3.57, -0.28 0.022 PFOA*TFJ flexion -0.08  -0.14, -0.02 0.011 -0.05  -0.12, 0.01 0.08   149  Table 5.6 Patellar medial spin results, mixed effects: model coefficients (β) with 95% confidence intervals (CI) and p-values. ‘Position’ represents standing compared to supine  β 95% CI p-value Position 1.25 0.31, 2.19 0.009 Tibiofemoral flexion 0.06 0.02, 0.09 0.001 Tibiofemoral flexion2 -0.002 -0.004, -0.000 0.024 Patellofemoral OA -0.73  -1.96, 0.49 0.24 PFOA* TFJ flexion2 0.004  0.001, 0.007 0.008         Figure 5.12 Patellar medial spin, standing (solid) and supine (dash), linear prediction from the fitted model (black) plus standard error (grey), by group:  a. patellofemoral OA; b. controls.. PFOA = patellofemoral OA; CON = control; UCI = upper bound of standard error; LCI = lower bound of standard error.  Table 5.7 Patellar medial tilt results, mixed effects: model coefficients (β) with 95% confidence intervals (CI) and p-values. ‘Position’ represents standing compared to supine  β 95% CI p-value Position -0.88 -1.96, 0.21 0.113 Tibiofemoral flexion 0.14 0.09, 0.18 <0.001 Tibiofemoral flexion2 -0.002 -0.004, 0.000 0.051 Patellofemoral OA -5.81  -6.86, -4.76 <0.001             Figure 5.13 Patellar medial tilt, standing (solid) and supine (dash), linear prediction from the fitted model (black) plus standard error (grey), by group:  a. patellofemoral OA; b. controls. PFOA = patellofemoral OA; CON = control; UCI = upper bound of standard error; LCI = lower bound of standard error.    150  Table 5.8 Patellar proximal translation results, mixed effects: model coefficients (β) with 95% confidence intervals (CI) and p-values. ‘Position’ represents standing compared to supine  β 95% CI p-value Position 1.40 0.06, 2.73 0.041 Tibiofemoral flexion -0.58 -0.63, -0.52 <0.001 Patellofemoral OA 3.93  1.74, 6.13 <0.001 PFOA*TFJ flexion 0.13 0.05, 0.21 0.002             Figure 5.14 Patellar proximal translation, standing (solid) and supine (dash), linear prediction from the fitted model (black) plus standard error (grey), by group:  a. patellofemoral OA; b. controls.  PFOA = patellofemoral OA; CON = control; UCI = upper bound of standard error; LCI = lower bound of standard error.  Table 5.9 Patellar lateral translation results, mixed effects: model coefficients (β) with 95% confidence intervals (CI) and p-values. ‘Position’ represents standing compared to supine  β 95% CI p-value Position 1.90 0.66, 3.14 0.003 Position*TFJ flexion -0.06 -0.11, -0.01 0.014 Tibiofemoral flexion -0.02 -0.06, 0.03 0.467 Tibiofemoral flexion2 0.002 0.000, 0.003 0.025 Patellofemoral OA 2.50  1.77, 3.24 <0.001         Figure 5.15 Patellar lateral translation, standing (solid) and supine (dash), linear prediction from the fitted model (black) plus standard error (grey), by group:  a. patellofemoral OA; b. controls. PFOA = patellofemoral OA; CON = control; UCI = upper bound of standard error; LCI = lower bound of standard error.    151  Table 5.10 Patellar anterior translation results, mixed effects: model coefficients (β) with 95% confidence intervals (CI) and p-values. ‘Position’ represents standing compared to supine  β 95% CI p-value Position 1.34 0.40, 2.28 0.005 Position*TFJ flexion -0.04  -0.08, -0.01 0.021 Tibiofemoral flexion -0.12 -0.16, -0.08 <0.001 Tibiofemoral flexion2 -0.002 -0.003, -0.001 <0.001 Patellofemoral OA -0.55  -1.51, 0.40 0.259 PFOA*TFJ flexion 0.06 0.03, 0.10 <0.001        Figure 5.16 Patellar anterior translation, standing (solid) and supine (dash), linear prediction from the fitted model (black) plus standard error (grey), by group:  a. patellofemoral OA; b. controls. PFOA = patellofemoral OA; CON = control; UCI = upper bound of standard error; LCI = lower bound of standard error.   Mixed effects models revealed that standing differed significantly from supine in all patellar alignment measures except tilt. The largest differences were with knees in full extension, where the patellae were approximately 4° more flexed, 2 mm more laterally translated, and 1 mm more anteriorly translated in standing compared to supine. The differences between standing and sitting did not change across tibiofemoral flexion angles for the remaining measures.  5.3.3.3 Two-dimensional alignment: malalignment subgrouping Using established reference values for 2D alignment (see Table 4.5) [250], six participants with patellofemoral OA and one control were defined as ‘malaligned’. Table 5.11 reports the   152  difference in 3D patellar alignment between positions (standing minus supine, full extension) as a function of malalignment. Qualitatively, those classified as malaligned on 2D measures showed greater within-group differences between supine and standing in patellar tilt, proximal and lateral translation, but less difference into flexion and anterior translation, compared to those without malalignment. The amount of within-group difference was not statistically different between those with and without malalignment.  Most notably, the within-group effect sizes in the malaligned group were in the same direction but larger than the within-group effect sizes in the patellofemoral OA group (Table 5.4) for patellar tilt, proximal translation, and lateral translation.   Table 5.11 Within-group difference in patellar alignment by position (standing value minus supine value), in full extension, in those malaligned (n=7) or not (n=23).  Malaligned  (n=7) Cohen’s d Non-malaligned  (n=23) Cohen’s d p-value Flexion 1.4 (4.2) 0.3 4.0 (4.2) 1.0 0.18 Medial tilt -3.5 (2.3) -1.5 -1.2 (3.4) -0.4 0.06 Proximal translation 3.5 (5.5) 0.6 0.2 (3.5) 0.1 0.18 Lateral translation 2.5 (4.1) 0.6 1.7 (2.8) 0.6 0.64 Anterior translation 0.7 (1.4) 0.5 1.4 (1.9) 0.8 0.29   5.3.4 Tibiofemoral alignment The only tibiofemoral alignment measure that differed between supine and standing was anterior translation, which for both OA and control groups was more anterior in standing compared to supine, and only in deeper flexion angles (see appendices E.1, E.6).     153  5.4 Discussion This study demonstrated good to excellent repeatability, with lower repeatability noted for spin in the asymptomatic group. Patellofemoral alignment differed between supine and standing positions within both groups, with at least moderate effect sizes towards increased patellar flexion, lateral tilt (OA group only), lateral translation and anterior translation (i.e. all towards greater malalignment) when standing (Table 5.4). Differences between supine and standing, within both groups, were greatest in full extension (compared to a flexed knee) for patellar flexion, lateral translation, and anterior translation (Figure 5.11, Figure 5.15, Figure 5.16).   Spin was previously reported as having low repeatability using the same methods but in a supine position [362]. The results of this study likely do not reflect high intra-subject spin variation in an upright environment, but rather are due to the methods used. For example, acquiring images in a coronal plane could have improved repeatability for spin, but this would have required more slices in the anteroposterior direction due to the thickness of the patella, resulting in increased scan duration. Given the importance of minimizing stand time for participants, this was not done. However, repeatability was not worse in the patellofemoral OA group compared to the asymptomatic group as may have been expected (i.e. had pain or muscle impairment increased the difficulty of adopting and maintaining each position).   Reporting SEM with the unitless ICC(2,1) enhances interpretation of repeatability in this study, providing 95% confidence that true alignment is within one SEM of the point estimate (Table 5.3). Lau et al. [373] evaluated 3D patellar alignment in participants with patellofemoral pain who were positioned supine and partially weightbearing (25% body weight). They reported SEM   154  values which are comparable to our results (flexion 2.40°, spin 0.94°, tilt 0.98°, proximal translation 1.43 mm, lateral translation 1.10 mm, anterior translation 1.67 mm) [373]. This suggests that acquiring images in upright positions does not affect repeatability. The additional statistic in Table 5.3, SDC95, estimates the smallest amount of change that could be measured on two occasions that would provide 95% confidence that true change has occurred. This statistic offers a starting point for estimating sample size calculations for future clinical trials.   The within-group comparisons in this chapter compare to similar studies. In a study investigating patellar instability, lateral tilt, lateral displacement, and proximal translation were qualitatively larger in upright compared to supine in full extension, with differences reduced in flexion and eventually reversed by 45° flexion [345]. Differences between positions were larger in the patellar instability group compared to a control group. In another study of samples with and without patellofemoral pain [374], at 45° knee flexion, patellae were less laterally tilted and laterally displaced in weightbearing compared to non-weightbearing, with larger differences within the patellofemoral pain group than controls. These are similar to findings in the present study for lateral displacement; however, there was no difference in lateral tilt or proximal translation by position at 45° flexion in the current study (Figure 5.13, Figure 5.14, Figure 5.15). This could be explained by methodological differences (e.g. 2D vs. 3D alignment). Alternatively it could indicate that in patellofemoral OA, lateral displacement differs more by position than lateral and proximal tilt, while in other patellar conditions, all three measures may differ by position.     155  In contrast, a study of real-time alignment in upright compared to supine showed a sample with patellofemoral pain was more laterally displaced in supine compared to upright, except for those who were also ‘maltrackers’ and were more laterally displaced in upright [283]. That study found no difference in tilt by position throughout range. In the present study, both groups were more laterally displaced in standing compared to supine (particularly in full extension), and the patellofemoral OA group was more laterally tilted in standing compared to supine in full extension only. Moreover, when divided into those with and without ‘malalignment’, those with malalignment had non-significantly greater differences than those without malalignment. This may suggest that individuals who are malaligned when supine are even more malaligned when standing. Thus, it is possible that standing images better quantify malalignment.  5.4.1 Limitations Alignment was evaluated on sequential static images rather than using dynamic (kinematic [347]) knee methods. Alignment differs when MRI scans are acquired statically compared to dynamically in a supine position [256 375], but it is unknown to what extent this differs in an upright position. Dynamic MR scans of upright squatting are typically limited to a single 2D slice [283], whereas in 3D, the entire knee can be evaluated, anatomical coordinate systems enable more accurate estimates of relative position of the bones, and alignment can be simultaneously calculated in all six degrees of freedom. Therefore, I selected a static method that enabled comprehensive 3D evaluation in a fully upright weightbearing position.  Only one participant reported pain during image acquisition in the standing position with knees flexed 30° or more. Lack of pain provocation in this study may have led to better repeatability than   156  if pain had been experienced by more participants during scanning. In this study, I used standardized positioning to ensure that knees were behind the great toe, which can reduce torque about the knee and reduces patellofemoral shear forces [376 377], which may reduce pain during squatting. If participants had self-selected their squat position in this study, this may have been pain provoking and may have altered alignment or repeatability. Future studies could compare self-selected positions with standardized positions to determine if alignment or symptoms differ. Importantly, the present study demonstrates that, even with standardized positioning that is not pain-provoking, alignment differs between those with and without patellofemoral OA.  Although the open-bore MRI scanner was of low field strength, a 3T scanner was employed to create the surface model for registration, and results demonstrated adequate intra-subject repeatability. Another advantage of using two scanners is that future studies can simultaneously assess cartilage morphology or biochemical composition, features that are not yet possible to evaluate in a 0.5T open-bore scanner.   Systematic error was introduced into the methods by using standardized foam wedges for supine positioning in the 0.5T open-bore scanner without checking alignment with a goniometer. I addressed this by analyzing within-group differences with unadjusted data in full extension only, and then also by modeling positions based on tibiofemoral flexion angle as a continuous variable using the processed alignment data.   This study had a relatively small sample size, which may have influenced the study results. When using mixed methods modeling, it has been suggested that a sample size of ten per   157  independent variable is adequate [378]. Acknowledging the sample size, results nonetheless showed both significant and clinically relevant effect sizes. Moreover, patterns of alignment were biologically plausible and similar to previously published studies of patellofemoral disease.    A final limitation is that this sample was underpowered for identifying and evaluating a ‘malalignment phenotype’, which was not a primary aim of this study. Seven participants were defined as ‘malaligned’ in this study. It is therefore possible that this patellofemoral OA sample was made up of more than one OA phenotype. For example, eight OA participants reported contralateral knee OA and three reported hand OA (only three of these 10 were also ‘malaligned’), possibly suggesting a polyarticular OA phenotype [96 205 206]. I did not assess for other knee OA phenotypes (e.g. chondrocalcinosis), though I did exclude those with a history of knee trauma (post-traumatic phenotype), those with BMI over 35 (obesity phenotype: only one participant, with OA, had BMI > 30), and those with inflammatory arthritis. In addition, no participants in this study were hypermobile and thus a hypermobility phenotype (e.g. such as Ehlers Danlos Syndrome), was also excluded from this study. Despite this, as mentioned, this sample may not represent a homogeneous phenotype. Alignment effect sizes of a pure malalignment OA phenotype may be underestimated in the present study.  5.4.2 Clinical implications Based on this study, the recommended position and knee angle selected for assessing alignment depends on the specific research question. Whether standing or supine, the greatest absolute values of most patellar alignment measures are captured in full extension. In full extension, greatest absolute values for all patellofemoral alignment measures are captured in standing   158  compared to supine. For those with patellofemoral OA (and more so, those with malalignment), within-group differences between standing and supine will be even greater than the difference within healthy knees in key measures (lateral tilt, proximal translation, lateral translation). This supports the theory that alignment in standing is more clinically important than supine, and that conventional supine images may be less sensitive for detecting malalignment.   To conclude, evaluating alignment in standing may be both more sensitive as well as more clinically important compared to evaluating alignment in supine as is typically done in conventional scanners. This conclusion supports part II of this study, in which between-group differences in alignment in standing and supine are reported.     159  Chapter 6 The MOOV study part II: 3D knee alignment differs between patellofemoral OA and matched controls.  6.1 Background A systematic review of the literature, presented in Chapter 2, revealed evidence of an association between knee alignment and patellofemoral OA features in people with knee OA (patellofemoral tibiofemoral, or combined) or other knee conditions [294]. Many of the reviewed studies evaluated alignment using MR images [146 258 259 261 276 280 281]. Several key knowledge gaps emerged from this study.  Individuals with patellofemoral OA commonly report pain during weightbearing tasks that increase the load at the patellofemoral joint (e.g. stairs, squatting) [41] – they are usually pain free in the supine position. However, no MRI studies were identified in the systematic review that investigated knee alignment in patellofemoral OA in an upright position [294]. In Chapter 5, it was demonstrated that knee alignment is altered as a function of body position (i.e. standing vs. supine). The greatest absolute values of patellar alignment (i.e. furthest towards the direction of malalignment) were in standing with knees in full extension. For individuals with patellofemoral OA (and even more so, with malalignment), within-group differences between standing and supine were greater than the within-group differences in control knees. This supports the theory that alignment in standing is more clinically important than supine, and that conventional supine images may be less sensitive for detecting malalignment.     160  Most studies included in the review did not specifically recruit participants diagnosed with patellofemoral OA – it was the target population in just two studies [146 277]. Rather, most studies evaluated participants with knee OA (most often recruited on the basis of tibiofemoral OA), and the remaining studies included individuals with various knee conditions. Moreover, only one study specifically recruited participants with anterior knee pain [146], and the rest either included people with general knee pain, or else pain was not included in the eligibility criteria at all. This highlights the need for studies that evaluate a well-defined patellofemoral OA sample, including more stringent criteria for defining anterior knee pain.  Another key limitation identified was that most studies did not include an asymptomatic control sample. Only one of the studies reviewed recruited a healthy comparison group [146]. Reference groups in the other studies reviewed were most commonly those without the structural outcome of interest, but they still had either tibiofemoral OA or another medical condition affecting the knee. While other studies have reported alignment in asymptomatic knees [286-289], direct comparison and generalizability is limited because alignment differs among studies due to methodological differences, including which alignment technique was used. This highlights the need for studies that include healthy controls for direct between-group comparison.  In upright MRI studies, patellofemoral pain was associated with greater lateral patellar displacement, lateral tilt and patellar height compared to controls [78 345 379], and was also associated with greater femoral internal rotation compared to controls [78 379]. In weight bearing, femur rotation may be a key driving mechanism of altered patellar alignment [79]. Femoral internal rotation may be related to hip muscle strength, which is lower in individuals   161  with patellofemoral OA [208 209]. In addition, no studies have compared the difference in alignment between two-legged stance and one-legged stance. One-legged stance increases the demand on the hip musculature, both in terms of strength (increased load) and motor control (narrower base of support). One leg stance also increases patellofemoral joint load as body weight is transferred onto a single limb. In a painful knee, this could invoke strategies to reduce patellofemoral joint reaction force that might result in further altering patellar alignment [148]. Knees with patellofemoral OA may therefore exhibit greater malalignment during one-legged stance.  6.1.1 Study aims The purpose of part II of this study was to: i) evaluate differences in standing 3D knee alignment between patellofemoral OA cases and matched controls; and (ii) evaluate how alignment differs between a two-legged squat and a one-legged squat in each group. A secondary exploratory aim was to evaluate these between-group differences in individuals with and without malalignment.   The primary hypothesis for this part of the study was that patellae would be more laterally displaced and tilted, and more proximally translated, in patellofemoral OA compared to matched controls in all positions [78 79 283 345 374 379]. In addition, between-group differences would be greater when transitioning from two-legged stance to one-legged stance, with OA knees demonstrating increased external tibial rotation30 when standing one-legged, compared to controls.                                                   30 External tibial rotation relative to the femur could be considered the distal consequence of  internal femoral rotation, when an individual is upright and the foot is in contact with the ground in a standardized position   162   6.2 Methods Methods for this study are reported in detail in part I of this study, Chapter 5, section 5.2. Additional methods added below describe statistical analyses for between-group comparisons only.  6.2.1 Statistical analyses 6.2.1.1 Inferential statistics: unadjusted difference method The difference method was used to estimate between-group effect sizes for the difference between patellofemoral OA cases and their matched controls (i.e. patellofemoral and tibiofemoral alignment [see Figure 5.8, Figure 5.9 to review alignment measures] for each patellofemoral OA case minus their respective matched control), in standing at four angles of knee flexion. Cohen’s d was also reported as a standardized effect size estimate of between-group differences within each of the four angles in standing [369 370].   6.2.1.2 Inferential statistics: mixed effects models To evaluate differences between cases and matched controls across all angles and positions in a single model, I constructed mixed effects models, as detailed in section 5.2.7.4.  6.2.1.3 Two-dimensional alignment Between-group differences in alignment were reported for 2D measures to enable qualitative comparison of 3D alignment results with more conventional alignment measures.     163  6.2.1.4 Two-legged vs. one-legged alignment I repeated all statistical analyses (unadjusted difference and mixed effects models) to model the difference between standing two-legged at 30° knee flexion compared to one-legged at 30° knee flexion.   6.3 Results 6.3.1 Patellofemoral alignment  6.3.1.1 Unadjusted between-group differences Table 6.1 reports mean ± SD differences in all four knee flexion angles in standing only (patellofemoral OA minus matched control), with supine at full extension included for comparison (given this is the traditional position in which alignment is measured on MRI [294]). Cohen’s d is also reported. Between-group effect sizes were consistently moderate to large in patellar tilt (patellofemoral OA more laterally tilted than controls at all angles), moderate in patellar proximal translation (OA more proximal), and moderate in patellar lateral translation in full extension only (supine and standing, OA more laterally displaced). On account of the systematic error that was introduced into the data collection in supine knee flexion positions (see section 5.3.3 for details), direct unadjusted between-group comparison of supine flexion alignment (i.e. using the difference method to account for matching) was not performed.  6.3.1.2 Mixed effects models Table 6.2, Table 6.3 and Table 6.4 report the coefficients for all variables included in the mixed effects models for three patellar rotations (these same values were also reported in Chapter 5).   164  Figure 6.1 presents three graphs of the linear prediction from the fitted model for three patellar rotations, combined with the upper and lower bounds of the standard error of the linear prediction (i.e. variability about the mean). The left graph of each figure shows the between-group differences in standing; and the right graph shows between-group differences in supine. Coefficients for three patellar translations are reported in Table 6.5, Table 6.6, and Table 6.7, and Figure 6.2 presents the models graphically.      165  Table 6.1 Between-group difference in alignment of matched pairs (patellofemoral OA [OA] minus control [CONT]) in supine full extension, and in standing across four angles.   Supine, 0° Two-legged, 0°  Two-legged, 15° Two-legged, 30° Two-legged, 45° Patellofemoral joint OA – CONT^  Cohen’s d OA - CONT Cohen’s d OA - CONT Cohen’s d  OA - CONT Cohen’s d OA - CONT Cohen’s d Flexion  0.4 (6.0) 0.1 -1.7 (7.6) -0.2 -3.0 (6.7) -0.4 -2.9 (8.2) -0.4 -2.3 (10.3) -0.2 Medial spin 1.5 (7.2) 0.2 0.7 (7.4) 0.1 -1.0 (6.2) -0.2 1.7 (6.9) 0.2 2.0 (6.5) 0.3 Medial tilt -4.8 (8.0) -0.6* -5.9 (9.1) -0.6 -7.2 (9.2) -0.8 -7.4 (7.7) -1.0 -4.6 (5.6) -0.8 Proximal translation 4.9 (9.3) 0.5 5.8 (12.5) 0.5 4.9 (11.3) 0.4 6.4 (12.2) 0.5 6.7 (10.4) 0.6 Lateral translation 3.0 (5.7) 0.5 2.4 (5.1) 0.5 2.5 (6.0) 0.4 2.4 (5.8) 0.4 1.7 (5.1) 0.3 Anterior translation 0.2 (3.0) 0.1 -0.1 (3.7) 0.0 0.3 (3.9) 0.1 1.3 (4.9) 0.3 2.5 (5.9) 0.4 Tibiofemoral joint           Adduction -0.4 (4.8) -0.1 -0.3 (4.5) -0.1 -0.8 (4.2) -0.2 -0.4 (4.0) -0.1 1.0 (4.3) 0.1 Internal rotation -1.8 (5.6) -0.3 -2.0 (7.0) -0.3 -2.2 (4.3) -0.5 -2.9 (5.3) -0.5 -2.7 (4.0) -0.7 Proximal translation 0.4 (2.1) 0.2 0.7 (2.1) 0.3 1.1 (2.1) 0.5 0.5 (2.4) 0.2 1.3 (3.6) 0.4 Lateral translation -0.1 (2.0) -0.1 0.0 (2.5) 0.0 0.3 (2.9) 0.1 -0.0 (2.6) 0.0 0.3 (3.2) 0.1 Anterior translation 1.1 (3.7) 0.3 1.8 (3.7) 0.5 1.2 (5.3) 0.2 2.3 (5.2) 0.4 1.9 (6.6) 0.3 ^OA-CONT = patellofemoral OA [OA] minus control [CONT] *bold indicates Cohen’s d ≥0.5      166  Table 6.2 Patellar flexion, mixed effects: model coefficients (β) with 95% confidence intervals (CI) and p-values. ‘Patellofemoral OA’ represents difference of OA minus matched control (previously reported in Table 5.5)   TFJ = tibiofemoral joint, PFOA = patellofemoral OA *one participant with patellofemoral OA who had Marie Charcot Tooth, and one control who had scoliosis were removed because their alignment values were influential   Table 6.3 Patellar medial spin, mixed effects: model coefficients (β) with 95% confidence intervals (CI) and p-values. ‘Patellofemoral OA’ represents difference of OA minus matched control (previously reported in Table 5.6)  β 95% CI p-value Patellofemoral OA -0.73  -1.96, 0.49 0.24 Tibiofemoral flexion 0.06 0.02, 0.09 0.001 Tibiofemoral flexion 2 -0.002 -0.004, -0.000 0.024 Position 1.25 0.31, 2.19 0.009 PFOA*TFJ flexion2 0.004  0.001, 0.007 0.008 TFJ = tibiofemoral joint, PFOA = patellofemoral OA   Table 6.4 Patellar medial tilt, mixed effects: model coefficients (β) with 95% confidence intervals (CI) and p-values. ‘Patellofemoral OA’ represents difference of OA minus matched control (previously reported in Table 5.7)  β 95% CI p-value Patellofemoral OA -5.81  -6.86, -4.76 <0.001 Tibiofemoral flexion 0.14 0.09, 0.18 <0.001 Tibiofemoral flexion 2 -0.002 -0.004, 0.000 0.051 Position -0.88 -1.96, 0.21 0.113     Full sample Two participants removed*  β 95% CI p-value β 95% CI p-value Patellofemoral OA -1.30  -2.95, 0.35 0.12 -1.93  -3.57, -0.28 0.022 Tibiofemoral flexion 0.68 0.62, 0.75 <0.001 0.70  0.63, 0.77 <0.001 Tibiofemoral flexion 2 0.006 0.004, 0.008 <0.001 0.006 0.004, 0.008 <0.001 Position 4.07 2.44, 5.71 <0.001 4.40 2.79, 6.02 <0.001 Position*TFJ flexion -0.15 -0.21, -0.08 <0.001 -0.16 -0.23, -0.10 <0.001 PFOA*TFJ flexion -0.08  -0.14, -0.02 0.011 -0.05  -0.12, 0.01 0.08   167    Figure 6.1 Between-group alignment in standing and supine, three rotations. Patellar flexion: a. standing, b. supine; Patellar spine: c. standing, d. supine; Patellar medial tilt: e. standing, f. supine. PFOA = patellofemoral OA; CON = control; UCI = upper bound of standard error; LCI = lower bound of standard error.    168   Table 6.5 Patellar proximal translation, mixed effects: model coefficients (β) with 95% confidence intervals (CI) and p-values. ‘Patellofemoral OA’ represents difference of OA minus matched control (previously reported in Table 5.8).   β 95% CI p-value Patellofemoral OA 3.93  1.74, 6.13 <0.001 Tibiofemoral flexion -0.58 -0.63, -0.52 <0.001 Position 1.40 0.06, 2.73 0.041 PFOA*TFJ flexion 0.13 0.05, 0.21 0.002 TFJ = tibiofemoral joint, PFOA = patellofemoral OA    Table 6.6 Patellar lateral translation, mixed effects: model coefficients (β) with 95% confidence intervals (CI) and p-values. ‘Patellofemoral OA’ represents difference of OA minus matched control (previously reported in Table 5.9).  β 95% CI p-value Patellofemoral OA 2.50  1.77, 3.24 <0.001 Tibiofemoral flexion -0.02 -0.06, 0.03 0.467 Tibiofemoral flexion 2 0.002 0.000, 0.003 0.025 Position 1.90 0.66, 3.14 0.003 Position*TFJ flexion -0.06 -0.11, -0.01 0.014 TFJ = tibiofemoral joint    Table 6.7 Patellar anterior translation, mixed effects: model coefficients (β) with 95% confidence intervals (CI) and p-values. ‘Patellofemoral OA’ represents difference of OA minus matched control (previously reported in Table 5.10).  β 95% CI p-value Patellofemoral OA -0.55  -1.51, 0.40 0.259 Tibiofemoral flexion -0.12 -0.16, -0.08 <0.001 Tibiofemoral flexion 2 -0.002 -0.003, -0.001 <0.001 Position 1.34 0.40, 2.28 0.005 Position*TFJ flexion -0.04  -0.08, -0.01 0.021 PFOA*TFJ flexion 0.06 0.03, 0.10 <0.001 TFJ = tibiofemoral joint, PFOA = patellofemoral OA       169   Figure 6.2 Between-group alignment in standing and supine, three translations. Proximal translation: a. standing, b. supine; Lateral translation: c. standing, d. supine; Anterior translation: e. standing, f. supine. PFOA = patellofemoral OA; CON = control; UCI = upper bound of standard error; LCI = lower bound of standard error.   170  The patellofemoral OA group had less patellar flexion than controls, most notably at higher degrees of knee flexion (Figure 6.1). The estimated effect size for OA was enhanced by the removal of the two participants in sensitivity analyses (one patellofemoral OA participant with Marie Charcot Tooth disease, one control with scoliosis – see section 5.3.3, Table 5.5). Effect sizes were small for patellar flexion (Cohen’s d ≤ 0.4, see Table 6.1). Patellar tilt was significantly and substantially different between groups, with patellofemoral OA knees 6° more laterally tilted than controls through all ranges of knee flexion (Table 5.7, Table 6.2). Importantly, patellar tilt was also the only measure with large effect sizes (Table 6.1).  The patellofemoral OA group had more proximally positioned patellae than controls, with larger differences at higher angles of knee flexion (Table 5.8, Figure 6.2). Effect sizes were moderate (Table 6.1). The OA group was 3 mm more laterally translated compared to controls through all knee flexion ranges (Table 5.9, Figure 6.2, Figure 5.15). Effect size for lateral translation was moderate in full extension only (in both standing and supine) (Table 6.1). Finally, the OA group was generally more anterior translated than controls when knees were flexed, with larger differences at higher angles of knee flexion (Table 5.10, Figure 6.2). Effect sizes for anterior translation were small (≤ 0.4) (Table 6.1).  6.3.1.3 Two-dimensional alignment between-group differences Two-dimensional alignment in a non-weightbearing supine position is reported in Table 6.8. Between-group differences were similar to 3D alignment results, in that the patellofemoral OA group was more laterally tilted (patellar tilt angle), laterally displaced (bisect offset), and proximally displaced (Insall Salvati ratio) compared to controls. However, only bisect offset   171  reached statistical significance. Bisect offset and patellar tilt angle both had moderate effect sizes.   Table 6.8 Two-dimensional alignment results, by group (in supine).   Patellofemoral OA Controls ∆ Cohen’s d p-value Patellar tilt angle (°) 12.61 (7.13) 8.76 (4.92) 3.84 (8.10) 0.5* 0.088 Bisect offset (%) 64.65 (14.55) 54.59 (4.77) 0.10 (0.18) 0.6 0.045 Insall-Salvati ratio 1.16 (0.13) 1.05 (0.16) 0.11 (0.26) 0.4 0.13 *bold indicates Cohen’s d ≥0.5 (i.e. moderate effect size)    6.3.1.4 Two legged vs. one legged squat There were no substantial changes in alignment within either group between two-legged squat at 30° knee flexion to one-legged squat at the same angle (see Table 6.9). The only key measure that reached within-group significance was patellar tilt for the patellofemoral OA group, though the amount of change was small (<1°).   Table 6.9 Within-group difference in alignment (one-legged minus two-legged), standing at 30° flexion. Mixed effects models.  Patellofemoral OA (n=15) Control (n=15)  Patellofemoral joint β (95% CI) p-value β (95% CI) p-value Flexion (°) -1.01 (-2.48, 0.46) 0.18 -1.40 (-2.66, 0.14) 0.029 Medial spin (°) -3.10 (-4.43, -1.77) <0.001 -0.55 (-1.91, 0.81) 0.429 Medial tilt (°) 0.85 (0.10, 1.61) 0.027 -0.20 (-1.43, 1.02) 0.744 Proximal translation (mm) 0.66 (-0.37, 1.69) 0.208 0.60 (-0.09, 1.29) 0.090 Lateral translation (mm) -0.16 (-0.72, 0.39) 0.563 0.10 (-0.43, 0.64) 0.706 Anterior translation (mm) -0.17 (-0.74, 0.40) 0.550 -0.48 (-0.88, -0.07) 0.021 Tibiofemoral joint     Adduction (°) 0.81 (0.11, 1.52) 0.024 0.35 (-0.18, 0.87) 0.197 Internal rotation (°) 0.08 (-1.28, 1.44) 0.907 0.36 (-0.96, 1.69) 0.593 Proximal translation (mm) 0.50 (-0.04, 1.04) 0.070 -0.03 (-0.52, 0.47) 0.920 Lateral translation (mm) -0.26 (-0.75, 0.23) 0.295 0.22 (-0.19, 0.64) 0.288 Anterior translation (mm) -0.24 (-1.21, 0.72) 0.623 1.08 (0.26, 1.89) 0.010     172  Between-group differences in one-legged stance are reported in Table 6.10. The amount of between group differences was not substantially different in one-legged stance compared to two-legged stance. In addition, when evaluating one-legged stance in those with and without malalignment, there were also no substantial differences (see Appendix F).  Table 6.10 Between group differences for patellofemoral OA vs. controls, one legged stance at 30° knee flexion. Mixed methods results:   Patellofemoral joint β (95% CI) p-value Flexion (°) -2.53 (-5.91, 0.86) 0.144 Medial spin (°) -0.62 (-3.69, 2.46) 0.694 Medial tilt (°) -6.30 (-10.44, -2.17) 0.003 Proximal translation (mm) 6.70 (1.72, 11.69) 0.008 Lateral translation (mm) 2.48 (-0.04, 5.00) 0.054 Anterior translation (mm) 1.55 (-0.11, 3.22) 0.068 Tibiofemoral joint   Adduction (°) -0.29 (-2.31, 1.73) 0.780 Internal rotation (°) -3.62 (-5.82, -1.41) 0.001 Proximal translation (mm) 0.76 (-0.40, 1.92)  0.200 Lateral translation (mm) -0.01 (-0.89, 0.87) 0.986 Anterior translation (mm) 1.96 (-0.86, 4.77) 0.173   6.3.2 Tibiofemoral alignment In two-legged stance, the patellofemoral OA group was overall 3° less internally rotated (i.e. more externally rotated) than controls, across all ranges (E.3). Effect sizes were moderate when the knees were flexed but small in full extension (Table 6.1). In one-legged stance at 30° knee flexion, the OA group was 4° more externally rotated than controls (Table 6.10)  In addition, the OA group was overall 2 mm more anterior translated than controls through all knee flexion angles (E.6). The effect size was moderate in full extension only (Table 6.1). This persisted in one-legged stance (2 mm between-group difference - Table 6.10).    173   6.4 Discussion In this chapter, differences in alignment between patellofemoral OA cases and their matched controls were presented. Large between-group differences were observed in several patellar alignment measures. In patellar flexion, the patellofemoral OA group was 1° less flexed than controls in full extension, and up to 5° less flexed by 45° knee flexion. In patellar tilt, the patellofemoral OA group was overall 6° more laterally tilted than the control group at all angles in both standing and supine. In patellar proximal translation, the OA group was 4 mm more proximally translated than controls in full extension, plus 1mm more proximal translation for every 10° of added knee flexion (achieving up to 10° difference by 45° flexion) in both standing and supine. In patellar lateral translation the OA group was 3 mm more laterally translated than controls, possibly more pronounced in full extension and in supine.  While 3D alignment has not previously been measured in standing in patellofemoral OA, the findings of this study can be compared with similar studies of patellofemoral disease. In patellofemoral pain, lateral translation, lateral tilt and proximal translation were all greater compared to controls in both partial and full weightbearing in supine through range, with larger differences in lateral displacement in full extension [317 379]. Individuals with patellar instability also had greater lateral translation, lateral tilt, and proximal translation compared to healthy controls in both supine and standing through range, with larger differences in all three measures in full extension [345]. In supine, a group with patellofemoral pain and maltracking were, in real time, more laterally tilted and translated and more proximally translated compared to healthy controls (larger difference in tilt in full extension, smaller difference in proximal   174  translation in extension) [349]. In the present study, direction of alignment for these three key measures was similar to these patellofemoral pain and instability study results, though only lateral translation was slightly greater in full extension in supine. This supports the theory that patellofemoral pain and OA may lie along the same disease spectrum (see section 1.7.3).  In tibial rotation, the OA group was 3° more externally rotated compared to controls, possibly more so when knees were flexed. This is in contrast with findings in a dynamic MRI study in non-weightbearing supine, in which a group with patellofemoral pain was more internally rotated than a healthy control group [349]. Authors of that study theorized this was a possible compensatory strategy for improving stability of the patella within the trochlea. The differences in the results from the present study may be partly explained by methodological differences (weight bearing status, positioning, and static vs. dynamic image acquisition). However, other studies have found individuals with patellofemoral pain to have tibial external rotation compared to controls [317], and similarly, internal femoral rotation compared to controls [78 79 379]. It is possible that the present study findings of increased external tibial rotation in patellofemoral OA may in part reflect internal rotation of the femur proximally.   The finding of greater tibial anterior translation in patellofemoral OA has not been reported previously in studies of patellofemoral pain or instability. It is possible that this finding represents a strategy to decrease patellofemoral joint reaction force. One speculation as to how this could be achieved would be a reduction in hamstring force that could allow the tibia to move anteriorly on the femur, while also potentially enabling the pelvis to tilt anteriorly, as has been measured previously in those with patellofemoral OA when climbing stairs [173].    175  Standing in 30° knee flexion, one-legged stance did not result in greater malalignment than two-legged stance in either group, even when divided into groups with and without malalignment. It is possible that offering assistive devices for balance (pressure bars) in this study reduced potential changes from two legs to one leg. Notably, the amount of force participants put through the pressure bars was not possible to control, and even though they were asked to use the bars minimally for balance, it is conceivable they relied on the bars more during one-legged stance. Alternatively, additional weightbearing and its associated motor control requirements at this task level do not substantially alter alignment. Rather, it might require a more difficult task such as jumping or running (which could not be measured in a static MR task) to increase load requirements sufficiently to elicit impairment-driven changes in alignment [320-322].   6.4.1 Limitations Methodological limitations for the present study are reported in detail in section 5.4.1.   Importantly, in addition to those described previously, the systematic error introduced by using foam wedges for supine positioning in the 0.5T open-bore scanner (without checking alignment with a goniometer) limited the ability to directly compare between-group alignment in supine at 15°, 30° and 45° knee flexion. This was in part addressed by using the tibiofemoral flexion angle as a continuous variable instead of the four angle indicator variable in mixed effects models. However, results should be interpreted with caution where comparisons between supine and standing in flexed knees are of interest. This chapter focused on standing alignment, which would not have been affected by this methodological error.   176  6.4.2 Clinical implications Out of the 11 degrees of freedom measured in this study, important measures for patellofemoral OA appear to be patellar flexion, patellar tilt, proximal translation, and lateral translation, as well as tibial internal rotation. Tibial anterior translation may also be important, though this is the first study to demonstrate this, and further evaluation is advised.   The position and knee angle recommended for assessment will depend on the specific research question. For example, the overall greatest absolute patellar alignment values (i.e. positioned furthest towards a direction of malalignment) are captured in standing at full extension (all values are more centralized, regardless of group, with either a flexed knee or in supine). In addition, those with patellofemoral OA (and more so those with malalignment) will exhibit greater malalignment when standing, compared to supine. Thus, if a researcher is interested in examining, for example, how much lateral translation individuals can have, it is best to measure this in standing and in full extension.   If wanting to detect the largest between-group differences, patellar flexion, proximal translation and anterior translation have larger between-group differences as the knee moves further into flexion. Generally, the remaining alignment measures have similar between-group differences throughout range. However, results using the difference method in standing (Table 6.1) suggest larger effect sizes may be found for patellar tilt at 30° knee flexion and for tibial rotation at 30° or more, and may be larger in or near full extension for lateral translation.     177  Based on these results, standing in full extension may be the best position for evaluating overall patellofemoral alignment for most research questions. However, if the research question is in regards to a comparison to healthy controls for patellar flexion, patellar proximal translation, patellar anterior translation, or possibly lateral tilt and tibial rotation, then flexion may detect larger between-group differences.   Another clinical implication is that this study did not detect significant findings in 2D proximal translation or lateral tilt, yet these were significant in 3D. This may reflect the fact that 2D alignment was evaluated on images taken unweighted and in supine. This may also reflect that 3D alignment may more accurately quantify alignment and may be a more sensitive measure of alignment than 2D.  A final clinical implication is that the results of this study do not currently justify the use of acquiring images in one-legged positions compared to two-legged at 30° knee flexion. Importantly, this has not yet been evaluated at deeper angles of knee flexion, or with participants self-selecting their position (since our standardization procedures may have influenced how participants stood during scanning). Since this is the first study to evaluate the difference between two-legged and once-legged stance, further evaluation is needed, and clinically this may still be important (i.e. dynamic valgus is common in those with or at risk of knee pathology, and biomechanical impairments are more pronounced with higher level functional tasks) [224 225 246 320-322]. One-legged stance may cause unnecessary discomfort for study participants who may find the task both fatiguing and/or painful. With no important differences found in the present study, two-legged image acquisition may be sufficient to detect and characterize   178  clinically important alignment differences in individuals with patellofemoral OA at the present level of weightbearing.    179  Chapter 7 Conclusion The overarching aim of my thesis was to characterize the associations between measures of patellofemoral alignment, tibiofemoral alignment, and/or trochlear morphology and patellofemoral OA. My primary hypothesis was that knee alignment (patellofemoral and tibiofemoral) and trochlear morphology are associated with prevalent patellofemoral OA. In this chapter, I provide a summary of key findings of each of four studies. I then integrate the findings into the broader clinical and research context. Finally, I suggest future research directions that will further this research field.   7.1 Study #1: Systematic review In my first study (Chapter 2), I conducted a systematic review of the literature to determine what is currently known about the relationship of knee alignment (in all three planes) or trochlear morphology to presence, severity, onset or progression of patellofemoral OA, using any imaging modality.   7.1.1 Key findings and contributions  This study contributes the following key findings to the literature: (i) In the axial plane, there is limited evidence that greater lateral patellar displacement and lateral tilt are associated with prevalence or severity of patellofemoral OA. These measures are more strongly associated with patellofemoral OA features in the lateral compartment. Only one longitudinal study has been published, limiting the ability to imply causation.   180  (ii) In the sagittal plane, there is limited evidence that a higher positioned patella is associated with prevalent patellofemoral OA. The association is more strongly associated with patellofemoral OA features in the lateral compartment. Only one longitudinal study has been published, limiting the ability to imply causation. (iii) In the frontal plane, there is strong evidence that frontal plane alignment is associated with prevalence or severity of patellofemoral OA. Direction of malalignment (varus vs. valgus) may be related to whether the tibiofemoral joint is also involved, and isolated patellofemoral OA may be more commonly valgus. Direction of alignment may be associated with compartment involvement, and two longitudinal studies suggest that frontal plane malalignment in either direction (varus or valgus) is associated with patellofemoral OA worsening. (iv) There is strong evidence of a relationship between trochlear morphology and patellofemoral OA. Specifically, a shallower trochlea is associated, in cross-sectional studies, with prevalence or severity of patellofemoral OA. Two longitudinal studies, however, have conflicting results.  An additional contribution of this study is that I identified several important knowledge gaps in this field of research: (i) Most studies did not specifically recruit individuals with patellofemoral OA.  (ii) Most studies did not include healthy comparison groups. (iii) Reference points identified in the literature to define ‘abnormal’ either did not explain or justify how they were derived [281 290], or were developed in a clinical knee sample [250]. No reference points were validated for patellofemoral OA.   181  (iv) While knee alignment has been evaluated in upright or full weightbearing using radiographs in this population, the same has not been done with MRI in a patellofemoral OA sample.  7.2 Study #2: Alignment and morphology one year following ACL reconstruction In my second study (Chapter 3), I investigated the relationship of patellofemoral alignment (in two planes, axial and sagittal) and trochlear morphology to early patellofemoral OA in a sample of individuals one year following ACLR. I hypothesized that following ACLR, alignment and morphology are associated with early patellofemoral OA, and that the nature of these associations are similar to those seen in non-traumatic patellofemoral OA.  7.2.1 Key findings and contributions  This study evaluated two planes and contributes the following key findings to the literature: (i) Axial plane alignment and morphology had not been previously reported in patellofemoral OA following ACLR (ii) One year post-ACLR, 17% had early radiographic patellofemoral OA (defined as definite patellofemoral osteophytes) (iii) In the axial plane, patellar lateral displacement (bisect offset) and trochlear morphology (sulcus angle, trochlear angle) were associated with radiographic indices of early patellofemoral OA one year after ACLR. Patellar tilt was not associated with OA. (iv) Direction of malalignment for these three measures was similar to that reported in my systematic review, which was of primarily non-traumatic patellofemoral OA.   182  (v) In the sagittal plane, there was no relationship between patellar height (Insall Salvati ratio) and patellofemoral OA one year following ACLR  7.3 Study #3: Framingham community cohort: reference values and dose-response patterns In my third study (Chapter 4), I first determined reference values for patellofemoral alignment (in the axial plane only) and morphology in a subsample of community-dwelling adults aged ≥50 years with no patellofemoral full thickness cartilage damage and no knee pain. To accomplish this I used an existing population-based cohort, the Framingham community cohort. Using the full sample (both with and without patellofemoral full thickness cartilage damage) I explored the dose-response pattern of alignment and morphology measures to the odds of having prevalent patellofemoral OA features across the range of values of each alignment or morphology measure.   7.3.1 Key findings and contributions  This study contributes the following key findings to the literature: (i) For all alignment and morphology measures, there is large within-group variability and substantial overlap between those with and without MRI-defined features of patellofemoral OA (ii) Reference values reported in this study can be used to identify individuals with abnormal alignment or morphology who may be at increased odds of having patellofemoral OA (iii) Two alignment measures (patellar tilt angle, bisect offset) demonstrated a consistent dose-response pattern with prevalent MRI-defined features of patellofemoral OA (full   183  thickness cartilage damage, any cartilage damage, osteophytes, BMLs), with and without knee pain. Odds were higher with increased lateral or medial patellar tilt, and with larger bisect offset (indicating greater lateral displacement) (iv) Three of the four morphology measures (sulcus angle, lateral and medial trochlear inclination) demonstrated a consistent dose-response pattern with prevalent MRI-defined features of patellofemoral OA, with and without knee pain. Odds were higher with a wider or shallower sulcus for all measures (i.e. larger sulcus angle, lower lateral and medial trochlear inclination) (v)  The associations between alignment or morphology and prevalent OA were generally strongest in the lateral patellofemoral compartment, compared to either the medial compartment or the entire patellofemoral joint (vi) The strength of the associations differed when pain was included in the outcome. Lateral trochlear inclination, bisect offset, and patellar tilt angle all had stronger associations with structure combined with pain as an outcome, compared to structure alone (vii) The morphology and alignment measures had similar dose-response patterns with knee pain or pain with stairs alone as compared to outcomes that included MRI-defined structural features  (viii) The strength of associations when pain was included in the outcome (e.g. full thickness cartilage damage plus pain) was qualitatively greater if knee pain was made worse with stairs (compared to knee pain alone)    184  7.4 Study #4: The “MRI open-bore for osteoarthritis Vancouver” (MOOV) study  In my fourth study (Chapter 5 and Chapter 6), I evaluated patellofemoral alignment in 3D in individuals standing and supine. I evaluated the repeatability of these methods in individuals with patellofemoral OA and in asymptomatic controls. In a cross-sectional study with matched controls, I then evaluated 3D patellofemoral alignment in individuals with patellofemoral OA and in asymptomatic controls. I hypothesized that (i) the patella would be more laterally displaced and tilted, and more proximally translated, when standing compared to supine in individuals with and without patellofemoral OA, (ii) that malalignment would be more pronounced in the OA group in standing compared to supine, that (iii) the patella would be more laterally displaced and tilted, and more proximally translated, in patellofemoral OA compared to matched controls in all positions, and (iv) that between-group differences would be greater when transitioning from two-legged stance to one-legged stance.   7.4.1 Key findings and contributions  This study contributes the following key findings to the literature: (i) Repeatability for 3D patellofemoral and tibiofemoral alignment was good to excellent in full weightbearing in individuals with and without patellofemoral OA, with the exception of patellar spin. (ii) Patellofemoral alignment differed between supine and standing within both groups. Both groups were more laterally displaced, more laterally tilted and more proximally translated in upright compared to supine. (iii) Differences between standing and supine, within both groups, were largest in full extension (compared to a bent knee) for key patellofemoral alignment measures. In   185  contrast, tibial anterior translation (the only tibiofemoral measure that differed) had larger differences in knee flexion than extension (iv)  Individuals defined as having malalignment based on 2D alignment measured on traditional non-weightbearing supine MR images were even more malaligned when standing, with qualitatively larger differences than those without malalignment (v) Results of this study support the theory that alignment in standing is more clinically important than supine, and that conventional supine images may not be adequately quantify the extent of malalignment (vi) The group with patellofemoral OA had substantially more patellar lateral tilt, lateral translation and proximal translation than the asymptomatic matched control group. The OA group also had greater tibial external rotation and anterior translation compared to the control group. Tibial anterior translation has not previously been reported in patellofemoral pain or OA. (vii) Findings were similar to previous studies of patellofemoral pain and instability, supporting the theory of patellofemoral pain and OA lying on a disease continuum (viii) Between-group 2D alignment results were similar to 3D alignment results for lateral tilt (patellar tilt angle), lateral translation (bisect offset), and proximal translation (Insall Salvati ratio), however only bisect offset reached statistical significance. 2D alignment based on traditional supine MR images may lack adequate sensitivity in detecting malalignment (ix) One-legged stance did not result in greater malalignment than two-legged stance in either group, even when divided into groups with and without malalignment.   186  (x) Standing in full extension may be the best position for addressing many research questions. Between group comparisons for patellar tilt, patellar proximal translation, or tibial rotation may be better assessed in knee flexion. Patellar lateral translation may be measured in supine unless malalignment is suspected, in which case standing may still be preferred.  7.5 Clinical implications In this dissertation, I addressed several key knowledge gaps that have potential clinical implications in this field. I synthesized the available literature evaluating the relationships between alignment or morphology and features of patellofemoral OA (Chapter 2) [294]. I further characterized the association between knee alignment or trochlear morphology and patellofemoral OA across several populations, including in younger adults following ACLR (Chapter 3) [380], middle aged individuals with patellofemoral OA (Chapter 5, Chapter 6), and in a large population-based cohort of adults aged over 50 (Chapter 4) [381]. Finally, I expanded on the literature by evaluating the influence on alignment of standing and squatting at various knee flexion angles, in comparison to supine positions commonly used in conventional MRI (Chapter 5, Chapter 6).   The results of my work suggest that patellofemoral OA is associated with increased patellar lateral displacement, lateral tilt (and less so medial tilt), proximal translation, tibial external rotation, and possibly decreased patellar flexion and increased tibial anterior translation (Chapter 2, Chapter 3, Chapter 4, Chapter 5, Chapter 6). Patellofemoral OA is also associated with a shallower trochlear groove. Such findings may be observable even very early in the OA process,   187  as evidenced by findings of increased lateral displacement, wider sulcus angle, and higher trochlear angle in patellofemoral OA as early as one year following a traumatic knee injury (Chapter 3) [380]. The relationship exhibits a dose-response pattern, with increased severity of malalignment or altered morphology associated with higher odds of prevalent patellofemoral OA and/or knee pain (Chapter 4) [381]. Having malalignment in supine may indicate even greater malalignment in standing (Chapter 5), suggesting the importance of evaluating alignment in upright, functional positions (Chapter 6).   The clinical implications of this work relate to how these findings contribute to understanding of the natural history of patellofemoral OA, physical examination, clinical reasoning, and primary and secondary prevention and treatment. Alignment and morphology may be clinically important along the spectrum of patellofemoral disease, from patellofemoral pain syndrome [249], to post-ACL injury or reconstruction (Chapter 3) [191 304 380], to early patellofemoral OA (Chapter 3, Chapter 4, Chapter 6) [380 381] and through to more advanced multicompartment knee OA (Chapter 2, Chapter 4) [294 381]. Addressing alignment as part of treatment for the patellofemoral joint has the potential for impact anywhere along this OA spectrum.   7.5.1 Possible clinical contributors to malalignment Alignment is potentially modifiable and may thus be important in the prevention or management of patellofemoral OA. The patella has considerable freedom of movement relative to the femur on account of the joint’s bony morphology and surrounding soft tissues (possibly more so in a dysplastic trochlea). The tibia also has considerable movement relative to the femur. Movement about these joints is controlled by a combination of active and passive mechanisms.   188  Understanding these mechanisms and how they contribute to malalignment may inform treatment.  The neuromuscular system actively controls lower limb position and thus alignment [293]. Impairments (e.g. muscle weakness or poor muscle coordination) of the lumbopelvic core and hip abductor/rotator muscles [208 209] may contribute to femoral internal rotation, decreased postural stability and even altered foot and ankle mechanics during functional tasks, which in turn may contribute to patellar alignment [78 79 382]. Quadriceps muscle impairments such as reduced muscle volume, weakness, or delayed muscle firing in vastus medialis obliquus (VMO) compared to lateralis (VML) [192 201 213 383] may influence patellar alignment through the direct attachment of the quadriceps to the patella and tibia (as well as femur and pelvis). Weakness or poor motor control of the intrinsic foot muscles may cause dynamic collapse (pronation) distally, which may result in internal tibial and femoral rotation [193].   Passive mechanisms such as bony morphology or soft tissue structure may also contribute to alignment. For example, reduced range of motion in ankle dorsiflexion (e.g. due to short gastro-soleus muscles or talocrural joint stiffness) could lead to dynamic collapse (pronation) through the midfoot, subtalar joint and the tibiofemoral joint during functional tasks such as stair descent [384]. Alternatively, midfoot ligamentous laxity, or altered foot alignment (e.g. increased forefoot varus) [193 384] could result in a similar pattern.     189  7.5.2 Clinical measurement of patellofemoral alignment Imaging will not generally be available outside of an orthopaedic setting. Therefore, clinicians will need other methods for evaluating alignment, or potential contributing factors, in their patients with patellofemoral OA. Unfortunately, physical examination of patellar position has not yet been found to be reliable [385 386], and there is a lack of affordable, reliable, valid and easy special tests for evaluating knee alignment. Based on best available evidence, clinicians could use Herrington’s test of patella medio-lateral position (see Figure 7.1) [387]. In the sagittal plane, a clinical screening tool has been described that uses calipers to measure the ratio of the patellar tendon length (patellar apex to tibial tuberosity) to patellar height (from base to apex), using a ratio ≥1.2 to define patella alta [241 250]. However, this tool has not been validated.  Figure 7.1 Clinical method for assessing patella position. Mark femoral condyles and mid-point of patella on tape. Calculate the distance from medial condyle to mid-patella, then subtract the distance from lateral condyle to mid-patella. Reprinted with permission31                                                  31 Reprinted with permission from Journal of Orthopaedic & Sports Physical Therapy, Volume 38, Herrington L. “The Difference in a Clinical Measure of Patella Lateral Position Between Individuals With Patellofemoral Pain and Matched Controls”, pp 59-62, 2008 doi:10.2519/jospt.2008.2660. Copyright ©Journal of Orthopaedic & Sports Physical Therapy®.   190  Trochlear morphology cannot be directly assessed in clinic because of the overlying patella. However, passive patellar mobility/stability can be evaluated, and this, combined with patellar alignment measures, can give an overall sense of what might be occurring at the patellofemoral joint functionally.  In addition to static alignment tests, a movement test such as a single leg squat [388 389] could be used to evaluate gross movement patterns that may contribute to malalignment (Figure 7.2). Patterns to look for include trunk lean, contralateral pelvis drop, dynamic knee valgus (femoral internal rotation and adduction, relative tibial external rotation), reduced ankle dorsiflexion, or foot pronation. Other contributing factors such as strength, range of motion, muscle length, and motor control during functional tasks (e.g. stairs, squat, gait – particularly pain-provoking tasks) should be evaluated.              Figure 7.2 Single-leg squat test, four types of performance. A. good performance; B. poor overall and trunk performance; C. poor pelvis and hip performance; D. poor hip and knee performance. Reprinted with permission32                                                   32 Reprinted from The American Journal of Sports Medicine, Volume 39, Crossley KM et al. “Performance on the single leg squat indicates hip abductor muscle function”, pp 866-873, 2011 doi: 10.1177/0363546510395456, with permission from SAGE Publications   191  7.5.3 Treating malalignment in patellofemoral OA – state of the art Effective treatments aimed at correcting abnormal alignment and/or movement patterns are needed, and may play a key role in reducing pain, improving function, improving quality of life, and ultimately preventing or slowing the progression of knee OA. The following section focuses on treatments aimed at correcting malalignment, or targeting contributors of malalignment.  7.5.3.1 Conservative approach is generally preferred over surgical  Conservative treatment is recommended as a first treatment approach for malalignment, and only after its failure should surgery be considered, if at all [390 391]. In fact, surgical options are usually considered only in cases of severe patellar instability and recurrent dislocations, or as ‘salvage’ surgeries [392]. Surgical options for correcting patellar instability include: medial patellofemoral ligament reconstruction, tibial tubercle osteotomy/transfer, medial soft tissue realignment, lateral retinacular repair, lateral soft tissue reconstruction, or trochleoplasty [393-399]. However, as with any knee trauma, there is a risk of provoking OA changes with surgery [392].   For treatment of patellofemoral OA, second generation patellofemoral arthroplasty may be considered [390]. Advantages of this procedure over total knee arthroplasty include that it is less invasive, involves quicker post-operative recovery, and preserves bone stock [400]. However, long-term follow-up studies are still needed and failure rates continue to be problematic, the most common failures being pain, OA progression and patellar maltracking [401-403].    192  7.5.3.2 Conservative patellofemoral OA treatments with or without alignment outcomes Currently, there is a paucity of research investigating non-surgical treatments for patellofemoral OA, and none that have targeted the subset of individuals with patellofemoral OA who also have knee malalignment or trochlear dysmorphia (see Table 7.1). Studies to date have evaluated short- to mid-term follow-up time periods with primary outcomes focused on pain and function, and less so on OA structural features or alignment.  Table 7.1 Overview of studies investigating interventions for patellofemoral OA (with or without alignment outcomes) and exercise interventions with alignment outcomesPaper Sample n   Treatment Results Quilty et al. 2003 [215] Patellofemoral OA 43 Rx /  40 No Rx RCT 10 weeks (PT 1/wk + daily home exercises): taping, VMO exercises, EMG biofeedback, posture retraining, footwear, weight loss  vs. no treatment 5 and 12 months: no change in pain or quadriceps strength; no differences between groups Crossley et al. 2015 [147] Patellofemoral OA 39 Rx /  42 Sham RCT  12 weeks (PT 1/wk x 4 wks; then 1/2wks x 8 wks, home exercise 4 days/wk): taping, functional exercises, quadriceps and hip strengthening exercises, manual therapy, education vs. education 3 months: lower pain in treatment group, no change in function either group  6 months: no longer differences between group Crossley et al. 2009 [146] Patellofemoral OA and healthy controls 14 OA /  14 CONT Single acute taping (medial glide, superior tilt, infrapatellar fat pad unload) vs. sham tape Tape reduced lateral displacement and lateral tilt, decreased pain with squatting Edmonds et al. 2016 [404] Knee OA 100 (61) Participants received all 3: single application of therapeutic tape vs. sham tape vs. no tape Therapeutic tape reduced pain during walking; no change in gait mechanics Hinman et al. 2003 [405] Knee OA 18 OA Participants received all 3: single application of therapeutic tape vs. sham tape vs. no tape Therapeutic tape reduced pain during stairs, walking and step test; only improved function on step test Hinman et al. 2003 [406] Knee OA 29 Rx 1 29 Rx 2 29 Rx 3 RCT 3 weeks of taping (either therapeutic or sham or none) 3 weeks: Therapeutic tape reduced pain and improved self-reported function (WOMAC) more so than sham tape    193  Paper Sample n   Treatment Results 6 weeks: therapeutic tape reduced pain and function; sham tape reduced function. Hunter et al. 2011[407]    McWalter et al. 2011 [364] Lateral patellofemoral OA 67 OA RCT crossover:  6 weeks wearing Bio Skin Q brace (Ossur, U.K.) with T strap vs. without T strap, at least 4 hours per day. 6 weeks: reduced pain in both treatments;     significant improvement in 3D alignment with brace donned properly vs. no brace (4° less patellar flexion, 1° less lateral tilt, 1 mm less proximal translation; 0.5 mm less lateral and anterior translation) Callaghan et al. 2015 [214]   Callaghan et al. 2016 [295]   Callaghan et al. 2016 [408] Patellofemoral OA 117 RCT  6 weeks Bio Skin Q brace vs. no brace (~7 hrs/day) 6 weeks: reduced pain, improved self-reported function (KOOS); reduced patellofemoral BMLs by 491 mm3  2D alignment patella 0.5 mm less anterior translation, larger patellofemoral contact area with brace vs. no brace  12 weeks of wear did not result in weaker quadriceps Yu et al. 2015  [409]     Zhang et al. 2017 [410] Patellofemoral OA 50 brace Pragmatic trial, up to 52 weeks: Tru-pull Lite brace (Don Joy, U.S.) in addition to outpatient multidisciplinary program (education, exercises, weight loss, diet advice, pain medication review, etc.) 52 weeks: brace made no difference over program without brace     26 weeks: no more likely to benefit from brace if patella is malaligned or trochlea shallow Wong et al. 2009 [293] Asymptomatic, healthy knees 48 RCT 8 weeks of hypertrophy vs. strength training 8 weeks: decreased lateral patellar glide, both groups; no change in alignment Pourahmadi et al. 2016 [411] Men with knee extension syndrome [412] 46 Pre-post 4 weeks passive lower extremity stretches Decreased Insall-Salvati ratio (1.28 ± 0.10 to 1.10 ± 0.08); decreased patellar tilt angle (15.19° ± 6.36 to 7.54° ± 4.90)  Studies investigating patellar taping plus infrapatellar fat pad taping (see Figure 7.3) showed improvements in patellar alignment [146] and reduced pain during functional tasks [146 404-  194  406]. Reported function was also improved after three weeks of taping, and balance was improved with wear [405 406] but not gait mechanics [404]. Long term use of tape (or incorrectly applied tape) may lead to skin breakdown, and some individuals are allergic to the tape’s adhesive. It may therefore be most clinically useful for acute exacerbations, or to enable patients to participate in therapeutic exercises that they would otherwise not do because of pain.        Figure 7.3 Taping to unload the infrapatellar fat pad. Reprinted with permission33  A similar intervention that is less prone to skin breakdown is knee bracing. Studies of a patellar knee brace (Bio Skin Q by Ossur, U.K. – see Figure 7.4) showed reduced pain, improved self-reported function and reduced volume of patellofemoral BMLs after six weeks of wear, whether or not the brace was donned properly with the realignment T strap affixed [214 407]. MR imaging showed small but significant changes in alignment, notably less anterior patellar translation and larger patellofemoral joint contact area [295 364]. A different brace (Tru-pull Lite by Don Joy, U.S.) did not show any benefit with up to a year of bracing plus outpatient multidisciplinary treatment over treatment without the brace [409], even if participants had                                                  33Reprinted from Clinics in Sports Medicine, Volume 21, McConnell J, “The physical therapist's approach to patellofemoral disorders”, pp. 363-387, 2002, with permission from Elsevier.     195  malalignment or a shallower trochlea at baseline [410]. It is unclear whether the differences among these studies were due to brace type or study design.              Figure 7.4 BioSkin Q Brace a. brace with the realigning T strap intact. b. diagram demonstrating medial glide action of the T strap cupping the lateral aspect of the patella. Reprinted with permission34  Two randomized controlled trials (RCT) investigated multi-modal physiotherapy treatment [147 215] which suggested that within a multi-modal program, targeted exercises including quadriceps and hip strengthening, as well as functional exercises may lower pain more than quadriceps exercises. Function did not improve in either study. Alignment was not measured in either of these two studies.                                                  34 Reprinted from Osteoarthritis and Cartilage, Volume 19, Hunter DJ et al, “A randomized trial of patellofemoral bracing for treatment of patellofemoral osteoarthritis”, pp. 792-800, 2011, with permission from Elsevier.     196  7.5.3.3 Treatments targeting patellar alignment  Beyond the OA literature, two studies investigated the effect of exercise therapy on patellar alignment. Strengthening and hypertrophy programs improved patellar stability in healthy knees [293], while in men with a movement impairment syndrome [412], four weeks of passive stretches improved patellar alignment [411]. These studies could inform approaches to treating patellofemoral OA. Quadriceps muscle length has not been evaluated in patellofemoral OA or in relation to patellar height, but it is short in individuals with patellofemoral pain [413]. In addition, in advanced knee OA, individuals with severely limited knee ROM have tethering adhesions of the quadriceps muscles to the underlying femur [414]. Furthermore, quadriceps weakness is associated with patellofemoral OA [192 196 201]. Therefore, quadriceps strengthening and stretching programs could beneficially impact patellar alignment or stability in patellofemoral OA.  7.5.3.4 Treatments for patellofemoral pain without alignment outcomes A considerable body of literature investigating treatments for patellofemoral pain, including systematic reviews and best practice guidelines [415-422], is available to inform treatment approaches for patellofemoral OA. While not measuring alignment outcomes, the studies do address features known to potentially alter alignment (e.g. hip weakness). Based on this literature, evidence supports strengthening exercises (both quadriceps and hip), neuromuscular exercises, taping, bracing, and/or foot orthoses (prefabricated or flat inserts suffice). Manual therapy of the knee and full kinetic chain in combination with physical therapy is beneficial [417]. Other suggestions include targeted distal and core strengthening, and stretching of tight lower extremity muscles [48 415].    197   Regarding type of exercises, it has not yet been confirmed whether open or closed kinetic chain exercises are superior for treating patellofemoral pain [422], but expert opinion suggests use of open chain exercises early in rehabilitation to target specific movements, and closed chain exercises overall for improving function [415]. Patellofemoral joint stress could be reduced by performing open chain exercises with ankle weights rather than machines, and limiting knee flexion range to between 45° and 90° [423], while for closed chain exercises, squats should be performed between 0° and 45° [423]. These could be important considerations for treating individuals with malalignment who may have reduced contact area and increased joint stress.   7.5.3.5 Who will benefit most from treating alignment? Treatment for patellofemoral OA may differ based on whether or not there is malalignment. In an individual with malalignment, improved alignment could result in greater patellofemoral joint contact area and reduced patellofemoral joint stress [71 81 214 242 424], which could improve pain and function, but might also be disease modifying via the reduced joint stress [214].  Some individuals are more likely to respond to interventions targeting alignment. For example, in women with patellofemoral pain, an eight week exercise program was more effective at reducing pain for those who had an increased patellar tilt angle that was reducible with a quadriceps contraction at baseline [425]. Other prognostic indicators suggest hypermobility may result in poorer outcomes to bracing and exercises, while smaller lateral patellofemoral angle may predict successful outcomes to taping [426]. Overall, however, there is low evidence regarding prognostic indicators to treatment in patellofemoral pain [427]. In the absence of such   198  evidence for patellofemoral OA, the best approach is likely one informed by the patellofemoral OA and patellofemoral pain research presented here, specifically targeting an individual’s risk factor profile [428].  7.5.4 The bigger picture – public health In Chapter 1, I reviewed the personal and societal burden associated with OA. To conclude this dissertation, I would like to return to the broader context of burden, and position the findings of this dissertation within the overall impact of OA. To reiterate, knee OA is responsible for more physical disability than any other disease among older community-dwelling adults in the U.S. [18 19]. On account of pain and physical impairment, individuals living with patellofemoral OA (or knee OA) over time will typically participate less in physical activity, including exercise, leisure, and activities of daily living [217 429]. This can lead to secondary deconditioning, weight gain, further degeneration and loss of productivity [429 430].   Physical inactivity is currently the fourth leading risk factor for global mortality [431]. Thus, deconditioning and weight gain associated with OA may represent an even more important public health risk than the OA itself. OA is substantially more prevalent in individuals with heart disease, metabolic syndrome, diabetes, hypertension, obesity, hyperglycemia, and hypercholesterolemia [199 432]. OA is also associated with increased risk of developing peptic ulcers and renal disease [199]. Even in early hip and knee OA (defined as hip or knee pain in adults aged 45 – 65), 67% have at least one comorbidity [433] and nearly all participants aged over 50 with clinical or radiographic hip or knee OA have at least one comorbidity [434]. The   199  comorbidity burden in turn is associated with worse OA-related outcomes, including worse pain, function, fatigue, insomnia, and longer hospital stays [433 435 436].   Both OA and associated comorbidities have the potential to be addressed through self-management education, physical activity, and healthy lifestyle choices [432]. The greatest potential for impacting knee OA (and possibly the associated sequelae) may lie in directing clinical resources and research efforts towards effective early intervention [2 15].            Figure 7.5 Obesity is associated with multiple comorbidities: OA, Alzheimer disease, stroke, cardiovascular disease, type 2 diabetes, liver disease, and colon cancer. Reprinted with permission35                                                    35 Reprinted from Exercise & Sport Science Reviews, Volume 33, Griffin T and Farshid G, “The role of mechanical loading in the onset and progression of osteoarthritis”, pp. 195-200, 2005, with permission from Wolters Kluwer Health, Inc.     200  7.6 Future directions The findings in this dissertation add to the existing literature. OA may begin in the patellofemoral joint and progress to multicompartment knee OA [43 45 64]. Knee OA is directly associated with substantial burden, and is also associated with comorbidities that further impact morbidity and mortality. Malalignment may be modifiable and thus treatable [146 295 411 425]. Thus, the findings of my research provide a strong rationale for prioritizing future research in this field.  Longitudinal studies are needed to confirm how and when patellofemoral OA develops following ACLR and how alignment and morphology contribute to OA onset and worsening following traumatic knee injury or surgery. In addition, the reference values identified in Chapter 4 [381] could be used in a large, longitudinal study to help determine the best alignment cut-points for predicting OA onset and worsening. This could be done within the context of a prospective study of multiple risk factors for better characterizing those most at risk for onset or worsening of patellofemoral OA as a distinct phenotype. Future studies should also aim to develop valid, reliable, and feasible clinical tests for evaluating patellar alignment. Clinical trials are also needed, and should include alignment as an outcome measure. Interventions would ideally be targeted towards individuals identified as having malalignment, though studies investigating patellofemoral OA are also needed.  These suggested future research questions are congruent with current research priorities identified by EULAR that include improving our understanding of the pathology of the earliest stages of OA; and identifying predictors of OA progression, particularly those that can assist   201  with developing targeted interventions [63]. Furthering the work in this field offers the potential to dramatically impact outcomes in knee OA, a medical condition of immense worldwide importance.     202  References 1. Centers for Disease Control and Prevention. Osteoarthritis fact sheet. Secondary osteoarthritis fact sheet  2017. https://www.cdc.gov/arthritis/basics/osteoarthritis.htm.  2. Lawrence RC, Felson DT, Helmick CG, Arnold LM, Choi H, Deyo RA et al. Estimates of the prevalence of arthritis and other rheumatic conditions in the United States. Part II. Arthritis & Rheumatism 2008;58(1):26-35 doi: http://dx.doi.org/10.1002/art.23176 3. Cisternas MG, Murphy L, Sacks JJ, Solomon DH, Pasta DJ, Helmick CG. Alternative methods for defining osteoarthritis and the impact on estimating prevalence in a US population-based survey. Arthritis Care & Research 2016;68(5):574-80 doi: https://dx.doi.org/10.1002/acr.22721 4. Arthritis Alliance of Canada. The impact of arthritis in Canada: today and over the next 30 years 2011. http://www.arthritisalliance.ca/images/PDF/eng/Initiatives/20111022_2200_impact_of_arthritis.pdf 5. Vos T, Flaxman AD, Naghavi M, Lozano R, Michaud C, Ezzati M et al. Years lived with disability (YLDs) for 1160 sequelae of 289 diseases and injuries 1990-2010: a systematic analysis for the Global Burden of Disease Study 2010.[Erratum appears in Lancet. 2013 Feb 23;381(9867):628 Note: AlMazroa, Mohammad A [added]; Memish, Ziad A [added]]. Lancet 2012;380(9859):2163-96 doi: http://dx.doi.org/10.1016/S0140-6736(12)61729-2. 6. Cross M, Smith E, Hoy D, Nolti S, Ackerman I, Fransen M et al. The global burden of hip and knee osteoarthritis: estimates from the global burden of disease 2010 study. Annals of the Rheumatic Diseases 2014:annrheumdis-2013-204763    203  7. Hootman JM, Helmick CG. Projections of US prevalence of arthritis and associated activity limitations. Arthritis & Rheumatism 2006;54(1):226-9 doi: https://dx.doi.org/10.1002/art.21562 8. Woolf AD, Pfleger B. Burden of major musculoskeletal conditions. Bulletin of the World Health Organization 2003;81(9):646-56 9. Ma VY, Chan L, Carruthers KJ. Incidence, prevalence, costs, and impact on disability of common conditions requiring rehabilitation in the United States: stroke, spinal cord injury, traumatic brain injury, multiple sclerosis, osteoarthritis, rheumatoid arthritis, limb loss, and back pain. Archives of Physical Medicine and Rehabilitation 2014;95(5):986-95. e1 10. Centers for Disease Control and Prevention. National and state medical expenditures and lost earnings attributable to arthritis and other rheumatic conditions--United States, 2003. MMWR Morb Mortal Wkly Rep 2007;56(1):4-7 11. Leigh JP, Seavey W, Leistikow B. Estimating the costs of job related arthritis. Journal of Rheumatology 2001;28(7):1647-54  12. Deshpande BR, Katz JN, Solomon DH, Yelin EH, Hunter DJ, Messier SP et al. Number of persons with symptomatic knee osteoarthritis in the US: impact of race and ethnicity, age, sex, and obesity. Arthritis Care & Research 2016;68(12):1743-50  13. Pereira D, Peleteiro B, Araujo J, Branco J, Santos RA, Ramos E. The effect of osteoarthritis definition on prevalence and incidence estimates: a systematic review. Osteoarthritis & Cartilage 2011;19(11):1270-85 doi: https://dx.doi.org/10.1016/j.joca.2011.08.009 14. Nelson A, Jordan J. Defining osteoarthritis: a moving target. Osteoarthritis & Cartilage 2012;20:1-3    204  15. Losina E, Weinstein AM, Reichmann WM, et al. Lifetime risk and age at diagnosis of symptomatic knee osteoarthritis in the US. Arthritis Care & Research 2013;65(5):703-11  16. Nguyen US, Zhang Y, Zhu Y, Niu J, Zhang B, Felson DT. Increasing prevalence of knee pain and symptomatic knee osteoarthritis: survey and cohort data. Ann Intern Med 2011;155(11):725-32 doi: https://dx.doi.org/10.7326/0003-4819-155-11-201112060-00004 17. Murphy L, Schwartz TA, Helmick CG, Benner JB, Tudor G, Koch G et al. Lifetime risk of symptomatic knee osteoarthritis. Arthritis & Rheumatism 2008;59(9):1207-13 doi: https://dx.doi.org/10.1002/art.24021 18. Guccione AA, Felson DT, Anderson JJ, Anthony JM, Zhang Y, Wilson PW et al. The effects of specific medical conditions on the functional limitations of elders in the Framingham Study. Am J Public Health 1994;84(3):351-8  19. Centers for Disease Control and Prevention. Prevalence of disabilities and associated health conditions among adults--United States, 1999. MMWR. Morbidity and mortality weekly report 2001;50(7):120  20. Alnahdi AH, Zeni JA, Snyder-Mackler L. Muscle impairments in patients with knee osteoarthritis. Sports Health 2012;4(4):284-92  21. Steultjens M, Dekker J, Van Baar M, Oostendorp R, Bijlsma J. Range of joint motion and disability in patients with osteoarthritis of the knee or hip. Rheumatology 2000;39(9):955-61  22. Hatfield GL, Morrison A, Wenman M, Hammond CA, Hunt MA. Clinical tests of standing balance in the knee osteoarthritis population: systematic review and meta-analysis. Physical Therapy 2016;96(3):324-37 doi: 10.2522/ptj.20150025   205  23. Ettinger WH, Davis MA, Neuhaus JM, Mallon KP. Long-term physical functioning in persons with knee osteoarthritis from NHANES. I: Effects of comorbid medical conditions. Journal of Clinical Epidemiology 1994;47(7):809-15  24. Ware JE. SF-36 health survey update. Spine  2000;25:3130-39  25. Alkan BM, Fidan F, Tosun A, Ardıçoğlu Ö. Quality of life and self-reported disability in patients with knee osteoarthritis. Modern Rheumatology 2014;24(1):166-71  26. Kiadaliri AA, Lamm CJ, de Verdier MG, Engstrom G, Turkiewicz A, Lohmander LS et al. Association of knee pain and different definitions of knee osteoarthritis with health-related quality of life: a population-based cohort study in southern Sweden. Health and Quality of Life Outcomes 2016;14(1):121  27. Culvenor AG, Collins NJ, Guermazi A, Cook JL, Vicenzino B, Khan KM et al. Early knee osteoarthritis is evident one year following anterior cruciate ligament reconstruction: a magnetic resonance imaging evaluation. Arthritis Rheumatol 2015;67(4):946-55 doi: http://dx.doi.org/10.1002/art.39005 28. Filbay SR, Ackerman IN, Russell TG, Macri EM, Crossley KM. Health-related quality of life after anterior cruciate ligament reconstruction: a systematic review. American Journal of Sports Medicine 2014;42(5):1247-55 doi: http://dx.doi.org/10.1177/0363546513512774 29. Ackerman IN, Bucknill A, Page RS, Broughton NS, Roberts C, Cavka B et al. The substantial personal burden experienced by younger people with hip or knee osteoarthritis. Osteoarthritis & Cartilage 2015;23(8):1276-84 doi: https://dx.doi.org/10.1016/j.joca.2015.04.008 30. Agency for Healthcare Research and Quality. The top five most expensive conditions treated in U.S. hospitals.   206  https://www.ahrq.gov/sites/default/files/wysiwyg/research/data/data-infographics/images/most-expensive-conditions.html 31. Agency for Healthcare Research and Quality. Most frequent conditions in U.S. hospitals, 2011. https://www.hcup-us.ahrq.gov/reports/statbriefs/sb162.pdf 32. Losina E, Paltiel AD, Weinstein AM, Yelin E, Hunter DJ, Chen SP et al. Lifetime medical costs of knee osteoarthritis management in the United States: impact of extending indications for total knee arthroplasty. Arthritis Care & Research 2015;67(2):203-15  33. Ruiz D, Koenig L, Dall TM, Gallo P, Narzikul A, Javad P et al. The direct and indirect costs to society of treatment for end-stage knee osteoarthritis. J Bone Joint Surg Am 2013;95(16):1473-80  34. Salmon J, Rat A, Sellam J, Michel M, Eschard JP, Guillemin F et al. Economic impact of lower-limb osteoarthritis worldwide: a systematic review of cost-of-illness studies. Osteoarthritis & Cartilage 2016;24(9):1500-08  35. Hubertsson J, Petersson IF, Thorstensson CA, Englund M. Risk of sick leave and disability pension in working-age women and men with knee osteoarthritis. Annals of the Rheumatic Diseases 2012:annrheumdis-2012-201472  36. Grotle M, Hagen KB, Natvig B, Dahl FA, Kvien TK. Prevalence and burden of osteoarthritis: results from a population survey in Norway. The Journal of Rheumatology 2008;35(4):677-84  37. Kotlarz H, Gunnarsson CL, Fang H, Rizzo JA. Osteoarthritis and absenteeism costs: evidence from US National Survey Data. J Occup Environ Med 2010;52(3):263-8 doi: https://dx.doi.org/10.1097/JOM.0b013e3181cf00aa   207  38. Agaliotis M, Fransen M, Bridgett L, Nairn L, Votrubec M, Jan S et al. Risk factors associated with reduced work productivity among people with chronic knee pain. Osteoarthritis & Cartilage 2013;21(9):1160-69  39. Gray H. Anatomy of the Human Body.: Philadelphia: Lea & Febiger, 1918. Bartleby.com, 2000 40. Thomas S, Rupiper D, Stacy GS. Imaging of the patellofemoral joint. Clinics in Sports Medicine 2014;33(3):413-36  41. Hinman RS, Crossley KM. Patellofemoral joint osteoarthritis: an important subgroup of knee osteoarthritis. Rheumatology 2007;46(7):1057-62  42. Crossley K, Hinman R. The patellofemoral joint: the forgotten joint in knee osteoarthritis. Osteoarthritis & Cartilage 2011;19:765-767 43. Ahlbäck S. Osteoarthrosis of the knee. A radiographic investigation. Acta Radiologica: Diagnosis 1968:Suppl 277: 7  44. McAlindon TE, Snow S, Cooper C, Dieppe PA. Radiographic patterns of osteoarthritis of the knee joint in the community: the importance of the patellofemoral joint. Annals of the Rheumatic Diseases 1992;51(7):844-9  45. Duncan R, Peat G, Thomas E, Hay EM, Croft P. Incidence, progression and sequence of development of radiographic knee osteoarthritis in a symptomatic population. Annals of the Rheumatic Diseases 2011;70(11):1944-8 doi: http://dx.doi.org/10.1136/ard.2011.151050 46. Kobayashi S, Pappas E, Fransen M, Refshauge K, Simic M. The prevalence of patellofemoral osteoarthritis: a systematic review. Osteoarthritis & Cartilage 2015;23:A189    208  47. Hart HF, Stefanik JJ, Wyndow N, Machotka Z, Crossley KM. The prevalence of radiographic and MRI-defined patellofemoral osteoarthritis and structural pathology: a systematic review and meta-analysis. Br J Sports Med 2017:bjsports-2017-097515  48. Crossley KM, Stefanik JJ, Selfe J, Collins NJ, Davis IS, Powers CM et al. 2016 Patellofemoral pain consensus statement from the 4th International Patellofemoral Pain Research Retreat, Manchester. Part 1: Terminology, definitions, clinical examination, natural history, patellofemoral osteoarthritis and patient-reported outcome measures. British Journal of Sports Medicine 2016;50(14):839-43 doi: https://dx.doi.org/10.1136/bjsports-2016-096384 49. Stefanik JJ, Niu J, Gross K, Roemer FW, Guermazi A, Felson DT. Using magnetic resonance imaging to determine the compartmental prevalence of knee joint structural damage. Osteoarthritis & Cartilage 2013;21(5):695-99  50. Ledingham J, Regan M, Jones A, Doherty M. Radiographic patterns and associations of osteoarthritis of the knee in patients referred to hospital. Annals of theRrheumatic Diseases 1993;52(7):520-26  51. Duncan RC, Hay EM, Saklatavala J, Croft PR. Prevalence of radiographic osteoarthritis - it all depends on your point of view. Rheumatology 2006;45:757-60  52. Wood L, Muller S, Peat G. The epidemiology of patellofemoral disorders in adulthood: a review of routine general practice morbidity recording. Prim Health Care Res Dev 2011;12(2):157-64 doi: http://dx.doi.org/10.1017/S1463423610000460 53. Cicuttini FM, Baker J, Hart DJ, Spector TD. Association of pain with radiological changes in different compartments and views of the knee joint. Osteoarthritis & Cartilage 1996;4(2):143-7    209  54. Kornaat PR, Bloem JL, Ceulemans RY, Riyazi N, Rosendaal FR, Nelissen RG et al. Osteoarthritis of the knee: association between clinical features and MR imaging findings Radiology 2006;239(3):811-17  55. Hunter DJ, March L, Sambrook PN. The association of cartilage volume with knee pain. Osteoarthritis & Cartilage 2003;11(10):725-9  56. Duncan R, Peat G, Thomas E, Wood L, Hay E, Croft P. Does isolated patellofemoral osteoarthritis matter? Osteoarthritis & Cartilage 2009;17:1151-5  57. Szebenyi B, Hollander AP, Dieppe P, Quilty B, Duddy G, Clarke S et al. Associations between pain, function, and radiographic features in osteoarthritis of the knee. Arthritis & Rheumatism 2006;54(1):230-5  58. Farrokhi S, Piva SR, Gil AB, Oddis CV, Brooks MM, Fitzgerald GK. Association of severity of coexisting patellofemoral disease with increased impairments and functional limitations in patients with knee osteoarthritis. Arthritis Care & Research 2013;65(4):544-51 59. Farrokhi S, Chen Y-F, Piva SR, Fitzgerald GK, Jeong J-H, Kwoh CK. The influence of knee pain location on symptoms, functional status, and knee-related quality of life in older adults with chronic knee pain: data from the Osteoarthritis Initiative. The Clinical Journal of Pain 2016;32(6):463-70  60. Iijima H, Fukutani N, Aoyama T, Fukumoto T, Uritani D, Kaneda E et al. Clinical impact of coexisting patellofemoral osteoarthritis in Japanese patients with medial knee osteoarthritis. Arthritis Care & Research 2016;68(4):493-501    210  61. Duncan R, Peat G, Thomas E, Wood L, Hay E, Croft P. How do pain and function vary with compartmental distribution and severity of radiographic knee osteoarthritis? Rheumatology 2008;47(11):1704-7  62. Culvenor AG, Lai CC, Gabbe BJ, Makdissi M, Collins NJ, Vicenzino B et al. Patellofemoral osteoarthritis is prevalent and associated with worse symptoms and function after hamstring tendon autograft ACL reconstruction. British Journal of Sports Medicine 2014;48(6):435-9 doi: http://dx.doi.org/10.1136/bjsports-2013-092975 63. Conaghan PG, Kloppenburg M, Schett G, Bijlsma JW, committee Eoah. Osteoarthritis research priorities: a report from a EULAR ad hoc expert committee. Annals of the Rheumatic Diseases 2014;73(8):1442-5 doi: http://dx.doi.org/10.1136/annrheumdis-2013-204660 64. Stefanik JJ, Guermazi A, Roemer FW, Peat G, Niu J, Segal NA et al. Changes in patellofemoral and tibiofemoral joint cartilage damage and bone marrow lesions over 7 years: the Multicenter Osteoarthritis Study. Osteoarthritis & Cartilage 2016;24(7):1160-6 doi: http://dx.doi.org/10.1016/j.joca.2016.01.981 65. Neumann DA. Kinesiology of the musculoskeletal system: foundations for physical rehabilitation: Mosby, 2002. 66. Wikipedia the free encyclopedia. Patella. Secondary Patella  2017. https://en.wikipedia.org/wiki/Patella 67. Tecklenburg K, Dejour D, Hoser C, Fink C. Bony and cartilaginous anatomy of the patellofemoral joint. Knee Surg Sports Traumatol Arthrosc 2006;14(3):235-40  68. Iranpour F, Merican AM, Amis AA, Cobb JP. The width: thickness ratio of the patella. Clinical Orthopaedics and Related Research 2008;466(5):1198-203    211  69. Magee DJ, Zachazewski JE, Quillen WS. Scientific foundations and principles of practice in musculoskeletal rehabilitation: Elsevier Health Sciences, 2007 70. Yamaguchi GT, Zajac FE. A planar model of the knee joint to characterize the knee extensor mechanism. Journal of Biomechanics 1989;22(1):1-10  71. Lee TQ, Morris G, Csintalan RP. The influence of tibial and femoral rotation on patellofemoral contact area and pressure. Journal of Orthopaedic & Sports Physical Therapy 2003;33(11):686-93  72. Reilly DT, Martens M. Experimental analysis of the quadriceps muscle force and patello-femoral joint reaction force for various activities. Acta Orthopaedica Scandinavica 1972;43(2):126-37  73. Cleather DJ, Goodwin JE, Bull AM. Hip and knee joint loading during vertical jumping and push jerking. Clinical Biomechanics 2013;28(1):98-103  74. D’Lima DD, Fregly BJ, Patil S, Steklov N, Colwell Jr CW. Knee joint forces: prediction, measurement, and significance. Proceedings of the Institution of Mechanical Engineers, Part H: Journal of Engineering in Medicine 2012;226(2):95-102  75. Makris EA, Hadidi P, Athanasiou KA. The knee meniscus: structure-function, pathophysiology, current repair techniques, and prospects for regeneration. Biomaterials 2011;32(30):7411-31 doi: 10.1016/j.biomaterials.2011.06.037 76. Connolly A, FitzPatrick D, Moulton J, Lee J, Lerner A. Tibiofemoral cartilage thickness distribution and its correlation with anthropometric variables. Proceedings of the Institution of Mechanical Engineers. Part H, Journal of Engineering in Medicine 2008;222(1):29-39 doi: 10.1243/09544119jeim306   212  77. Hungerford DS, Barry M. Biomechanics of the patellofemoral joint. Clinical Orthopaedics and Related Research 1979;144:9-15  78. Souza RB, Draper CE, Fredericson M, Powers CM. Femur rotation and patellofemoral joint kinematics: a weight-bearing magnetic resonance imaging analysis. Journal of Orthopaedic & Sports Physical Therapy 2010;40(5):277-85  79. Powers CM, Ward SR, Fredericson M, Guillet M, Shellock FG. Patellofemoral kinematics during weight-bearing and non-weight-bearing knee extension in persons with lateral subluxation of the patella: a preliminary study. Journal of Orthopaedic & Sports Physical Therapy 2003;33(11):677-85  80. Witvrouw E, Callaghan MJ, Stefanik JJ, Noehren B, Bazzett-Jones DM, Willson JD et al. Patellofemoral pain: consensus statement from the 3rd International Patellofemoral Pain Research Retreat held in Vancouver, September 2013. British Journal of Sports Medicine 2014;48(6):411-14  81. Besier TF, Gold GE, Delp SL, Fredericson M, Beaupré GS. The influence of femoral internal and external rotation on cartilage stresses within the patellofemoral joint. Journal of Orthopaedic Research 2008;26(12):1627-35  82. Wyndow N, Collins N, Vicenzino B, Tucker K, Crossley K. Is there a biomechanical link between patellofemoral pain and osteoarthritis? A narrative review. Sports Medicine 2016;46(12):1797-808 doi: 10.1007/s40279-016-0545-6 83. Felson DT. Osteoarthritis as a disease of mechanics. Osteoarthritis & Cartilage 2013;21(1):10-5 doi: http://dx.doi.org/10.1016/j.joca.2012.09.012 84. Loeser RF, Goldring SR, Scanzello CR, Goldring MB. Osteoarthritis: a disease of the joint as an organ. Arthritis & Rheumatism 2012;64(6):1697-707    213  85. Porcheret M. Osteoarthritis: a new approach to an old problem? Secondary Osteoarthritis: a new approach to an old problem? 2012. http://www.arthritisresearchuk.org/arthritis-information/arthritis-today-magazine/156-spring-2012/osteoarthritis.aspx 86. van der Kraan PM, van den Berg WB. Osteophytes: relevance and biology. Osteoarthritis & Cartilage 2007;15(3):237-44  87. Majumdar S, editor. Advances in MRI of the Knee for Osteoarthritis. Singapore: World Scientific Publishing Co. Pte. Ltd., 2010 88. Felson DT, Niu J, Guermazi A, Sack B, Aliabadi P. Defining radiographic incidence and progression of knee osteoarthritis: suggested modifications of the Kellgren and Lawrence scale. Annals of the Rheumatic Diseases 2011;70(11):1884-6 doi: http://dx.doi.org/10.1136/ard.2011.155119 89. McAlindon T, Zhang Y, Hannan M, Naimark A, Weissman B, Castelli W et al. Are risk factors for patellofemoral and tibiofemoral knee osteoarthritis different? Journal of Rheumatology 1996;23(2):332-7  90. Peterfy CG, Guermazi A, Zaim S, Tirman PFJ, Miaux Y, White D et al. Whole-Organ Magnetic Resonance Imaging Score (WORMS) of the knee in osteoarthritis. Osteoarthritis & Cartilage 2004;12(3):177-90  91. Pollard T, Gwilym S, Carr A. The assessment of early osteoarthritis. Bone & Joint Journal 2008;90(4):411-21  92. Larsson S. Aggrecan fragments as biomarkers in osteoarthritis. Lund University, Faculty of Medicine Doctoral Dissertation Series 2010;2010(55)    214  93. Wilson DR, McWalter EJ, Johnston JD. The measurement of joint mechanics and their role in osteoarthritis genesis and progression. Rheumatic Diseases Clinics of North America 2013;39:21-44  94. Cotofana S, Eckstein F, Wirth W, Souza RB, Li X, Wyman B et al. In vivo measures of cartilage deformation: patterns in healthy and osteoarthritic female knees using 3T MR imaging. Eur Radiol 2011;21(6):1127-35  95. Frobell RB. Change in cartilage thickness, posttraumatic bone marrow lesions, and joint fluid volumes after acute ACL disruption. J Bone Joint Surg Am 2011;93(12):1096-103  96. Bijlsma JW, Berenbaum F, Lafeber FP. Osteoarthritis: an update with relevance for clinical practice. The Lancet 2011;377(9783):2115-26  97. Froimson MI, Ratcliffe A, Gardner TR, Mow VC. Differences in patellofemoral joint cartilage material properties and their significance to the etiology of cartilage surface fibrillation. Osteoarthritis & Cartilage 1997;5(6):377-86  98. Teichtahl AJ, Wluka AE, Cicuttini FM. Frontal plane knee alignment is associated with a longitudinal reduction in patella cartilage volume in people with knee osteoarthritis. Osteoarthritis and Cartilage 2008;16(7):851-54  99. Gilbertson E. Development of periarticular osteophytes in experimentally induced osteoarthritis in the dog. A study using microradiographic, microangiographic, and fluorescent bone-labelling techniques. Annals of the Rheumatic Diseases 1975;34(1):12-25  100. Hunter DJ. Insights from imaging on the epidemiology and pathophysiology of osteoarthritis. Radiol Clin North Am 2009;47(4):539-51  101. Abrahim-Zadeh R, Yu JS, Resnick D. Central (interior) osteophytes of the distal femur. Imaging and pathologic findings. Invest Radiol 1994;29(11):1001-5    215  102. McCauley TR, Kornaat PR, Jee WH. Central osteophytes in the knee: prevalence and association with cartilage defects on MR imaging. American Journal of Roentgenology 2001;176(2):359-64 doi: https://dx.doi.org/10.2214/ajr.176.2.1760359 103. Boegård T, Rudling O, Petersson IF, Jonsson K. Correlation between radiographically diagnosed osteophytes and magnetic resonance detected cartilage defects in the tibiofemoral joint. Annals of the Rheumatic Diseases 1998;57(7):401-07  104. Taljanovic MS, Graham AR, Benjamin JB, Gmitro AF, Krupinski EA, Schwartz SA et al. Bone marrow edema pattern in advanced hip osteoarthritis: quantitative assessment with magnetic resonance imaging and correlation with clinical examination, radiographic findings, and histopathology. Skeletal Radiology 2008;37(5):423-31  105. Eriksen EF. Treatment of bone marrow lesions (bone marrow edema). BoneKEy reports 2015;4  106. Zanetti M, Bruder E, Romero J, Hodler J. Bone marrow edema pattern in osteoarthritic knees: correlation between MR imaging and histologic findings 1. Radiology 2000;215(3):835-40  107. Del Grande F, Farahani SJ, Carrino JA, Chhabra A. Bone marrow lesions: A systematic diagnostic approach. The Indian Journal of Radiology & Imaging 2014;24(3):279  108. Sofat N, Assi L, Kuttapitiya A, Boyde A, Ejindu V, Heron C. How do bone marrow lesions cause osteoarthritis pain? A structural and functional tissue based study. ACR/ARHP Annual Meeting 2016 https://qmro.qmul.ac.uk/xmlui/handle/123456789/18311 109. Crema MD, Roemer FW, Zhu Y, Marra MD, Niu J, Zhang Y et al. Subchondral cystlike lesions develop longitudinally in areas of bone marrow edema–like lesions in patients with   216  or at risk for knee osteoarthritis: detection with MR imaging—The MOST Study. Radiology 2010;256(3):855-62  110. Chen Y, Wang T, Guan M, Zhao W, Leung FKL, Pan H et al. Bone turnover and articular cartilage differences localized to subchondral cysts in knees with advanced osteoarthritis. Osteoarthritis & Cartilage 2015;23(12):2174-83 doi: https://dx.doi.org/10.1016/j.joca.2015.07.012 111. Tavares Júnior WC, Faria FMd, Figueiredo R, Matushita JPK, Silva LC, Kakehasi AM. Bone attrition: a cause of knee pain in osteoarthritis. Radiologia Brasileira 2012;45(5):273-78  112. Torres L, Dunlop D, Peterfy C, Guermazi A, Prasad P, Hayes KW et al. The relationship between specific tissue lesions and pain severity in persons with knee osteoarthritis. Osteoarthritis & Cartilage 2006;14(10):1033-40  113. Crema MD, Cibere J, Sayre EC, Roemer FW, Wong H, Thorne A et al. The relationship between subchondral sclerosis detected with MRI and cartilage loss in a cohort of subjects with knee pain: the knee osteoarthritis progression (KOAP) study. Osteoarthritis & Cartilage 2014;22(4):540-6 doi: https://dx.doi.org/10.1016/j.joca.2014.01.006 114. Kellgren JH, Lawrence JS. Radiological assessment of osteoarthrosis. Annals of the Rheumatic Diseases 1957;16(4):494-502  115. Cox LG, van Donkelaar CC, van Rietbergen B, Emans PJ, Ito K. Decreased bone tissue mineralization can partly explain subchondral sclerosis observed in osteoarthritis. Bone 2012;50(5):1152-61 doi: https://dx.doi.org/10.1016/j.bone.2012.01.024   217  116. Atukorala I, Kwoh CK, Guermazi A, Roemer FW, Boudreau RM, Hannon MJ et al. Synovitis in knee osteoarthritis: a precursor of disease? Annals of the Rheumatic Diseases 2014:annrheumdis-2014-205894  117. Guermazi A, Roemer FW, Hayashi D, Crema MD, Niu J, Zhang Y et al. Assessment of synovitis with contrast-enhanced MRI using a whole-joint semiquantitative scoring system in people with, or at high risk of, knee osteoarthritis: the MOST study. Annals of the Rheumatic Diseases 2011;70(5):805-11  118. Roemer FW, Jarraya M, Felson DT, Hayashi D, Crema MD, Loeuille D et al. Magnetic resonance imaging of Hoffa's fat pad and relevance for osteoarthritis research: a narrative review. Osteoarthritis & Cartilage 2016;24(3):383-97 doi: https://dx.doi.org/10.1016/j.joca.2015.09.018 119. Cowan SM, Hart HF, Warden SJ, Crossley KM. Infrapatellar fat pad volume is greater in individuals with patellofemoral joint osteoarthritis and associated with pain. Rheumatology International 2015;35(8):1439-42 doi: https://dx.doi.org/10.1007/s00296-015-3250-0 120. Chuckpaiwong B, Charles HC, Kraus VB, Guilak F, Nunley JA. Age‐associated increases in the size of the infrapatellar fat pad in knee osteoarthritis as measured by 3T MRI. Journal of Orthopaedic Research 2010;28(9):1149-54  121. Pan F, Han W, Wang X, Liu Z, Jin X, Antony B et al. A longitudinal study of the association between infrapatellar fat pad maximal area and changes in knee symptoms and structure in older adults. Annals of the Rheumatic Diseases 2014:annrheumdis-2013-205108    218  122. Klein-Wieringa I, Kloppenburg M, Bastiaansen-Jenniskens Y, Yusuf E, Kwekkeboom JC, El-Bannoudi H et al. The infrapatellar fat pad of patients with osteoarthritis has an inflammatory phenotype. Annals of the Rheumatic Diseases 2011;70(5):851-57  123. Hill CL, Seo GS, Gale D, Totterman S, Gale ME, Felson DT. Cruciate ligament integrity in osteoarthritis of the knee. Arthritis & Rheumatism 2005;52(3):794-99  124. Frallonardo P, Oliviero F, Peruzzo L, Tauro L, Scanu A, Galozzi P et al. Detection of calcium crystals in knee osteoarthritis synovial fluid: a comparison between polarized light and Scanning electron microscopy. Journal of Clinical Rheumatology 2016;22(7):369-71 doi: https://dx.doi.org/10.1097/RHU.0000000000000416 125. Abhishek A. Calcium pyrophosphate deposition disease: a review of epidemiologic findings. Curr Opin Rheumatol 2016;28(2):133-39  126. Altman RD, Hochberg M, Murphy WA, Jr., Wolfe F, Lequesne M. Atlas of individual radiographic features in osteoarthritis. Osteoarthritis & Cartilage 1995;3 Suppl A:3-70  127. Altman RD, Gold GE. Atlas of individual radiographic features in osteoarthritis, revised. Osteoarthritis & Cartilage 2007;15 Suppl A:A1-56  128. Zhang W, Doherty M, Peat G, Beirma-Zeinstra SMA, Arden NK, Bresnihan B et al. EULAR evidence based recommendations for the diagnosis of knee osteoarthritis. Annals of the Rheumatic Diseases 2009  129. Schiphof D, Boers M, Bierma-Zeinstra SM. Differences in descriptions of Kellgren and Lawrence grades of knee osteoarthritis. Annals of the Rheumatic Diseases 2008;67(7):1034-36  130. Litwic A, Edwards MH, Dennison EM, Cooper C. Epidemiology and burden of osteoarthritis. Br Med Bull 2013;105(1):185-99    219  131. Lynch JA, Buckland-Wright JC, Macfarlane DG. Precision of joint space width measurement in knee osteoarthritis from digital image analysis of high definition macroradiographs. Osteoarthritis & Cartilage 1993;1(4):209-18  132. Kellgren J, Jeffrey M, Ball J. The epidemiology of chronic rheumatism. Atlas of standard radiographs of arthritis. Oxford, UK  Blackwell Scientific Publications, 1963. 133. Culvenor AG, Engen CN, Oiestad BE, Engebretsen L, Risberg MA. Defining the presence of radiographic knee osteoarthritis: a comparison between the Kellgren and Lawrence system and OARSI atlas criteria. Knee Surg Sports Traumatol Arthrosc 2015;23(12):3532-9 doi: https://dx.doi.org/10.1007/s00167-014-3205-0 134. Englund M, Lohmander LS. Patellofemoral osteoarthritis coexistent with tibiofemoral osteoarthritis in a meniscectomy population. Annals of the Rheumatic Diseases 2005;64(12):1721-6 doi: https://dx.doi.org/10.1136/ard.2005.035568 135. Dillon CF, Rasch EK, Gu Q, Hirsch R. Prevalence of knee osteoarthritis in the United States: arthritis data from the Third National Health and Nutrition Examination Survey 1991-94. Journal of Rheumatology 2006;33(11):2271-9  136. Altman R, Asch E, Bloch D, Bole G, Borenstein D, Brandt K et al. Development of criteria for the classification and reporting of osteoarthritis. Classification of osteoarthritis of the knee. Diagnostic and Therapeutic Criteria Committee of the American Rheumatism Association. Arthritis Rheum 1986;29(8):1039-49  137. Jones A, Ledingham J, McAlindon T, Regan M, Hart D, MacMillan PJ et al. Radiographic assessment of patellofemoral osteoarthritis. Annals of the Rheumatic Diseases 1993;52(9):655-58    220  138. Chaisson CE, Gale DR, Gale E, Kazis L, Skinner K, Felson DT. Detecting radiographic knee osteoarthritis: what combination of views is optimal? Rheumatology 2000;39(11):1218-21  139. Cicuttini FM, Baker J, Hart DJ, Spector TD. Choosing the best method for radiological assessment of patellofemoral osteoarthritis. Annals of the Rheumatic Diseases 1996;55(2):134-36  140. Bhattacharya R, Kumar V, Safawi E, Finn P, Hui AC. The knee skyline radiograph: its usefulness in the diagnosis of patello-femoral osteoarthritis. International Orthopaedics 2007;31(2):247-52  141. Peat G, Thomas E, Duncan R, Wood L, Hay E, Croft P. Clinical classification criteria for knee osteoarthritis: performance in the general population and primary care. Annals of the Rheumatic Diseases 2006;65(10):1363-7  142. Peat G, Duncan RC, Wood LRJ, Thomas E, Muller S. Clinical features of symptomatic patellofemoral joint osteoarthritis. Arthritis Research & Therapy 2012;14(2):R63 doi: http://dx.doi.org/10.1186/ar3779 143. Schiphof D, van Middelkoop M, de Klerk BM, Oei EHJ, Hofman A, Koes BW et al. Crepitus is a first indication of patellofemoral osteoarthritis (and not of tibiofemoral osteoarthritis). Osteoarthritis & Cartilage 2014;22(5):631-8 doi: 10.1016/j.joca.2014.02.008 144. Stefanik JJ, Neogi T, Niu J, Roemer FW, Segal NA, Lewis CE et al. The diagnostic performance of anterior knee pain and activity-related pain in identifying knees with structural damage in the patellofemoral joint: the Multicenter Osteoarthritis Study. The Journal of Rheumatology 2014;41(8):1695-702    221  145. Felson DT. Challenges of identifying and treating patellofemoral osteoarthritis [Editorial] Br J Sports Med, 2016;50(14):832-33. http://dx.doi.org/10.1136/bjsports-2016-096294 146. Crossley K, Marino G, Macilquham M, Schache A, Hinman R. Can patellar tape reduce the patellar malalignment and pain associated with patellofemoral osteoarthritis? Arthritis & Rheumatism 2009;61(12):1719-25  147. Crossley KM, Vicenzino B, Lentzos J, Schache AG, Pandy MG, Ozturk H et al. Exercise, education, manual-therapy and taping compared to education for patellofemoral osteoarthritis: A blinded, randomised clinical trial. Osteoarthritis & Cartilage 2015;23(9):1457-64  148. Fok LA, Schache AG, Crossley KM, Lin YC, Pandy MG. Patellofemoral joint loading during stair ambulation in people with patellofemoral osteoarthritis. Arthritis & Rheumatism 2013;65(8):2059-69  149. Shapiro LM, McWalter EJ, Son MS, Levenston M, Hargreaves BA, Gold GE. Mechanisms of osteoarthritis in the knee: MR imaging appearance. Journal of Magnetic Resonance Imaging 2014;39(6):1346-56  150. International Atomic Energy Agency. X rays: what patients need to know. https://rpop.iaea.org/RPOP/RPoP/Content/Documents/Whitepapers/leaflet-x-rays.pdf. 151. Hart DJ, Spector TD. Kellgren & Lawrence grade 1 osteophytes in the knee - doubtful or definite? Osteoarthritis & Cartilage 2002;11(2):149-50  152. Hunter DJ, Niu J, Felson DT, Harvey WF, Gross KD, McCree P et al. Knee alignment does not predict incident osteoarthritis: the Framingham osteoarthritis study. Arthritis & Rheumatism 2007;56(4):1212-8    222  153. Mazzuca SA, Brandt KD, Katz BP, Ding Y, Lane KA, Buckwalter KA. Risk factors for progression of tibiofemoral osteoarthritis: an analysis based on fluoroscopically standardised knee radiography. Ann Rheum Dis 2006;65:515-9  154. Kwoh CK. Epidemiology of osteoarthritis. The Epidemiology of Aging: Springer, 2012:523-36. 155. Hawker GA, Badley EM, Croxford R, Coyte CP, Glazier RH, Guan J et al. A population-based nested case-control study of the costs of hip and knee replacement surgery. Med Care 2009;47(7):732-41  156. Link TM. MR imaging in osteoarthritis: hardware, coils, and sequences. Radiol Clin North Am 2009;47(4):617-32  157. Hunter DJ, Guermazi A, Lo GH, Grainger AJ, Conaghan PG, Boudreau RM et al. Evolution of semi-quantitative whole joint assessment of knee OA: MOAKS (MRI Osteoarthritis Knee Score). Osteoarthritis & Cartilage 2011;19(8):990-1002  158. Hunter DJ, Lo GH, Gale D, Grainger AJ, Guermazi A, Conaghan PG. The reliability of a new scoring system for knee osteoarthritis MRI and the validity of bone marrow lesion assessment: BLOKS (Boston–Leeds Osteoarthritis Knee Score). Annals of the Rheumatic Diseases 2008;67(2):206-11  159. van der Heijden RA, de Kanter JL, Bierma-Zeinstra SM, Verhaar AJN, van Veldhoven PLJ, Krestin GP et al. Structural abnormalities on magnetic resonance imaging in patients with patellofemoral pain a cross-sectional case-control study. The American Journal of Sports Medicine 2016;44(9):2339-46   223  160. Sharma L, Chmiel JS, Almagor O, Dunlop D, Guermazi A, Bathon JM et al. Significance of preradiographic magnetic resonance imaging lesions in persons at increased risk of knee osteoarthritis. Arthritis Rheumatol 2014;66(7):1811-19  161. Yusuf E, Kortekaas MC, Watt I, Huizinga TW, Kloppenburg M. Do knee abnormalities visualised on MRI explain knee pain in knee osteoarthritis? A systematic review. Annals of the Rheumatic Diseases 2011;70(1):60-67  162. Guermazi A, Niu J, Hayashi D, Roemer FW, Englund M, Neogi T et al. Prevalence of abnormalities in knees detected by MRI in adults without knee osteoarthritis: population based observational study (Framingham Osteoarthritis Study). BMJ 2012;345:e5339 doi: http://dx.doi.org/10.1136/bmj.e5339 163. Hunter DJ, Arden N, Conaghan PG, Eckstein F, Gold G, Grainger A et al. Definition of osteoarthritis on MRI: results of a Delphi exercise. Osteoarthritis & Cartilage 2011;19(8):963-9  164. Amin S, LaValley MP, Guermazi A, Grigoryan M, Hunter DJ, Clancy M et al. The relationship between cartilage loss on magnetic resonance imaging and radiographic progression in men and women with knee osteoarthritis. Arthritis & Rheumatism 2005;52(10):3152-59  165. Gudbergsen H, Lohmander L, Jones G, Christensen R, Bartels EM, Danneskiold-Samsoe B et al. Correlations between radiographic assessments and MRI features of knee osteoarthritis–a cross-sectional study. Osteoarthritis & Cartilage 2013;21(4):535-43  166. Menashe L, Hirko K, Losina E, Kloppenburg M, Zhang W, Li L et al. The diagnostic performance of MRI in osteoarthritis: a systematic review and meta-analysis. Osteoarthritis & Cartilage 2012;20(1):13-21    224  167. Quatman CE, Hettrich CM, Schmitt LC, Spindler KP. The clinical utility and diagnostic performance of magnetic resonance imaging for identification of early and advanced knee osteoarthritis: a systematic review. The American Journal of Sports Medicine 2011;39(7):1557-68  168. Pihlajamäki HK, Kuikka P-I, Leppänen V-V, Kiuru MJ, Mattila VM. Reliability of clinical findings and magnetic resonance imaging for the diagnosis of chondromalacia patellae. The Journal of Bone & Joint Surgery 2010;92(4):927-34  169. Cook C, Mabry L, Reiman MP, Hegedus EJ. Best tests/clinical findings for screening and diagnosis of patellofemoral pain syndrome: a systematic review. Physiotherapy 2012;98(2):93-100  170. van Oudenaarde K, Jobke B, Oostveen A, Marijnissen ACA, Wolterbeek R, Wesseling J et al. Predictive value of MRI features for development of radiographic osteoarthritis in a cohort of participants with pre-radiographic knee osteoarthritis—the CHECK study. Rheumatology 2017;56(1):113-20  171. Cicuttini F, Spector T, Baker J. Risk factors for osteoarthritis in the tibiofemoral and patellofemoral joints of the knee. The Journal of Rheumatology 1997;24(6):1164-67  172. Cooper C, McAlindon T, Snow S, Vines K, Young P, Kirwan J et al. Mechanical and constitutional risk factors for symptomatic knee osteoarthritis: differences between medial tibiofemoral and patellofemoral disease. The Journal of Rheumatology 1994;21(2):307-13  173. Srikanth VK, Fryer JL, Zhai G, Winzenberg TM, Hosmer D, Jones G. A meta-analysis of sex differences prevalence, incidence and severity of osteoarthritis. Osteoarthritis & Cartilage 2005;13(9):769-81    225  174. Silverwood V, Blagojevic-Bucknall M, Jinks C, Jordan JL, Protheroe J, Jordan KP. Current evidence on risk factors for knee osteoarthritis in older adults: a systematic review and meta-analysis. Osteoarthritis & Cartilage 2015;23(4):507-15 doi: http://dx.doi.org/10.1016/j.joca.2014.11.019 175. Blagojevic M, Jinks C, Jeffery A, Jordan K. Risk factors for onset of osteoarthritis of the knee in older adults: a systematic review and meta-analysis. Osteoarthritis & Cartilage 2010;18(1):24-33  176. Hart DJ, Doyle DV, Spector TD. Incidence and risk factors for radiographic knee osteoarthritis in middle-aged women. Arthritis Rheum 1999;42(1):17-24  177. Richmond SA, Fukuchi RK, Ezzat A, Schneider K, Schneider G, Emery CA. Are joint injury, sport activity, physical activity, obesity, or occupational activities predictors for osteoarthritis? A systematic review. Journal of Orthopaedic & Sports Physical Therapy 2013;43(8):515-B19  178. Ratzlaff CR, Koehoorn M, Cibere J, Kopec JA. Is lifelong knee joint force from work, home, and sport related to knee osteoarthritis? Int J Rheumatol 2012;2012:584193 doi: https://dx.doi.org/10.1155/2012/584193 179. Ratzlaff CR. Good news, bad news: sports matter but occupational and household activity really matter - sport and recreation unlikely to be a panacea for public health. British Journal of Sports Medicine 2012;46(10):699-701 doi: https://dx.doi.org/10.1136/bjsports-2011-090800 180. Øiestad B, Juhl C, Eitzen I, Thorlund J. Knee extensor muscle weakness is a risk factor for development of knee osteoarthritis. A systematic review and meta-analysis. Osteoarthritis & Cartilage 2015;23(2):171-77    226  181. Segal NA, Glass NA, Torner J, Yang M, Felson DT, Sharma L et al. Quadriceps weakness predicts risk for knee joint space narrowing in women in the MOST cohort. Osteoarthritis & Cartilage 2010;18(6):769-75  182. Felson DT, Niu J, Gross KD, Englund M, Sharma L, Cooke TDV et al. Valgus malalignment is a risk factor for lateral knee osteoarthritis incidence and progression: findings from the Multicenter Osteoarthritis Study and the Osteoarthritis Initiative. Arthritis & Rheumatism 2013;65(2):355-62 doi: http://dx.doi.org/10.1002/art.37726 183. Sharma L, Song J, Dunlop D, Felson D, Lewis CE, Segal N et al. Varus and valgus alignment and incident and progressive knee osteoarthritis. Annals of the Rheumatic Diseases 2010;69(11):1940-5 doi: http://dx.doi.org/10.1136/ard.2010.129742 184. Chapple CM, Nicholson H, Baxter GD, Abbott JH. Patient characteristics that predict progression of knee osteoarthritis: a systematic review of prognostic studies. Arthritis Care & Research 2011;63(8):1115-25  185. Bastick AN, Belo JN, Runhaar J, Bierma-Zeinstra SMA. What are the prognostic factors for radiographic progression of knee osteoarthritis? A meta-analysis. Clinical Orthopaedics & Related Research 2015;473(9):2969-89 doi: http://dx.doi.org/10.1007/s11999-015-4349-z 186. Tanamas S, Hanna F, Cicuttini F, Wluka A, Berry P, Urquhart D. Does knee malalignment increase the risk of development and progression of knee osteoarthritis? A systematic review. Arthritis & Rheumatism 2009;61(4):459-67  187. Messier SP. Diet and exercise for obese adults with knee osteoarthritis. Clinics in Geriatric Medicine 2010;26(3):461-77    227  188. Griffin TM, Guilak F. The role of mechanical loading in the onset and progression of osteoarthritis. Exercise and Sport Sciences Reviews 2005;33(4):195-200  189. Messier SP, Gutekunst DJ, Davis C, DeVita P. Weight loss reduces knee‐joint loads in overweight and obese older adults with knee osteoarthritis. Arthritis & Rheumatism 2005;52(7):2026-32  190. Ukachukwu V, Duncan R, Belcher J, Marshall M, Stefanik J, Crossley K et al. Clinical significance of medial vs. lateral compartment patellofemoral osteoarthritis: cross‐sectional analyses in an adult population with knee pain. Arthritis Care & Research 2016p;69(7):943-51  191. Culvenor AG, Cook JL, Collins NJ, Crossley KM. Is patellofemoral joint osteoarthritis an under-recognised outcome of anterior cruciate ligament reconstruction? A narrative literature review. British Journal of Sports Medicine 2013;47(2):66-70 doi: http://dx.doi.org/10.1136/bjsports-2012-091490 192. Stefanik JJ, Guermazi A, Zhu Y, Zumwalt AC, Gross KD, Clancy M et al. Quadriceps weakness, patella alta, and structural features of patellofemoral osteoarthritis. Arthritis Care & Research 2011;63(10):1391-7 doi: http://dx.doi.org/10.1002/acr.20528 193. Lufler RS, Stefanik JJ, Niu J, Sawyer FK, Hoagland TM, Gross KD. The association of forefoot varus deformity with patellofemoral cartilage damage in older adult cadavers. Arthritis Care & Research (Hoboken) 2017;300(6):1032-38 doi: https://dx.doi.org/10.1002/ar.23524 194. Cicuttini F, Wluka A, Wang Y, Stuckey S. The determinants of change in patella cartilage volume in osteoarthritic knees. The Journal of Rheumatology 2002;29(12):2615-19    228  195. Roemer FW, Kwoh CK, Hannon MJ, Green SM, Jakicic JM, Broudrea R et al. Risk factors for magnetic resonance imaging–detected patellofemoral and tibiofemoral cartilage loss during a six‐month period: The Joints On Glucosamine study. Arthritis & Rheumatism 2012;64(6):1888-98  196. Amin S, Baker K, Niu J, Clancy M, Goggins J, Guermazi A et al. Quadriceps strength and the risk of cartilage loss and symptom progression in knee osteoarthritis. Arthritis & Rheumatism 2009;60(1):189-98  197. Cahue S, Dunlop D, Hayes K, Song J, Torres L, Sharma L. Varus-valgus alignment in the progression of patellofemoral osteoarthritis. Arthritis & Rheumatism 2004;50(7):2184-90  198. Chakravarty EF, Hubert HB, Lingala VB, Zatarain E, Fries JF. Long distance running and knee osteoarthritis. A prospective study. American Journal of Preventive Medicine 2008;35(2):133–8  199. Murphy L, Helmick CG. The impact of osteoarthritis in the United States: a population-health perspective: A population-based review of the fourth most common cause of hospitalization in US adults. Orthopaedic Nursing 2012;31(2):85-91  200. Ezzat AM. Cumulative occupational physical load as risk factor for knee osteoarthritis. Arthritis Rheum; 2012; 64(10):S473. 201. Baker KR, Xu L, Zhang Y, Nevitt M, Niu J, Aliabadi P et al. Quadriceps weakness and its relationship to tibiofemoral and patellofemoral knee osteoarthritis in Chinese: the Beijing osteoarthritis study. Arthritis & Rheumatism 2004;50(6):1815-21  202. Wang J, Antony B, Zhu Z, Han W, Pan F, Wang X et al. Association of patellar bone marrow lesions with knee pain, patellar cartilage defect and patellar cartilage volume loss in older adults: a cohort study. Osteoarthritis & Cartilage 2015;23(8):1330-36    229  203. Stefanik JJ, Wang K, Gross KD, Roemer FW, Lynch JA, Segal N et al. The association of demographic, modifiable, structural and biomechanical risk factors with medial and lateral patellofemoral joint structural damage on MRI: The Multicenter Osteoarthritis Study. Arthritis Rheum 2012;64:S474 doi: http://dx.doi.org/10.1002/art.37735 204. Wikipedia. Phenotype. 21 Feb 2017 2017. https://en.wikipedia.org/wiki/Phenotype. 205. Driban JB, Sitler MR, Barbe MF, Balasubramanian E. Is osteoarthritis a heterogeneous disease that can be stratified into subsets? Clin Rheumatol 2010;29(2):123-31 doi: http://dx.doi.org/10.1007/s10067-009-1301-1 206. Bierma-Zeinstra S, Verhagen AP. Osteoarthritis subpopulations and implications for clinical trial design. Arthritis Res Ther 2011;13:213  207. Crema MD, Guermazi A, Sayre EC, Roemer FW, Wong H, Thorne A et al. The association of magnetic resonance imaging (MRI)-detected structural pathology of the knee with crepitus in a population-based cohort with knee pain: the MoDEKO study. Osteoarthritis & Cartilage 2011;19(12):1429-32 doi: 10.1016/j.joca.2011.09.003 208. Crossley KM, Dorn TW, Ozturk H, van den Noort J, Schache AG, Pandy MG. Altered hip muscle forces during gait in people with patellofemoral osteoarthritis. Osteoarthritis & Cartilage 2012;20(11):1243-9 doi: http://dx.doi.org/10.1016/j.joca.2012.07.011 209. Pohl MB, Patel C, Wiley JP, Ferber R. Gait biomechanics and hip muscular strength in patients with patellofemoral osteoarthritis. Gait & Posture 2013;37(3):440-4 doi: http://dx.doi.org/10.1016/j.gaitpost.2012.08.01 210. Teng HL, MacLeod TD, Link TM, Majumdar S, Souza RB. Higher knee flexion moment during the second half of the stance phase of gait is associated with the progression of osteoarthritis of the patellofemoral joint on magnetic resonance imaging. Journal of   230  Orthopaedic & Sports Physical Therapy 2015;45(9):656-64 doi: https://dx.doi.org/10.2519/jospt.2015.5859 211. Farrokhi S, O’Connell M, Fitzgerald GK. Altered gait biomechanics and increased knee-specific impairments in patients with coexisting tibiofemoral and patellofemoral osteoarthritis. Gait & Posture 2015;41(1):81-85  212. Thoma L, McNally M, Chaudhari A, Best TM, Flanigan DC, Siston RA et al. Differential knee joint loading patterns during gait for individuals with tibiofemoral and patellofemoral articular cartilage defects in the knee. Osteoarthritis & Cartilage 2017;25(7):1046-54  213. Hart HF, Ackland DC, Pandy MG, Crossley KM. Quadriceps volumes are reduced in people with patellofemoral joint osteoarthritis. Osteoarthritis & Cartilage 2012;20(8):863-8 doi: http://dx.doi.org/10.1016/j.joca.2012.04.009 214. Callaghan MJ, Parkes MJ, Hutchinson CE, Gait AD, Forsythe LM, Marjanovic EJ et al. A randomised trial of a brace for patellofemoral osteoarthritis targeting knee pain and bone marrow lesions. Annals of the Rheumatic Diseases 2015;74:1164-70 215. Quilty B, Tucker M, Campbell R, Dieppe P. Physiotherapy, including quadriceps exercises and patellar taping, for knee osteoarthritis with predominant patello-femoral joint involvement: randomized controlled trial. Journal of Rheumatology 2003;30(6):1311-7  216. Collins NJ, Vicenzino B, Van Der Heijden RA, Van Middelkoop M. Pain during prolonged sitting is a common problem in persons with patellofemoral pain. Journal of Orthopaedic & Sports Physical Therapy 2016;46(8):658-63  217. Crossley KM, Callaghan MJ, van Linschoten R. Patellofemoral pain. bmj 2015;351:h3939    231  218. Nunes GS, Stapait EL, Kirsten MH, de Noronha M, Santos GM. Clinical test for diagnosis of patellofemoral pain syndrome: Systematic review with meta-analysis. Physical Therapy in Sport 2013;14(1):54-59  219. Rathleff MS, Roos EM, Olesen JL, Rasmussen S. Early intervention for adolescents with patellofemoral pain syndrome--a pragmatic cluster randomised controlled trial. BMC Musculoskeletal Disorders 2012;13:9  220. Hinman RS, Lentzos J, Vicenzino B, Crossley KM. Is patellofemoral osteoarthritis common in middle-aged people with chronic patellofemoral pain? Arthritis Care & Research 2014;66(8):1252-7 doi: http://dx.doi.org/10.1002/acr.22274 221. van Middelkoop M, van Linschoten R, Berger MY, Koes BW, Bierma-Zeinstra SM. Knee complaints seen in general practice: active sport participants versus non-sport participants. BMC Musculoskeletal Disorders 2008;9(1):36  222. Lankhorst NE, Bierma-Zeinstra SM, van Middelkoop M. Risk factors for patellofemoral pain syndrome: a systematic review. Journal of Orthopaedic & Sports Physical Therapy 2012;42(2):81-A12  223. Lankhorst NE, Bierma-Zeinstra SM, van Middelkoop M. Factors associated with patellofemoral pain syndrome: a systematic review. British Journal of Sports Medicine 2013;47(4):193-206 doi: http://dx.doi.org/10.1136/bjsports-2011-090369 224. Petersen W, Ellermann A, Gösele-Koppenburg A, Best R, Rembitzki IV, Bruggemann GD et al. Patellofemoral pain syndrome. Knee Surg Sports Traumatol Arthrosc 2014;22(10):2264-74  225. Graci V, Salsich GB. Trunk and lower extremity segment kinematics and their relationship to pain following movement instruction during a single-leg squat in females with dynamic   232  knee valgus and patellofemoral pain. Journal of Science and Medicine in Sport 2015;18(3):343-47  226. Giles LS, Webster KE, McClelland JA, Cook J. Does quadriceps atrophy exist in individuals with patellofemoral pain? A systematic literature review with meta-analysis. Journal of Orthopaedic & Sports Physical Therapy 2013;43(11):766-76  227. Prins MR, Van Der Wurff P. Females with patellofemoral pain syndrome have weak hip muscles: a systematic review. Australian Journal of Physiotherapy 2009;55(1):9-15  228. Barton CJ, Lack S, Malliaras P, Morrissey D. Gluteal muscle activity and patellofemoral pain syndrome: a systematic review. British Journal of Sports Medicine 2013;47(4):207-14 doi: http://dx.doi.org/10.1136/bjsports-2012-090953 229. Rathleff MS, Rathleff CR, Crossley KM, Barton CJ. Is hip strength a risk factor for patellofemoral pain? A systematic review and meta-analysis. British Journal of Sports Medicine 2014;48(14):1088 doi: http://dx.doi.org/10.1136/bjsports-2013-093305 230. Crossley KM. Is patellofemoral osteoarthritis a common sequela of patellofemoral pain? British Journal of Sports Medicine 2014;48(6):409-10 doi: http://dx.doi.org/10.1136/bjsports-2014-093445 231. Thomas MJ, Wood L, Selfe J, Peat G. Anterior knee pain in younger adults as a precursor to subsequent patellofemoral osteoarthritis: a systematic review. BMC Musculoskeletal Disorders 2010;11(1):201  232. Nimon G, Murray D, Sandow M, Goodfellow J. Natural history of anterior knee pain: a 14-to 20-year follow-up of nonoperative management. Journal of Pediatric Orthopaedics 1998;18(1):118-22    233  233. Utting M, Davies G, Newman J. Is anterior knee pain a predisposing factor to patellofemoral osteoarthritis? The Knee 2005;12(5):362-65  234. Collins N, Vicenzino B, Macri E, Crossley K. Prevalence and factors associated with radiographic PFJ OA in young to middle-aged adults with chronic patellofemoral pain. Journal of Science and Medicine in Sport 2015;19:e85  235. Hunter DJ. Lower extremity osteoarthritis management needs a paradigm shift. British Journal of Sports Medicine 2011;45(4):283-8 doi: http://dx.doi.org/10.1136/bjsm.2010.081117 236. Moyer RF, Ratneswaran A, Beier F, Birmingham TB. Osteoarthritis year in review 2014: mechanics--basic and clinical studies in osteoarthritis. Osteoarthritis & Cartilage 2014;22(12):1989-2002 doi: http://dx.doi.org/10.1016/j.joca.2014.06.034 237. Horisberger M, Fortuna R, Valderrabano V, Herzog W. Long-term repetitive mechanical loading of the knee joint by in vivo muscle stimulation accelerates cartilage degeneration and increases chondrocyte death in a rabbit model. Clinical Biomechanics 2013;28(5):536-43  238. Heino BJ, Powers CM. Patellofemoral stress during walking in persons with and without patellofemoral pain. Medicine and science in sports and exercise 2002;34(10):1582-93  239. Mills K, Hunter DJ. Patellofemoral joint osteoarthritis: an individualised pathomechanical approach to management. Best Practice & Research Clinical Rheumatology 2014;28(1):73-91  240. van Haver A, De Roo K, De Beule M, Labey L, de Baets P, Dejour D, Claessens T et al. The effect of trochlear dysplasia on patellofemoral biomechanics: a cadaveric study with   234  simulated trochlear deformities. American Journal of Sports Medicine 2015;43(6):1354-61 doi: http://dx.doi.org/10.1177/0363546515572143 241. Ward SR, Terk MR, Powers CM. Patella alta: association with patellofemoral alignment and changes in contact area during weight-bearing. Journal of Bone & Joint Surgery - American Volume 2007;89(8):1749-55  242. Farrokhi S, Keyak J, Powers C. Individuals with patellofemoral pain exhibit greater patellofemoral joint stress: a finite element analysis study. Osteoarthritis & Cartilage 2011;19(3):287-94  243. McConnell J. The physical therapist's approach to patellofemoral disorders. Clinics in Sports Medicine 2002;21(3):363-87  244. Souza RB, Powers CM. Predictors of hip internal rotation during running: an evaluation of hip strength and femoral structure in women with and without patellofemoral pain. Am J Sports Med 2009;37(3):579-87 doi: https://dx.doi.org/10.1177/0363546508326711 245. Bagwell JJ, Fukuda TY, Powers CM. Sagittal plane pelvis motion influences transverse plane motion of the femur: kinematic coupling at the hip joint. Gait & Posture 2016;43:120-4 doi: https://dx.doi.org/10.1016/j.gaitpost.2015.09.010 246. Hoglund LT, Hillstrom HJ, Barr-Gillespie AE, Lockard MA, Barbe MF, Song J. Frontal plane knee and hip kinematics during sit-to-stand and proximal lower extremity strength in persons with patellofemoral osteoarthritis: a pilot study. Journal of Applied Biomechanics 2014;30(1):82-94  247. Elias JJ, Kirkpatrick MS, Saranathan A, Mani S, Smith LG, Tanaka MJ. Hamstrings loading contributes to lateral patellofemoral malalignment and elevated cartilage pressures:   235  an in vitro study. Clinical Biomechanics 2011;26(8):841-6 doi: https://dx.doi.org/10.1016/j.clinbiomech.2011.03.016 248. Farrokhi S, Colletti PM, Powers CM. Differences in patellar cartilage thickness, transverse relaxation time, and deformational behavior a comparison of young women with and without patellofemoral pain. The American Journal of Sports Medicine 2011;39(2):384-91  249. Drew BT, Redmond A, Smith T, Penny F, Conaghan P. Which patellofemoral joint imaging features are associated with patellofemoral pain? Systematic review and meta-analysis. Osteoarthritis & Cartilage 2016;24(2):224-36  250. Insall J, Salvati E. Patella position in the normal knee joint Radiology 1971;101(1):101-04  251. Otsuki S, Nakajima M, Okamoto Y, Oda S, Hoshiyama Y, Iida G et al. Correlation between varus knee malalignment and patellofemoral osteoarthritis. Knee Surg Sports Traumatol Arthrosc 2014;24:1-6  252. Fregly BJ, Rahman HA, Banks SA. Theoretical accuracy of model-based shape matching for measuring natural knee kinematics with single-plane fluoroscopy. J Biomech Eng 2005;127(4):692-99  253. Saggin PRF, Saggin JI, Dejour D. Imaging in patellofemoral instability: an abnormality-based approach. Sports Medicine and Arthroscopy Review 2012;20(3):145-51  254. Patel VV, Hall K, Ries M, Lindsey C, Ozhinsky E, Lu Y et al. Magnetic resonance imaging of patellofemoral kinematics with weight-bearing. J Bone Joint Surg Am 2003;85(12):2419-24  255. McWalter EJ, Hunter DJ, Wilson DR. The effect of load magnitude on three-dimensional patellar kinematics in vivo. Journal of Biomechanics 2010;43(10):1890-7    236  256. d'Entremont AG, Nordmeyer-Massner JA, Bos C, Wilson DR, Pruessmann KP. Do dynamic-based MR knee kinematics methods produce the same results as static methods?.[Erratum appears in Magn Reson Med. 2013 Jun;69(6):1813]. Magn Reson Med 2013;69(6):1634-44 doi: http://dx.doi.org/10.1002/mrm.24425 257. Hunter DJ, Zhang YQ, Niu JB, Felson DT, Kwoh K, Newman A et al. Patella malalignment, pain and patellofemoral progression: the Health ABC Study. Osteoarthritis & Cartilage 2007;15(10):1120-7  258. Stefanik JJ, Roemer FW, Zumwalt AC, Zhu Y, Gross KD, Lynch JA et al. Association between measures of trochlear morphology and structural features of patellofemoral joint osteoarthritis on MRI: the MOST study. Journal of Orthopaedic Research 2012;30(1):1-8  259. Kalichman L, Zhang Y, Niu J, Goggins J, Gale D, Felson DT et al. The association between patellar alignment and patellofemoral joint osteoarthritis features—an MRI study. Rheumatology 2007;46(8):1303-08  260. Davies-Tuck M, Teichtahl AJ, Wluka AE, Wang Y, Urquhart DM, Cui J et al. Femoral sulcus angle and increased patella facet cartilage volume in an osteoarthritic population. Osteoarthritis & Cartilage 2008;16(1):131-5  261. Jungmann PM, Tham SC, Liebl H, Nevitt MC, McCulloch CE, Lynch J et al. Association of trochlear dysplasia with degenerative abnormalities in the knee: data from the Osteoarthritis Initiative. Skeletal Radiology 2013;42(10):1383-92 doi: http://dx.doi.org/10.1007/s00256-013-1664-x 262. National Institutes for Health Research. PROSPERO International Prospective Register of Systematic Reviews. http://www.crd.york.ac.uk/prospero    237  263. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gotzsche PC, Ioannidis JPA et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. Journal of Clinical Epidemiology 2009;62(10):e1-34 doi: http://dx.doi.org/10.1016/j.jclinepi.2009.06.006 264. Downs SH, Black N. The feasibility of creating a checklist for the assessment of the methodological quality both of randomised and non-randomised studies of health care interventions. Journal of Epidemiology & Community Health 1998;52(6):377-84  265. Langham S, Langham J, Goertz HP, Ratcliffe M. Large-scale, prospective, observational studies in patients with psoriasis and psoriatic arthritis: A systematic and critical review. BMC Medical Research Methodology 2011;11:32 doi: http://dx.doi.org/10.1186/1471-2288-11-32 266. Mills K, Hunt MA, Ferber R. Biomechanical deviations during level walking associated with knee osteoarthritis: a systematic review and meta-analysis. Arthritis care & research 2013;65(10):1643-65 doi: http://dx.doi.org/10.1002/acr.22015 267. Deeks JJ, Dinnes J, D'Amico R, Sowden AJ, Sakarovitch C, Song F et al. Evaluating non-randomised intervention studies. Health Technol Assess 2003;7(27):iii-x, 1-173  268. Sanderson S, Tatt ID, Higgins JP. Tools for assessing quality and susceptibility to bias in observational studies in epidemiology: a systematic review and annotated bibliography. Int J Epidemiol 2007;36(3):666-76  269. von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, Vandenbrouke JP et al. The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet 2007;370(9596):1453-7    238  270. Slavin RE. Best evidence synthesis: an intelligent alternative to meta-analysis. Journal of Clinical Epidemiology 1995;48(1):9-18  271. von Eisenhart-Rothe R, Graichen H, Hudelmaier M, Vogl T, Sharma L, Eckstein F. Femorotibial and patellar cartilage loss in patients prior to total knee arthroplasty, heterogeneity, and correlation with alignment of the knee. Annals of the Rheumatic Diseases 2006;65(1):69-73  272. Harrison MM, Cooke TD, Fisher SB, Griffin MP. Patterns of knee arthrosis and patellar subluxation. Clinical Orthopaedics & Related Research 1994(309):56-63  273. Iwano T, Kurosawa H, Tokuyama H, Hoshikawa Y. Roentgenographic and clinical findings of patellofemoral osteoarthrosis. With special reference to its relationship to femorotibial osteoarthrosis and etiologic factors. Clinical Orthopaedics & Related Research 1990(252):190-7  274. Rose CP, Cockshott WP. Anterior femoral erosion and patello-femoral osteoarthritis. J Can Assoc Radiol 1982;33(1):32-4  275. Stefanik J, Zhu Y, Zumwalt A, Gross KD, Clancy M, Lynch JA et al. Association between patella alta and the prevalence and worsening of structural features of patellofemoral joint osteoarthritis: the multicenter osteoarthritis study. Arthritis Care & Research 2010;62(9):1258-65  276. Kalichman L, Zhang Y, Niu J, Goggins J, Gale D, Zhu Y et al. The association between patellar alignment on magnetic resonance imaging and radiographic manifestations of knee osteoarthritis. Arthritis Research & Therapy 2007;9(2):R26  277. Mofidi A, Veravalli K, Jinnah RH, Poehling GG. Association and impact of patellofemoral dysplasia on patellofemoral arthropathy and arthroplasty. The Knee 2014;21(2):509-13    239  278. Elahi S, Cahue S, Felson DT, Engelman L, Sharma L. The association between varus-valgus alignment and patellofemoral osteoarthritis. Arthritis Rheum 2000;43(8):1874-80 doi: 10.1002/1529-0131(200008)43:8<1874::aid-anr25>3.0.co;2-2 279. Im GI, Kim MK, Lee SH. Relationship between knee alignment and radiographic markers of osteoarthritis: a cross‐sectional study from a Korean population. Int J Rheum Dis 2013  280. Tsavalas N, Katonis P, Karantanas AH. Knee joint anterior malalignment and patellofemoral osteoarthritis: an MRI study. Eur Radiol 2012;22(2):418-28 doi: http://dx.doi.org/10.1007/s00330-011-2275-3 281. Ali SA, Helmer R, Terk MR. Analysis of the patellofemoral region on MRI: Association of abnormal trochlear morphology with severe cartilage defects. American Journal of Roentgenology 2010;194(3):721-27 doi: http://dx.doi.org/10.2214/AJR.09.3008 282. Hodges PW, Smeets RJ. Interaction between pain, movement, and physical activity: short-term benefits, long-term consequences, and targets for treatment. The Clinical Journal of Pain 2015;31(2):97-107  283. Draper CE, Besier TF, Fredericson M, Santos JM, Beaupre GS, Delp SL et al. Differences in patellofemoral kinematics between weight-bearing and non-weight-bearing conditions in patients with patellofemoral pain. Journal of Orthopaedic Research 2011;29:312-7  284. Vittinghoff E, Glidden DV, Shiboski SC, McCulloch CE. Regression methods in biostatistics: linear, logistic, survival, and repeated measures models. New York: Springer Science+Business Media, Inc., 2005. 285. Greenland S. Dose-response and trend analysis in epidemiology: alternatives to categorical analysis. Epidemiology 1995:356-65    240  286. Aglietti P, Insall JN, Cerulli G. Patellar pain and incongruence: I: Measurements of incongruence. Clinical Orthopaedics and Related Research 1983;176:217-24  287. Tanamas SK, Teichtahl AJ, Wluka AE, Wang Y, Davies-Tuck M, Urquhart DM et al. The associations between indices of patellofemoral geometry and knee pain and patella cartilage volume: a cross-sectional study. BMC Musculoskeletal Disorders 2010;11(1):87  288. Teichtahl AJ, Parkins K, Hanna F, Wluka AE, Urquhart DM, English DR et al. The relationship between the angle of the trochlear groove and patella cartilage and bone morphology–a cross-sectional study of healthy adults. Osteoarthritis & Cartilage 2007;15(10):1158-62  289. Bellemans J, Colyn W, Vandenneucker H, Victor J. The Chitranjan Ranawat Award: is neutral mechanical alignment normal for all patients? The concept of constitutional varus. Clinical Orthopaedics and Related Research 2012;470(1):45-53  290. Dejour H, Walch G, Nove-Josserand L, Guier C. Factors of patellar instability: an anatomic radiographic study. Knee Surg Sports Traumatol Arthrosc 1994;2(1):19-26  291. The Cochrane Collaboration. Cochrane Handbook for Systematic Reviews of Interventions: version 5.1.0 10.2.1 publication bias. 2011. http://handbook.cochrane.org/chapter_10/10_2_1_publication_bias.htm Publication Bias. 292. Cohen AM, Stavri PZ, Hersh WR. A categorization and analysis of the criticisms of Evidence-Based Medicine. Int J Med Inf 2004;73(1):35-43  293. Wong YM, Chan ST, Tang KW, Ng GY. Two modes of weight training programs and patellar stabilization. J Athlet Train 2009;44(3):264-71 doi: http://dx.doi.org/10.4085/1062-6050-44.3.264   241  294. Macri E, Stefanik J, Khan K, Crossley K. Is tibiofemoral or patellofemoral alignment or trochlear morphology associated with patellofemoral osteoarthritis? A systematic review. Arthritis Care & Research 2016;68(10):1453-70 doi: 10.1002/acr.22842 295. Callaghan M, Guney H, Reeves N, Bailey D, Doslikova K, Maganaris CN et al. A knee brace alters patella position in patellofemoral osteoarthritis: a study using weight bearing magnetic resonance imaging. Osteoarthritis & Cartilage 2016;24(12):2055-60  296. Friel NA, Chu CR. The role of ACL injury in the development of posttraumatic knee osteoarthritis. Clinics in Sports Medicine 2013;32(1):1-12 doi: http://dx.doi.org/10.1016/j.csm.2012.08.017 297. Risberg MA, Oiestad BE, Gunderson R, Aune AK, Engebretsen L, Culvenor A et al. Changes in knee osteoarthritis, symptoms, and function after anterior cruciate ligament reconstruction: a 20-year prospective follow-up study. The American Journal of Sports Medicine 2016;44(5):1215-24  298. Leys T, Salmon L, Waller A, Linklater J, Pinczewski L. Clinical results and risk factors for Reinjury 15 Years After anterior cruciate ligament reconstruction a prospective study of hamstring and patellar tendon grafts. The American Journal of Sports Medicine 2012;40(3):595-605  299. Keays SL, Bullock-Saxton JE, Keays AC, Newcombe PA, Bullock MI. A 6-year follow-up of the effect of graft site on strength, stability, range of motion, function, and joint degeneration after anterior cruciate ligament reconstruction patellar tendon versus semitendinosus and gracilis tendon graft. The American Journal of Sports Medicine 2007;35(5):729-39    242  300. Neuman P, Englund M, Kostogiannis I, Fridén T, Roos H, Dahlberg LE. Prevalence of tibiofemoral osteoarthritis 15 years after nonoperative treatment of anterior cruciate ligament injury. A prospective cohort study. The American Journal of Sports Medicine 2008;36(9):1717-25  301. Neuman P, Kostogiannis I, Friden T, Roos H, Dahlberg LE, Englund M. Patellofemoral osteoarthritis 15 years after anterior cruciate ligament injury--a prospective cohort study. Osteoarthritis & Cartilage 2009;17(3):284-90 doi: http://dx.doi.org/10.1016/j.joca.2008.07.005 302. Øiestad BE, Holm I, Engebretsen L, Risberg MA. The association between radiographic knee osteoarthritis and knee symptoms, function and quality of life 10–15 years after anterior cruciate ligament reconstruction. British Journal of Sports Medicine 2011;45(7):583-88  303. Ackerman IN, Kemp JL, Crossley KM, Culvenor AG, Hinman RS. Hip and Knee Osteoarthritis Affects Younger People, Too. Journal of Orthopaedic & Sports Physical Therapy 2017;47(2):67-79 doi: https://dx.doi.org/10.2519/jospt.2017.7286 304. Culvenor AG, Schache AG, Vicenzino B, Pandy MG, Collins NJ, Cook JL et al. Are knee biomechanics different in those with and without patellofemoral osteoarthritis after anterior cruciate ligament reconstruction? Arthritis Care & Research 2014;66(10):1566-70  305. Hart HF, Culvenor AG, Collins NJ, Ackland DJ, Cowan SM, Machotka Z et al. Knee kinematics and joint moments during gait following anterior cruciate ligament reconstruction: a systematic review and meta-analysis. Br J Sports Med 2015 doi: 10.1136/bjsports-2015-094797   243  306. Papannagari R, Gill TJ, Defrate LE, Moses JM, Petruska AJ, Li G. In vivo kinematics of the knee after anterior cruciate ligament reconstruction: a clinical and functional evaluation. American Journal of Sports Medicine 2006;34(12):2006-12  307. van de Velde SK, Gill TJ, DeFrate LE, Papannagari R, Li G. The effect of anterior cruciate ligament deficiency and reconstruction on the patellofemoral joint. American Journal of Sports Medicine 2008;36(6):1150-9 doi: http://dx.doi.org/10.1177/0363546508314404 308. Cameron ML, Briggs KK, Steadman JR. Reproducibility and reliability of the outerbridge classification for grading chondral lesions of the knee arthroscopically. Am J Sports Med 2003;31(1):83-6  309. Stefanik JJ, Zumwalt AC, Segal NA, Lynch JA, Powers CM. Association between measures of patella height, morphologic features of the trochlea, and patellofemoral joint alignment: the MOST study. Clinical Orthopaedics & Related Research 2013;471(8):2641-8 doi: http://dx.doi.org/10.1007/s11999-013-2942-6 310. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;i:307 - 10  311. Portney LG, Watkins MP. Foundations of Clinical Research: Applications to Practice. Third Edition ed. Upper Saddle River, New Jersey: Pearson Eduction, Inc., 2009. 312. Streiner DL, Norman GR. Health Measurement Scales: a practical guide to their development and use. Fourth edition. Fourth ed. Oxford: Oxford University Press, 2008. 313. Muellner T, Kaltenbrunner W, Nikolic A, Mittlboeck M, Schabus R, Vecsei V. Anterior cruciate ligament reconstruction alters the patellar alignment. Arthroscopy 1999;15(2):165-8   244  314. Shabshin N, Schweitzer ME, Morrison WB, Parker L. MRI criteria for patella alta and baja. Skeletal Radiology 2004;33(8):445-50 doi: 10.1007/s00256-004-0794-6 315. Jarvela T, Paakkala T, Kannus P, Jarvinen M. The incidence of patellofemoral osteoarthritis and associated findings 7 years after anterior cruciate ligament reconstruction with a bone-patellar tendon-bone autograft. American Journal of Sports Medicine 2001;29(1):18-24  316. Powers CM, Shellock FG, Beering TV, Garrido DE, Goldbach RM, Molnar T. Effect of bracing on patellar kinematics in patients with patellofemoral joint pain. Medicine & Science in Sports & Exercise 1999;31(12):1714-20  317. Salsich GB, Perman WH. Tibiofemoral and patellofemoral mechanics are altered at small knee flexion angles in people with patellofemoral pain. J Sci Med Sport 2013;16(1):13-7 doi: http://dx.doi.org/10.1016/j.jsams.2012.04.003 318. Botchu R, Obaid H, Rennie WJ. Correlation between trochlear dysplasia and anterior cruciate ligament injury. Journal of Orthopaedic Surgery 2013;21(2):185-8  319. Ntagiopoulos PG, Bonin N, Sonnery-Cottet B, Badet R, Dejour D. The incidence of trochlear dysplasia in anterior cruciate ligament tears. International Orthopaedics 2014;38(6):1269-75 doi: http://dx.doi.org/10.1007/s00264-014-2291-4 320. Ristanis S, Giakas G, Papageorgiou CD, Moraiti T, Stergiou N, Georgoulis AD. The effects of anterior cruciate ligament reconstruction on tibial rotation during pivoting after descending stairs. Knee Surg Sports Traumatol Arthrosc 2003;11(6):360-5  321. Stergiou N, Ristanis S, Moraiti C, Georgoulis AD. Tibial rotation in anterior cruciate ligament (ACL)-deficient and ACL-reconstructed knees: a theoretical proposition for the development of osteoarthritis. Sports Medicine 2007;37(7):601-13    245  322. Webster KE, Feller JA. Tibial rotation in anterior cruciate ligament reconstructed knees during single limb hop and drop landings. Clinical Biomechanics 2012;27(5):475-9 doi: http://dx.doi.org/10.1016/j.clinbiomech.2011.12.008 323. Powers CM, Ward SR, Chan L-D, Chen Y-J, Terk MR. The effect of bracing on patella alignment and patellofemoral joint contact area. Medicine and Science in Sports and Exercise. 2004;36(7):1226-32  324. Pasco J, Lane S, Brennan S, Timney EN, Bucki-Smith G, Dobbins AG et al. Fracture risk among older men: osteopenia and osteoporosis defined using cut-points derived from female versus male reference data. Osteoporosis International 2014;25(3):857-62  325. Hall M, Heavens J, Cullum I, Ell P. The range of bone density in normal British women. The British Journal of Radiology 1990;63(748):266-69  326. Organization WH. Assessment of fracture risk and its application to screening for postmenopausal osteoporosis: report of a WHO study group [meeting held in Rome from 22 to 25 June 1992]. 1994  327. Laborie L, Lehmann T, Engesæter I, Sera F, Engesæter L, Rosendahl K. The alpha angle in cam-type femoroacetabular impingement. Bone Joint J 2014;96(4):449-54  328. Lepage-Saucier M, Thiéry C, Larbi A, Lecouvet FE, Berg BCV, Omoumi P. Femoroacetabular impingement: normal values of the quantitative morphometric parameters in asymptomatic hips. Eur Radiol 2014;24(7):1707-14  329. Thomas GE, Palmer AJ, Batra RN, Kiran A, Hart D, Spector T et al. Subclinical deformities of the hip are significant predictors of radiographic osteoarthritis and joint replacement in women. A 20 year longitudinal cohort study. Osteoarthritis & Cartilage 2014;22(10):1504-10 doi: http://dx.doi.org/10.1016/j.joca.2014.06.038   246  330. Song X, Jousilahti P, Stehouwer CD, Soderberg S, Onat A, Laatikainen T et al. Cardiovascular and all-cause mortality in relation to various anthropometric measures of obesity in Europeans. Nutr Metab Cardiovasc Dis 2015;25(3):295-304 doi: http://dx.doi.org/10.1016/j.numecd.2014.09.004 331. Rothman KJ, Moore LL, Singer MR, Nguyen U-SD, Mannino S, Milunsky A. Teratogenicity of high vitamin A intake. New England Journal of Medicine 1995;333(21):1369-73  332. Englund M, Guermazi A, Gale D, Hunter DJ, Aliabadi P, Clancy M et al. Incidental meniscal findings on knee MRI in middle-aged and elderly persons. New England Journal of Medicine 2008;359(11):1108-15 doi: http://dx.doi.org/10.1056/NEJMoa0800777 333. Cooper C, Cushnaghan J, Kirwan J, Rogers J, McAlindon T, McCrae F et al. Radiographic assessment of the knee joint in osteoarthritis. Annals of the Rheumatic Diseases 1992;51(1):80-82  334. Stefanik JJ, Gross KD, Guermazi A, Felson DT, Roemer FW, Zhang Y et al. The relation of MRI-detected structural damage in the medial and lateral patellofemoral joint to knee pain: the Multicenter and Framingham Osteoarthritis Studies. Osteoarthritis & Cartilage 2015;23(4):565-70 doi: http://dx.doi.org/10.1016/j.joca.2014.12.023 335. McGraw KO, Wong SP. Forming Inferences About Some Intraclass Correlation Coefficients. Psychological Methods 1996;1(1):30 - 46  336. Harrell F. Regression modeling strategies: with applications to linear models, logistic and ordinal regression, and survival analysis. New York: Springer, 2001. 337. Greenland S. Avoiding power loss associated with categorization and ordinal scores in dose-response and trend analysis. Epidemiology 1995;6(4):450-4    247  338. Kraemer HC, Morgan GA, Leech NL, Gliner JA, Vaske JJ, Harmon RJ. Measures of clinical significance. Journal of the American Academy of Child & Adolescent Psychiatry 2003;42(12):1524-29  339. Carrillon Y, Abidi H, Dejour D, Fantino O, Moyen B, Tran-Minh VA. Patellar instability: assessment on MR images by measuring the lateral trochlear inclination-initial experience. Radiology 2000;216(2):582-5  340. Sutter R, Dietrich TJ, Zingg PO, Pfirrmann CW. How useful is the alpha angle for discriminating between symptomatic patients with cam-type femoroacetabular impingement and asymptomatic volunteers? Radiology 2012;264(2):514-21  341. Kalichman L, Zhu Y, Zhang Y, Niu J, Gale D, Felson DT et al. The association between patella alignment and knee pain and function: an MRI study in persons with symptomatic knee osteoarthritis. Osteoarthritis & Cartilage 2007;15(11):1235-40  342. Zhang Y, Nevitt M, Niu J, Lewis C, Torner J, Guermazi A et al. Fluctuation of knee pain and changes in bone marrow lesions, effusions, and synovitis on magnetic resonance imaging. Arthritis Rheumatol 2011;63(3):691-99  343. Pal S, Draper CE, Fredericson M, Gold GE, Delp SL, Beaupre GS et al. Patellar maltracking correlates with vastus medialis activation delay in patellofemoral pain patients. The American Journal of Sports Medicine 2011;39(3):590-98  344. Noehren B, Barrance PJ, Pohl MP, Davis IS. A comparison of tibiofemoral and patellofemoral alignment during a neutral and valgus single leg squat: an MRI study. Knee 2012;19(4):380-6 doi: 10.1016/j.knee.2011.05.012 345. Becher C, Fleischer B, Rase M, Schumacher T, Ettinger M, Ostermeier S et al. Effects of upright weight bearing and the knee flexion angle on patellofemoral indices using magnetic   248  resonance imaging in patients with patellofemoral instability. Knee Surg Sports Traumatol Arthrosc 2015:1-9  346. Mariani S, La Marra A, Arrigoni F, Necozione S, Splendiani A, Cesare ED et al. Dynamic measurement of patello-femoral joint alignment using weight-bearing magnetic resonance imaging (WB-MRI). European Journal of Radiology 2015;84(12):2571-78  347. Shellock FG, Mink JH, Fox JM. Patellofemoral joint: kinematic MR imaging to assess tracking abnormalities. Radiology 1988;168(2):551-53  348. Draper CE, Besier TF, Santos JM, Jennings F, Fredricson M, Gold GE et al. Using real-time MRI to quantify altered joint kinematics in subjects with patellofemoral pain and to evaluate the effects of a patellar brace or sleeve on joint motion. Journal of Orthopaedic Research 2008;27(5):571-7  349. Sheehan FT, Derasari A, Brindle TJ, Alter KE. Understanding patellofemoral pain with maltracking in the presence of joint laxity: complete 3D in vivo patellofemoral and tibiofemoral kinematics. Journal of Orthopaedic Research 2009;27(5):561-70  350. International Atomic Energy Agency. Information for Patients. Secondary Information for Patients  2013. https://rpop.iaea.org/RPOP/RPoP/Content/InformationFor/Patients/information-patients/. 351. Zhang Y, Hunter DJ, Nevitt MC, Xu L, Niu J, Lui LY et al. Association of squatting with increased prevalence of radiographic tibiofemoral knee osteoarthritis: the Beijing Osteoarthritis Study. Arthritis & Rheumatism 2004;50(4):1187-92  352. Craig CL, Marshall AL, Sjostrom M, Bauman AE, Booth ML, Ainsworth BE et al. International physical activity questionnaire: 12-country reliability and validity. Medicine & Science in Sports & Exercise 2003;35(8):1381-95    249  353. Beighton P, Solomon L, Soskolne C. Articular mobility in an African population. Annals of the Rheumatic Diseases 1973;32(5):413  354. Herdman M, Gudex C, Lloyd A, Janssen MF, Kind P, Parkin D et al. Development and preliminary testing of the new five-level version of EQ-5D (EQ-5D-5L). Quality of Life Research 2011;20(10):1727-36 doi: http://dx.doi.org/10.1007/s11136-011-9903-x 355. Janssen MF, Pickard AS, Golicki D, Gudex C, Niewada M, Scalone L et al. Measurement properties of the EQ-5D-5L compared to the EQ-5D-3L across eight patient groups: a multi-country study. Quality of Life Research 2013;22(7):1717-27 doi: http://dx.doi.org/10.1007/s11136-012-0322-4 356. Roos EM, Roos HP, Lohmander LS, Ekdahl C, Beynnon BD. Knee Injury and Osteoarthritis Outcome Score (KOOS)--development of a self-administered outcome measure. Journal of Orthopaedic & Sports Physical Therapy 1998;28(2):88-96  357. Crossley KM, Macri EM, Cowan SM, Collins NJ, Roos EM. The patellofemoral pain and osteoarthritis subscale of the KOOS (KOOS-PF): development and validation using the COSMIN checklist. Br J Sports Med 2017:bjsports-2016-096776  358. Kujala UM, Jaakkola LH, Koskinen SK, Taimela S, Hurme M, Nelimarkka O. Scoring of patellofemoral disorders. Arthroscopy 1993;9(2):159-63  359. Swinkels-Meewisse EJ, Swinkels RA, Verbeek AL, Vlaeyen JW, Oostendorp RA. Psychometric properties of the Tampa Scale for kinesiophobia and the fear-avoidance beliefs questionnaire in acute low back pain. Manual Therapy 2003;8(1):29-36  360. Thomee P, Wahrborg P, Borjesson M, Thomee R, Eriksson BI, Karlsson J. A new instrument for measuring self-efficacy in patients with an anterior cruciate ligament injury. Scand J Med Sci Sports 2006;16(3):181-7    250  361. d'Entremont AG, McCormack RG, Horlick SG, Stone TB, Manzary MM, Wilson DR. Effect of opening-wedge high tibial osteotomy on the three-dimensional kinematics of the knee. Bone & Joint Journal 2014;96-B(9):1214-21 doi: https://dx.doi.org/10.1302/0301-620X.96B9.32522 362. Fellows RA, Hill NA, Macintyre NJ, Harrison MM, Ellis RE, Wilson DR. Repeatability of a novel technique for in vivo measurement of three-dimensional patellar tracking using magnetic resonance imaging. Journal of Magnetic Resonance Imaging 2005;22:145-53  363. Macintyre NJ, Hill NA, Fellows RA, Ellis RE, Wilson DR. Patellofemoral joint kinematics in individuals with and without patellofemoral pain syndrome. The Journal of Bone & Joint Surgery 2006;88:2596-605 doi: 10,2106/JBJS.E.00674 364. McWalter EJ, Hunter DJ, Harvey WF, McCree P, Hirko KA, Felson DT et al. The effect of a patellar brace on three-dimensional patellar kinematics in patients with lateral patellofemoral osteoarthritis. Osteoarthritis & Cartilage 2011;19(7):801-8 doi: http://dx.doi.org/10.1016/j.joca.2011.03.003 365. Grood ES, Suntay WJ. A joint coordinate system for the clinical description of three-dimensional motions: application to the knee. J Biomech Eng 1983;105(2):136-44  366. Fellows RA, Hill NA, Macintyre NJ, Harrison MM, Ellis RE, Wilson DR. Magnetic resonance imaging for in vivo assessment of three-dimensional patellar tracking. Journal of Biomechanics 2005;38:1643-52  367. Besl PJ, McKay ND. Method for registration of 3-D shapes. Robotics-DL tentative: International Society for Optics and Photonics 1992;586-606. 368. Cleveland WS. LOWESS: A Program for Smoothing Scatterplots by Robust Locally Weighted Regression. The American Statistician 1981;35(1):54-54 doi: 10.2307/2683591   251  369. Cohen J. Statistical power analysis for the behavioral sciences Lawrence Earlbaum Associates. Hillsdale, NJ 1988:20-26  370. Morris SB, DeShon RP. Combining effect size estimates in meta-analysis with repeated measures and independent-groups designs. Psychological methods 2002;7(1):105  371. Garson GD. Hierarchical linear modeling: Guide and applications: Sage, 2012. 372. Akaike H. A new look at the statistical model identification. IEEE transactions on automatic control 1974;19(6):716-23  373. Lau BC, Thuillier DU, Pedoia V, Chen EY, Zhang Z, Feeley BT et al. Inter-and intra-rater reliability of patellofemoral kinematic and contact area quantification by fast spin echo MRI and correlation with cartilage health by quantitative T1ρ MRI. The Knee 2016;23(1):13-19  374. Kim TH, Sobti A, Lee SH, Lee JS, Oh KJ. The effects of weight-bearing conditions on patellofemoral indices in individuals without and with patellofemoral pain syndrome. Skeletal Radiology 2014;43(2):157-64 doi: http://dx.doi.org/10.1007/s00256-013-1756-7 375. Freedman BR, Sheehan FT. Predicting three-dimensional patellofemoral kinematics from static imaging-based alignment measures. Journal of Orthopaedic Research 2013;31(3):441-7 doi: http://dx.doi.org/10.1002/jor.22246 376. Fry AC, Smith JC, Schilling BK. Effect of knee position on hip and knee torques during the barbell squat. The Journal of Strength & Conditioning Research 2003;17(4):629-33  377. Ariel B. Biomechanical analysis of the knee joint during deep knee bends with heavy load. Biomechanics IV: Springer, 1974:44-52. 378. Raudenbush SW, Bryk AS. Hierarchical linear models: applications and data analysis methods. 2nd edition. Thousand Oaks, CA: Sage Publications, 2002.   252  379. Brazaitis A, Tamošiūnas A, Tutkuvienė J, Gocentas A. Predictors of Patellofemoral Pain Applying Full Weight Bearing Kinematic MRI. Central European Journal of Sport Sciences and Medicine 2015;12(4):43-51  380. Macri EM, Culvenor AG, Morris HG, Whitehead TS, Russell TG, Khan KM et al. Lateral displacement, sulcus angle and trochlear angle are associated with early patellofemoral osteoarthritis following anterior cruciate ligament reconstruction. Knee Surg Sports Traumatol Arthrosc 2017:1-8  381. Macri EM, Felson DT, Zhang Y, Guermazi A, Roemer FW, Crossley KM et al. Patellofemoral morphology and alignment: reference values and dose-response patterns for the relation to MRI features of patellofemoral osteoarthritis. Osteoarthritis & Cartilage 2017: doi: 10.1016/j.joca.2017.06.005  382. Lee SP, Powers CM. Individuals with diminished hip abductor muscle strength exhibit altered ankle biomechanics and neuromuscular activation during unipedal balance tasks.[Erratum appears in Gait Posture. 2014 May;40(1):278]. Gait & Posture 2014;39(3):933-8 doi: https://dx.doi.org/10.1016/j.gaitpost.2013.12.004 383. Chester R, Smith T, Sweeting D, Dixon J, Wood S, Song F. The relative timing of VMO and VL in the aetiology of anterior knee pain: a systematic review and meta-analysis. BMC Musculoskeletal Disorders 2008;9:64  384. Wyndow N, De Jong A, Rial K, Tucker K, Collins N, Vicenzino B et al. The relationship of foot and ankle mobility to the frontal plane projection angle in asymptomatic adults. Journal of Foot & Ankle Research 2016;9:3 doi: https://dx.doi.org/10.1186/s13047-016-0134-9   253  385. Tomsich DA, Nitz AJ, Threlkeld AJ, Shapiro R. Patellofemoral alignment: reliability. Journal of Orthopaedic & Sports Physical Therapy 1996;23(3):200-08  386. Fitzgerald G, McClure PW. Reliability of measurements obtained with four tests for patellofemoral alignment. Physical Therapy 1995;75(2):84-90  387. Smith TO, Davies L, Donell ST. The reliability and validity of assessing medio-lateral patellar position: a systematic review. Manual Therapy 2009;14(4):355-62  388. Crossley KM, Zhang W-J, Schache AG, Bryant A, Cowan SM. Performance on the single-leg squat task indicates hip abductor muscle function. The American Journal of Sports Medicine 2011;39(4):866-73  389. Willson JD, Davis IS. Utility of the frontal plane projection angle in females with patellofemoral pain. Journal of Orthopaedic & Sports Physical Therapy 2008;38(10):606-15  390. Kim Y-M, Joo Y-B. Patellofemoral osteoarthritis. Knee Surgery & Related Research 2012;24(4):193-200 doi: 10.5792/ksrr.2012.24.4.193 391. Erickson BJ, Mascarenhas R, Sayegh ET, Saltzman B, Verma NN, Bush-Joseph CA et al. Does operative treatment of first-time patellar dislocations lead to increased patellofemoral stability? A systematic review of overlapping meta-analyses. Arthroscopy 2015;31(6):1207-15 doi: https://dx.doi.org/10.1016/j.arthro.2014.11.040 392. Sanchis-Alfonso V. How to deal with chronic patellar instability: What does the literature tell us? Sports & Health 2016;8(1):86-90 doi: https://dx.doi.org/10.1177/1941738115604156 393. Burnham JM, Howard JS, Hayes CB, Lattermann C. Medial patellofemoral ligament reconstruction with concomitant tibial tubercle transfer: a systematic review of outcomes   254  and complications. Arthroscopy 2016;32(6):1185-95 doi: https://dx.doi.org/10.1016/j.arthro.2015.11.039 394. Chouteau J. Surgical reconstruction of the medial patellofemoral ligament. Orthop Traumatol Surg Res 2016;102(1 Suppl):S189-94 doi: https://dx.doi.org/10.1016/j.otsr.2015.06.030 395. Grawe B, Stein BS. Tibial tubercle osteotomy: indication and techniques. J Knee Surg 2015;28(4):279-84 doi: https://dx.doi.org/10.1055/s-0035-1544973 396. Longo UG, Berton A, Salvatore G, Migliorini F, Ciuffreda M, Nazarian A et al. Medial patellofemoral ligament reconstruction combined with bony procedures for patellar instability: current indications, outcomes, and complications. Arthroscopy 2016;32(7):1421-7 doi: https://dx.doi.org/10.1016/j.arthro.2016.01.013 397. Mackay ND, Smith NA, Parsons N, Spalding T, Thompson P, Sprowson AP. Medial patellofemoral ligament reconstruction for patellar dislocation: a systematic review. Orthop 2014;2(8):2325967114544021 doi: https://dx.doi.org/10.1177/2325967114544021 398. Ntagiopoulos PG, Dejour D. Current concepts on trochleoplasty procedures for the surgical treatment of trochlear dysplasia. Knee Surgery, Sports Traumatology, Arthroscopy 2014;22(10):2531-9 doi: https://dx.doi.org/10.1007/s00167-014-3013-6 399. Song JG, Kang SB, Oh SH, Han JH, Shah D, Park HJ et al. Medial soft-tissue realignment versus medial patellofemoral ligament reconstruction for recurrent patellar dislocation: systematic review. Arthroscopy 2016;32(3):507-16 doi: https://dx.doi.org/10.1016/j.arthro.2015.08.012   255  400. Tarassoli P, Punwar S, Khan W, Johnstone D. Patellofemoral arthroplasty: a systematic review of the literature. Open Orthop J 2012;6:340-7 doi: https://dx.doi.org/10.2174/1874325001206010340 401. Parratte S, Ollivier M, Lunebourg A, Abdel MP, Argenson JN. Long-term results of compartmental arthroplasties of the knee: long term results of partial knee arthroplasty. Bone & Joint Journal 2015;97-B(10 Suppl A):9-15 doi: https://dx.doi.org/10.1302/0301-620X.97B10.36426. 402. Rodriguez-Merchan EC. Surgical treatment of isolated patellofemoral osteoarthritis. Hss J 2014;10(1):79-82 doi: https://dx.doi.org/10.1007/s11420-013-9375-3 403. van der List JP, Chawla H, Villa JC, Pearle AD. Why do patellofemoral arthroplasties fail today? A systematic review. Knee 2017;24(1):2-8 doi: https://dx.doi.org/10.1016/j.knee.2015.11.002 404. Edmonds DW, McConnell J, Ebert JR, Ackland TR, Donnelly CJ. Biomechanical, neuromuscular and knee pain effects following therapeutic knee taping among patients with knee osteoarthritis during walking gait. Clinical Biomechanics 2016;39:38-43  405. Hinman RS, Bennell KL, Crossley KM, McConnell J. Immediate effects of adhesive tape on pain and disability in individuals with knee osteoarthritis. Rheumatology 2003;42(7):865-9 doi: https://dx.doi.org/10.1093/rheumatology/keg233 406. Hinman RS, Crossley KM, McConnell J, Bennell KL. Efficacy of knee tape in the management of osteoarthritis of the knee: blinded randomised controlled trial. BMJ 2003;327(7407):135 doi: https://dx.doi.org/10.1136/bmj.327.7407.135   256  407. Hunter DJ, Harvey W, Gross KD, Felson DT, McCree P, Li L et al. A randomized trial of patellofemoral bracing for treatment of patellofemoral osteoarthritis. Osteoarthritis & Cartilage 2011;19(7):792-800  408. Callaghan MJ, Parkes MJ, Felson DT. The effect of knee braces on quadriceps strength and inhibition in subjects with patellofemoral osteoarthritis. Journal of Orthopaedic & Sports Physical Therapy 2016;46(1):19-25  409. Yu SP, Williams M, Eyles JP, Chen JS, Makovey J, Hunter DJ. Effectiveness of knee bracing in osteoarthritis: pragmatic trial in a multidisciplinary clinic. Int J Rheum Dis 2015  410. Zhang X, Eyles JP, Makovey J, Williams MJ, Hunter DJ. Is the effectiveness of patellofemoral bracing modified by patellofemoral alignment and trochlear morphology? BMC Musculoskeletal Disorders 2017;18(1):168  411. Pourahmadi MR, Takamjani IE, Hesampour K, Shah-Hosseini GR, Jamshidi AA, Shamsi MB. Effects of static stretching of knee musculature on patellar alignment and knee functional disability in male patients diagnosed with knee extension syndrome: A single-group, pretest–posttest trial. Manual Therapy 2016;22:179-89  412. Sahrmann S. Movement system impairment syndromes of the extremities, cervical and thoracic spines: Elsevier Health Sciences, 2010. 413. Piva SR, Goodnite EA, Childs JD. Strength around the hip and flexibility of soft tissues in individuals with and without patellofemoral pain syndrome. Journal of Orthopaedic & Sports Physical Therapy 2005;35(12):793-801 doi: https://dx.doi.org/10.2519/jospt.2005.35.12.793   257  414. Tarabichi S, Tarabichi Y. Can an anterior quadriceps release improve range of motion in the stiff arthritic knee? Journal of Arthroplasty 2010;25(4):571-5 doi: https://dx.doi.org/10.1016/j.arth.2009.04.015 415. Barton CJ, Lack S, Hemmings S, Tufail S, Morrissey D. The best practice guide to conservative management of patellofemoral pain: incorporating level 1 evidence with expert clinical reasoning. British Journal of Sports Medicine 2015;49(14):923-34 doi: https://dx.doi.org/10.1136/bjsports-2014-093637 416. Thomson C, Krouwel O, Kuisma R, Hebron C. The outcome of hip exercise in patellofemoral pain: A systematic review. Manual Therapy 2016;26:1-30  417. Espí-López GV, Arnal-Gómez A, Balasch-Bernat M, Inglés M. Effectiveness of manual therapy combined with physical therapy in treatment of patellofemoral pain syndrome: systematic review. Journal of Chiropractic Medicine 2016  418. Lack S, Barton C, Sohan O, Crossley K, Morrissey D. Proximal muscle rehabilitation is effective for patellofemoral pain: a systematic review with meta-analysis. British Journal of Sports Medicine 2015:bjsports-2015-094723  419. van der Heijden RA, Lankhorst NE, van Linschoten R, Bierma-Zeinstra S, van Middelkoop M. Exercise for treating patellofemoral pain syndrome. An abridged version of Cochrane systematic review. Eur J Phys Rehabil Med 2015;52(1):110-33  420. Kooiker L, Van De Port IG, Weir A, Moen MH. Effects of physical therapist–guided quadriceps-strengthening exercises for the treatment of patellofemoral pain syndrome: a systematic review. Journal of Orthopaedic & Sports Physical Therapy 2014;44(6):391-B1    258  421. Santos TR, Oliveira BA, Ocarino JM, Holt KG, Fonseca ST. Effectiveness of hip muscle strengthening in patellofemoral pain syndrome patients: a systematic review. Brazilian Journal of Physical Therapy 2015;19(3):167-76  422. Alrshood A, El Alwani A, Amein N, El Mughrabi M, Enaiba S, Almangoush A. A systematic review of the effect of open and closed kinetic chain exercises on the vastus medialis oblique and vastus lateralis muscles of patients with patellofemoral pain syndrome. International Journal of Orthopaedics 2017;3(1):152-61  423. Powers CM, Ho K-Y, Chen Y-J, Souza RB, Farrokhi S. Patellofemoral joint stress during weight-bearing and non—weight-bearing quadriceps exercises. Journal of Orthopaedic & Sports Physical Therapy 2014;44(5):320-27  424. Liao T-C, Yang N, Ho K-Y, Farrokhi S, Powers CM. Femur rotation increases patella cartilage stress in females with patellofemoral pain. Medicine and Sscience in Sports and Exercise 2015;47(9):1775-80  425. Peng H-T, Song C-Y. Predictors of treatment response to strengthening and stretching exercises for patellofemoral pain: an examination of patellar alignment. The Knee 2015;22(6):494-98  426