B7de PDF

DETERMINANTS OF STRUCTURAL CHANGES IN
THE KNEE JOINT IN A MIDDLE-AGED COHORT

WITH A LOW PREVALENCE OF OSTEOARTHRITIS
!
by
Hussain Ijaz Khan, MBBS
Menzies Institute for Medical Research
University of Tasmania
Submitted in fulfilment of the requirements for the degree of
Doctor of Philosophy
University of Tasmania
August 2016
Declaration of Originality
"This thesis contains no material which has been accepted for a degree or diploma by the
University or any other institution, except by way of background information and duly
acknowledged in the thesis, and to the best of my knowledge and belief no material
previously published or written by another person except where due acknowledgement is
made in the text of the thesis, nor does the thesis contain any material that infringes
copyright."
.
Signed: .. Date: . ? �/�t4zo f b
Authority of Access
This thesis may be made available for loan. Copying and communication of any part of this
thesis is prohibited for two years from the date this statement was signed; after that time
limited copying and communication is permitted in accordance with the Copyright Act 1968.
11
Statement Regarding Published Work Contained in Thesis
"The publishers of the papers comprising Chapters 4, 5, 6, 7 and 9 hold the copyright for that
content, and access to the material should be sought from the respective journals. The
remaining non published content of the thesis may be made available for loan and limited
copying and communication in accordance with the Copyright Act 1968."
Signed
iii
Statement of Ethical Conduct
The research associated with this thesis abides by the international and Australian codes on
human and animal experimentation, the guidelines by the Australian Government's Office of
the Gene Technology Regulator and the rulings of the Safety, Ethics and Institutional
Biosafety Committees of the University.
Signed: . Date: . ?. �.J _1:J o I b
iv
Magnetic Resonance Imaging and Radiographic Osteoarthritis Changes Over Ten Years in a
Midlife Cohort. Arthritis Care Res (Hoboken). 2016 Jul;68(7):958-64.
The Contribution of each author:
• IIlK was responsible for the analysis and interpretation of data, preparation of initial
manuscript and revisions of the manuscript.
• LC was responsible for data collection, preparation of initial manuscript and revisions of the
manuscript.
• DA was responsible for data cleaning and management, data interpretation and drafting of
the manuscript.
• AM was responsible for data collection and drafting of the manuscript.
• LB was responsible for data analysis and drafting of the manuscript.
• JPP and JMP were responsible for the measurements of femoral cartilage volume, meniscal
tears and meniscal extrusion, and drafting of the manuscript.
• CD, FC, and GJ were responsible for protocol development, data acquisition and drafting
of the manuscript.
Signed: ..... .................... . z�/o1-/;;,ot/:,

Date .................
Hussain Ijaz Khan, MBBS
(Candidate)
Signed: ...... .. . . Prof. Graeme Jones

Date: :1:-�;;I:.':"':.� '- 0 1�•
(Primary Supervisor)
viii
Publication arising from the thesis
Chapter 4: Khan HI, Aitken D, Blizzard L, Ding C, Pelletier JP, Pelletier JM, Cicuttini
F, Jones G. History of knee injury and MRI-assessed knee structures in middle- and older-
aged adults: a cross-sectional study. Clin Rheumatol. 2015 Aug;34(8):1463-72.
Chapter 5: Khan HI, Aitken D, Chou L, McBride A, Ding C, Blizzard L, Pelletier JP,
Pelletier JM, Cicuttini F, Jones G. A family history of knee joint replacement increases the
progression of knee radiographic osteoarthritis and medial tibial cartilage volume loss over
10 years. Osteoarthr. Cartil. 2015 Feb;23(2):203-9.
Chapter 6: Foong YC*, Khan HI*, Blizzard L, Ding C, Cicuttini F, Jones G, Aitken D. The
clinical significance, natural history and predictors of bone marrow lesion change over eight
years. Arthritis Res Ther. 2014 Jul 14;16(4):R149. *: Co-first authors
Chapter 7: Khan HI, Aitken D, Ding C, Blizzard L, Pelletier JP, Martel-Pelletier J, Cicuttini
F, Jones G. Natural history and clinical significance of meniscal tears over 8 years in a
midlife cohort. BMC Musculoskelet Disord. 2016 Jan 5;17:4.
Chapter 9: Ijaz Khan H, Chou L, Aitken D, McBride A, Ding C, Blizzard L, Pelletier JP,
Martel-Pelletier J, Cicuttini F, Jones G. Correlation Between Changes in Global Knee
Structures Assessed by Magnetic Resonance Imaging and Radiographic Osteoarthritis
Changes Over Ten Years in a Midlife Cohort. Arthritis Care Res (Hoboken). 2016
Jul;68(7):958-64.
Articles under-going peer-review:
Chapter 8: Natural history and clinical significance of cartilage defects over 10 years and the
relevance of cartilage defects in the progression of disease in the lateral compartment
ix
Manuscripts published during candidature, but external to thesis material:
McBride A*, Khan HI*, Aitken D, Chou L, Ding C, Blizzard L, Pelletier JP, Martel-Pelletier
J, Cicuttini F, Jones G. Does cartilage volume measurement or radiographic osteoarthritis at
baseline independently predict ten-year cartilage volume loss? BMC Musculoskelet
Disord. 2016 Feb 2;17:54.*Co-first authors
Khan HI, Aitken D, Chou L, McBride A, Ding C, Blizzard L, Pelletier JP, Martel-Pelletier J,
Cicuttini F, Jones G. Reply Letter to the Editor: Knee joint replacement and individual
susceptibility for progression of knee osteoarthritis and tibial cartilage volume loss: not only
genes run in the family. Osteoarthritis Cartilage. 2015 Oct;23 (10):
Khan HI, Dore D, Zhai G, Ding C, Pelletier JP, Martel-Pelletier J, Cicuttini F, Jones G.
Association between hip and knee cartilage measured using radiographs and magnetic
resonance imaging: the Tasmanian Older Adult Cohort Study. Rheumatology (Oxford). 2013
Nov;52 (11):2009-15.
Pan F, Ding C, Winzenberg T, Khan HI, Martel-Pelletier J, Pelletier JP, Cicuttini F, Jones G.
Response to: 'Does it make sense to investigate whether the offspring of people with a total
knee replacement for severe primary knee osteoarthritis have a higher risk of worsening knee
pain?' by Lei et al. Ann Rheum Dis. 2015 Aug;74 (8):e45.
Pan F, Ding C, Winzenberg T, Khan HI, Martel-Pelletier J, Pelletier JP, Cicuttini F, Jones G.
The offspring of people with a total knee replacement for severe primary knee osteoarthritis
have a higher risk of worsening knee pain over 8 years. Ann Rheum Dis. 2016 Feb;75
(2):368-73.
Pan F, Khan HI, Ding C, Winzenberg T, Martel-Pelletier J, Pelletier JP, Cicuttini F, Jones G.
Familial effects on structural changes relevant to knee osteoarthritis: a prospective cohort
study. Osteoarthritis Cartilage. 2015 Apr;23(4):559-64.
x
Scientific presentation arising from this thesis
2012 ¥ Association between hip and knee cartilage measured using radiographs and
magnetic resonance imaging: the Tasmanian Older Adult Cohort Study.
American College of Rheumatology Conference 2012 Washington DC,
USA. (Poster Presentation)
2013 ¥ A family history of knee joint replacement increases the progression of knee
radiographic osteoarthritis and cartilage volume loss over 10 years.
American College of Rheumatology Conference 2013 in San Diego, USA.
(Poster Presentation)
¥ History of knee injury and MRI-assessed knee structure in middle and older
aged adults. American College of Rheumatology Conference 2013 in San
Diego, USA. (Poster Presentation)
¥ Meniscal extrusion is a better predictor of radiographic osteoarthritis change
over ten years compared to cartilage volume. American College of
Rheumatology Conference 2013 in San Diego, USA. (Poster Presentation)
¥ Predictors of cartilage volume loss over 10 years in a mid-life cohort.
American College of Rheumatology Conference 2013 in San Diego, USA.
(Poster Presentation)
¥ The clinical significance, natural history and predictors of bone marrow
lesion change over eight years. American College of Rheumatology
Conference 2013 in San Diego, USA. (Oral Presentation)
2014 ¥ Natural history and clinical significance of meniscal tears in a mid-life

cohort. Australian Rheumatology Association Conference in Hobart 2014,
Australia. (Oral Presentation)
¥ Meniscal tracking in a midlife cohort. American College of Rheumatology
Conference 2014 in Boston, USA. (Poster Presentation)
¥ Predictors of cartilage volume loss over 10 years in a midlife cohort.
Australian Orthopaedic Association Conference 2014 in Melbourne,
Australia (Oral Presentation)
2015 ¥ Natural history and clinical significance of cartilage defects in a mid-life

cohort. Osteoarthritis Research Society International (OARSI) World
Congress 2015 in Seattle, Washington, USA. (Oral Presentation)
¥ Natural history of cartilage defects over 10 years and the role of sub-
chondral bone. Bone and Mineral Society Annual Meeting 2015 in Hobart,
Australia. (Poster Presentation)
xi
Awards resulting from thesis material
2012 International Postgraduate Research Scholarship (IPRS)
2014 Finalist for Young Investigator award at Australian Rheumatology

Association Conference
2015 Travel Award, Graduate Research Travel Fund, University of Tasmania,

Australia
2015 Finalist for Young Investigator award at Bone and Mineral Society Conference
2016 World Osteoarthritis Research Society Young Researcher Scholarship
2016 European Rheumatology Association Travel Grant- EULAR congress 2016
xii
Acknowledgement
I would like to start by thanking my primary supervisor, Professor Graeme Jones. I am

extremely grateful to him for taking me on as a PhD candidate and showing great belief in me
throughout the candidature. Graeme is one of the leading scientists in the field of
osteoarthritis. He has a wealth of knowledge in our field and I am fortunate to have had the
opportunity to do my PhD under him. His critical evaluation and intellectual input into my
research have contributed greatly to my research papers and my success as a student. More
importantly Graeme provided me with a lot of freedom to work the way I wanted and
encouraged me to provide my own input in all the projects we worked together. He always
showed great flexibility that allowed me to take my professional qualifying clinical exams
during my PhD candidature. His door is always open for his students. He is a great mentor
and I am extremely fortunate to have had him as my primary supervisor during my PhD. I
thank him for the many contributions he has made.
I am also extremely grateful to my co-supervisors, Dr Dawn Aitken and Associate Professor

Leigh Blizzard. Dawn became my co-supervisor right after the completion of her own PhD
candidature. She exactly knew what issues a student faces early on in the PhD candidature.
Dawn was a hands-on and pro-active supervisor especially early on in my candidature. I
personally felt I lacked certain skills compared to other PhD candidates in our group at the
start of my candidature due to my clinical background. Dawn was extremely patient and
spent a lot of time teaching me basics of scientific writing and statistical analysis. She is
extremely professional, organized and has great attention to detail with drafts of manuscripts.
In many ways I have tried to model my research career on hers. She has been an amazing
teacher and mentor to me.
Associate Professor Leigh Blizzard was my statistical supervisor. Early on in my candidature,

he was willing to sit with me while I ran new analyses to help me identify and solve
problems, and I thank him for his patience and time in this early period. Leigh is a very
talented teacher, has a great rapport with all his students and is extremely passionate about
biostatistics. He always made me think about the basics of bio-statistical concepts rather than
just focusing on concepts relevant to my thesis only. Leigh and I never missed an opportunity
for healthy sporting banter as well.
I would like to thank the numerous researchers who have given me advice and helped me
during my candidature. I would like to thank Dr Laura Laslett. Laura was at times like my
xiii
fourth supervisor, especially when Dawn left for her maternity leave. Laura always advised
me well during my candidature with especially valuable input in my final talk and drafting of
this thesis. I would also like to thank Professor Changhai Ding for his contribution. Changhai
was one of the chief investigators in the Offspring study and his feedback on all my papers
was extremely valuable. Changhai also trained me to assess cartilage defects on MRI and I
greatly appreciate his mentorship. Thank you to Professor Flavia Cicuttini, Professor Johanne
Martel-Pelletier and Professor Jean-Pierre Pelletier for their feedback on all my papers.
Thank you to Dr Jane Zochling for the engaging casual work in Tasmanian Ankylosing
Spondylitis Study. Thank you to my fellow PhD students and friends, especially Dr Benny
Antony whom I frequently bounced ideas off.
Thank you to the funding organisations especially National Health and Medical Research
Council of Australia and Masonic Centenary Medical Research Foundation, which supported
the Offspring study. During my candidature, I have had financial support from a number of
sources. I acknowledge the University of Tasmania for awarding me International
Postgraduate Research Scholarship (IPRS). Thank you to the many Offspring study staff and
volunteers for the tremendous work you have put into this project. I would like to say a huge
thank you to the Offspring study participants who generously gave their time to make this
research possible.
I would like to say a big thank you to my partner Farzaneh Atashrazm for her love, patience,
friendship, and amazing support throughout these years. Farzaneh and I started our PhD
candidatures around the same time and sharing these years with her made the whole
experience even more enjoyable. She listened to my presentations hundreds of times with
patience, supported me every step of the way and always encouraged me in tough times.
InshAllah Farzaneh and I shall graduate together later this year and I cannot think of a more
fitting end to this amazing journey.
Finally I would like to thank my parents, Samra and Ijaz Khan. I am who I am today because
of their numerous sacrifices. My parents always made sure that my two brothers and I got the
best education. They invested emotionally and financially to provide us every thing possible
to succeed not only in our careers but also in personal and social lives. I cannot thank my
parents enough for their unconditional love and support. I hope this thesis in some way
justifies all the sacrifices they made for me. I could not have done this without their prayers.
xiv
List of Abbreviations
2D Two-dimensional
3D Three-dimensional
ACR American College of Rheumatology
AQoL Assessment of quality of life
BME Bone marrow edema
BMI Body mass index
BML Bone marrow lesion
CI Confidence interval
CTX-II Type II collagen C-terminal telopeptide
CV Coefficient of variation
dGEMRIC Delayed gadolinium-enhanced magnetic resonance imaging of cartilage
DMOAD Disease-modifying osteoarthritis drug
FDA Food and Drug Administration
FFQ Food frequency questionnaire
FS Fat suppressed
FSE Fast spin echo
GRE Gradient-recalled echo
HDL High-density lipoprotein
ICC Intraclass correlation coefficient
IL Interleukin
JSN Joint space narrowing
JSW Joint space width
K/L Kellgren and Lawrence
KOOS Knee injury and Osteoarthritis Outcome Score
LF Lateral femoral
xv
LSC Least significant criterion
LT Lateral tibial
MF Medial femoral
MOST Multicentre Osteoarthritis Study
MRI Magnetic resonance imaging
MT Medial tibial
NHMRC The National Health and Medical Research Council
NSAIDS non-steroidal anti-inflammatory drugs
OA Osteoarthritis
OR Odds ratio
Pa Per annum
RA Rheumatoid arthritis
RCT Randomised controlled trial
ROA Radiographic osteoarthritis
ROI Region of interest
sBMD Subchondral bone mineral density
SD Standard deviation
STIR Short tau inversion recovery
TASOAC Tasmanian Older Adult Cohort study
TKR Total knee replacement
THR Total hip replacement
US United States
VAS Visual analog score
WOMAC Western Ontario and McMasters Universities Osteoarthritis Index
WORMS Whole-Organ magnetic resonance imaging score
ZA Zoledronic acid
xvi
Table of Content
Declarations ............................................................................................................................... i
Acknowledgement…………………………………………………………………………………..…...…xiii
List of Abbreviations ............................................................................................................... xv
Table of Content………………………………………...……………………….…………xvii
List of Tables………………………………………………………………………………xxiii
List of Figures………………………………………………………………………………xxv
Abstract ...................................................................................................................................... 1
Chapter One-Literature Review.................................................................................... 4

1.1 Osteoarthritis ....................................................................................................................... 5
1.2 Knee OA............................................................................................................................... 5
1.2.1 Burden of disease....................................................................................................................... 6
1.2.2 Applied anatomy of the knee ..................................................................................................... 7
1.2.3 Knee OA pathophysiology ........................................................................................................ 9
1.2.4 Symptoms .................................................................................................................................. 9
1.2.5 Risk factors .............................................................................................................................. 10
1.2.6 Diagnosis ................................................................................................................................. 12
1.2.6.1 Clinical criteria ................................................................................................................................ 12
1.2.6.2 Imaging ........................................................................................................................................... 13
1.2.7 Radiography versus MRI ......................................................................................................... 17
1.2.8 Why study middle-aged adults?............................................................................................... 20
1.3 Offspring Study ................................................................................................................. 20
1.3.1 Evidence from earlier phases of Offspring study .................................................................... 21
1.4 Summary............................................................................................................................ 23
Chapter Two-Research Questions .............................................................................. 24
xvii
2.1 Research Questions ........................................................................................................... 25
2.2 Key Hypothesis .................................................................................................................. 25
Chapter Three-Methodology ........................................................................................ 26

3.1 Prelude............................................................................................................................... 27
3.2 Study population and design ............................................................................................ 27
3.3 Anthropometrics ................................................................................................................ 30
3.4 Leg strength ....................................................................................................................... 30
3.5 Knee joint injury and surgery ........................................................................................... 30
3.6 Knee pain ........................................................................................................................... 31
3.7 Magnetic Resonance Imaging .......................................................................................... 31
3.7.1 Cartilage volume...................................................................................................................... 32
3.7.2 Cartilage defects ...................................................................................................................... 33
3.7.3 Bone marrow lesions ............................................................................................................... 33
3.7.4 Meniscal tears .......................................................................................................................... 34
3.7.5 Meniscal extrusion ................................................................................................................... 35
3.7.6 Tibial bone area ....................................................................................................................... 36
3.7.7 Effusion ................................................................................................................................... 37
3.8 Radiology ........................................................................................................................... 37
3.9 Summary of outcome factors, study factors and covariates ............................................ 38
3.10 Sample size and role of the candidate in the Offspring study ....................................... 40
3.11 Ethical considerations .................................................................................................... 40
3.12 Statistical analysis ........................................................................................................... 41
Chapter Four-Injury and MRI-assessed Knee Structure .................................. 42

4.1 Introduction....................................................................................................................... 43
4.2 Methods ............................................................................................................................. 45
4.2.1 Subjects.................................................................................................................................... 45
4.2.2 Anthropometrics ...................................................................................................................... 46
4.2.3 Knee joint injury and surgery .................................................................................................. 46
xviii
4.2.4 MRI.......................................................................................................................................... 46
4.2.5 TASOAC ................................................................................................................................. 47
4.2.6 Offspring.................................................................................................................................. 47
4.2.7 Cartilage volume...................................................................................................................... 47
4.2.10 Tibial bone area ..................................................................................................................... 49
4.2.11 Meniscal damage ................................................................................................................... 49
4.2.12 Data analysis .......................................................................................................................... 49
4.3 Results................................................................................................................................ 51
4.4 Discussion.......................................................................................................................... 58
Chapter Five-Role of Family History in Osteoarthritis Progression ............. 62

5.1 Introduction....................................................................................................................... 63
5.2 Methods ............................................................................................................................. 65
5.2.1 Study subjects .......................................................................................................................... 65
5.2.3 Knee pain ................................................................................................................................. 65
5.2.4 Leg strength ............................................................................................................................. 66
5.2.5 Magnetic resonance imaging ................................................................................................... 66
5.2.7 Bone area ................................................................................................................................. 68
5.2.8 Meniscal tears .......................................................................................................................... 68
5.2.10 Radiology............................................................................................................................... 69
5.2.11 Statistical analysis.................................................................................................................. 69
5.3 Results................................................................................................................................ 71
5.4 Discussion.......................................................................................................................... 76
Chapter Six-Natural History of Bone Marrow Lesions ...................................... 80
xix
6.1 Introduction....................................................................................................................... 81
6.2 Materials and methods ...................................................................................................... 83
6.2.1 Study subjects .......................................................................................................................... 83
6.2.3 Leg strength ............................................................................................................................. 84
6.2.4 Knee pain ................................................................................................................................. 84
6.2.5 Radiography............................................................................................................................. 84
6.2.6 Magnetic resonance imaging ................................................................................................... 85
6.2.7 Statistical analysis.................................................................................................................... 87
6.3 Results................................................................................................................................ 88
6.3.1 Participant characteristics ........................................................................................................ 88
6.3.2 Natural history ......................................................................................................................... 88
6.3.3 Pain .......................................................................................................................................... 91
6.3.4 Factors affecting bone marrow lesion change ......................................................................... 91
6.4 Discussion.......................................................................................................................... 94
Chapter Seven-Natural History of Meniscal Tears .............................................. 98

7.1 Background ....................................................................................................................... 99
7.2 Methods ........................................................................................................................... 100
7.2.1 Knee pain ............................................................................................................................... 100
7.2.2 Knee joint injury .................................................................................................................... 101
7.2.3 Magnetic resonance imaging ................................................................................................. 101
7.2.4 Meniscal tears ........................................................................................................................ 101
7.2.5 Meniscal extrusion ................................................................................................................. 102
7.2.6 Cartilage volume.................................................................................................................... 102
7.2.7 Cartilage defects .................................................................................................................... 102
7.2.8 Bone marrow lesions ............................................................................................................. 103
7.2.9 Effusion ................................................................................................................................. 103
7.2.10 Radiography......................................................................................................................... 103
xx
7.2.11 Statistical analysis................................................................................................................ 104
7.3 Results.............................................................................................................................. 105
7.3.1 Natural History ...................................................................................................................... 105
7.3.2 Predictors of change .............................................................................................................. 108
7.3.3 Pain ........................................................................................................................................ 108
7.3.4 Structural changes.................................................................................................................. 110
7.4 Discussion........................................................................................................................ 111
Chapter Eight-Natural Hisotry of Cartilage Defects ......................................... 115

8.1 Introduction..................................................................................................................... 116
8.2 Methods ........................................................................................................................... 117
8.2.1 Study subjects ........................................................................................................................ 117
8.2.2 Anthropometrics .................................................................................................................... 117
8.2.3 Knee pain ............................................................................................................................... 118
8.2.4 Magnetic resonance imaging ................................................................................................. 118
8.2.5 Cartilage defects .................................................................................................................... 118
8.2.6 Cartilage volume.................................................................................................................... 119
8.2.7 Meniscal tears ........................................................................................................................ 119
8.2.8 Meniscal extrusion ................................................................................................................. 119
8.2.9 Bone marrow lesions ............................................................................................................. 120
8.2.10 Effusion ............................................................................................................................... 120
8.2.11 Radiography......................................................................................................................... 120
8.2.12 Statistical analysis................................................................................................................ 121
8.3 Results.............................................................................................................................. 122
8.4 Discussion........................................................................................................................ 130
Chapter Nine-Correlation between MRI and X-Ray Knee Structures ....... 133
9.1 Introduction..................................................................................................................... 134
9.2 Patients and Methods ...................................................................................................... 135
9.2.1 Study Population.................................................................................................................... 135
xxi
9.2.2 Anthropometrics .................................................................................................................... 135
9.2.3 Knee joint injury and surgery ................................................................................................ 136
9.2.4 Imaging .................................................................................................................................. 136
MRI ........................................................................................................................................................... 136
Cartilage volume ....................................................................................................................................... 137
Cartilage defects ........................................................................................................................................ 137
Meniscal tears ........................................................................................................................................... 138
Meniscal extrusion .................................................................................................................................... 138
9.2.5 Radiology............................................................................................................................... 139
9.2.6 Statistical analysis.................................................................................................................. 139
9.3 Results.............................................................................................................................. 141
9.4 Discussion........................................................................................................................ 145
Chapter Ten-Summary and Future Direction ..................................................... 148

10.1 Summary........................................................................................................................ 149
10.2 Future Direction ........................................................................................................... 151
References .......................................................................................................................... 154

Appendix A ............................................................................................................................ 169
Appendix B ............................................................................................................................ 170
Appendix C ............................................................................................................................ 178
xxii
List of Tables
Table 1.1: Kellgren and Lawrence (K/L) grading system…………………………………. 14
Table 1.2: Individual radiographic features measured by the OARSI atlas for tibiofemoral
Osteoarthritis…………………………………….…………………………………….….….15
Table 3.1: Baseline characteristics of the participants who were followed-up and who were
lost to follow up………………………………………………………………………………29
Table 3.2: Summary of outcome factors, study factors, and covariates used in this thesis….39
Table 4.1: Characteristics of the study participants………………………………………….51
Table 4.2: Association between injury and cartilage volume in the knee………………...…52
Table 4.3: Association between injury and cartilage defects in the knee……………………53
Table 4.4: Association between injury and bone marrow lesions in the knee……………….54
Table 4.5: Association between injury and tibial bone area in the knee………………….…55
Table 4.6: Association between injury and meniscal pathology in the knee………….…….56
lost to follow up…………………………………….…………………………………….….72
Table 5.2: Baseline characteristics of the study participants……………………….……….73
Table 5.3: Comparison of radiographic changes and cartilage loss (absolute) between
offspring and controls…………………………………….………………………….………74
Table 5.4: Multivariable analyses of differences between offspring and controls in changes in
radiographic changes and cartilage loss (absolute) ………………………………………….75
Table 6.1: Participant characteristics…………………………………………………….….89
Table 6.2: Relationship between change in WOMAC and change in BML size……...…….92
Table 6.3: Relationship between change in WOMAC and site-specific change in BML
size……………………………………………………………………………………………92
xxiii
Table 6.4: Predictors of change in bone marrow lesion size………………………………...93
Table 7.1: Characteristics (at visit-2) of participants with and without any change (incident
tears and increase in score) in tears over 8 years………………………………………..….107
Table 7.2: Predictors of change in total knee meniscal tears over 8 years…………………108
Table 7.3: Association between change in meniscal tears and change in pain over 8
years………………………………………………………………………………………...109
Table 7.4: Association between change in meniscal tears and knee structures on MRI over 8
Years…………………………………….………………………………………………….110
Table 8.1: Characteristics of participants with and without any increase in mean cartilage
defects score over 10 years…………………………………….……………………………124
Table 8.2: Predictors of change in cartilage defects………………………………….…….126
Table 8.3: Association between change in cartilage defects and structural changes assessed
on MRI…………….…………………………………….…………….……………………128
Table 8.4: Association between change in cartilage defects and change in pain…………..129
Table 9.1: Characteristics of the study participants…………….…………….………….…148
Table 9.2: Correlation between structural changes on MRI over 8-10 years and radiographic
changes over 10 years…………….…………….…………………….……………….……144
xxiv
List of Figures
Figure 1.1: Knee joint anatomy. A) Anterior view B) superior view of the tibial plateau.….8
Figure 1.2: Radiographic joint space narrowing…………….……………..…..….…...……16
Figure 1.3: Radiographic osteophytes…………….…………….…………….……..………16
Figure 3.1: Flow chart Offspring study…………….…………….…………….……………28
Figure 3.2: Cartilage Volume Segmentation …………….…………………………….……32
Figure 3.3: Grade-2 cartilage defect …….…………………….………………..……...……33
Figure 3.4: Bone marrow lesion ……….…………………….…….…………………..……34
Figure 3.5: A. Normal meniscus B. meniscal tear …….…………………………..…...……35
Figure 3.6: Medial and lateral tibial bone area …….…………….…………...….…….……36
Figure 3.7: Joint space narrowing .………..…………….……………………….…….……38
Figure 3.8: Osteophyte …………….…………….…………………….…………...….……38
Figure 6.1: Natural history of bone marrow lesions…………….……………….….………90
Figure 7.1: Prevalence and natural history of meniscal tears. A) Prevalence of meniscal tears
at visit 2, B) Site-specific distribution of meniscal tears at visit 2……………………..…..106
Figure 8.1: Natural history of cartilage defects. A) Site specific prevalence of cartilage
defects at baseline. B) Site-specific change in cartilage defects over 10 years…………..…123
Figure 8.2: Association between cartilage defects at baseline and cartilage volume loss over
10 years. A) Medial tibio-femoral compartment. B) Lateral tibio-femoral compartment. C)
Patellar compartment…………….…………….…………………….…………….…..……128
Figure 9.1: Flow chart of the Offspring study…………….……………….………….……141
xxv
Abstract
Osteoarthritis (OA) is the most common joint disorder in adults around the world. Knee OA
is the most common form of OA in weight-bearing joints and results in deterioration of knee
structures and function in older adults for which there is no cost-effective treatment currently
available. The natural history of knee OA is highly variable and can involve any part of the
joint including the articular cartilage, meniscus, sub-chondral bone and synovium. Use of
MRI has revolutionised the understanding of knee OA disease process but there is limited
long-term data available in middle-aged adults with early disease changes, as most studies
have focused on older adults with established disease. Identifying modifiable risk factors
early in life has the potential to prevent or delay the development of knee OA in later life.
This thesis aims to investigate the long-term knee structural natural history data in middle-
aged adults and subsequent correlations with frequent knee symptoms.
A population-based sample of middle-aged adults (mean age 45(26–61) years; 58% females
participated at baseline and approximately 2 and 10 years later. Matched sampling was used
to recruit the study participants. Half of the participants were the adult offspring of patients
who had a knee replacement performed for idiopathic knee OA at any Hobart hospital from
1996 to 2000. The other half were age and sex matched controls, randomly selected from the
population (using electoral rolls) with no history of knee OA in either parent. Cartilage
volume, cartilage defects, bone area, bone marrow lesions (BMLs), meniscal tears, meniscal
extrusion and effusion were determined using magnetic resonance imaging (MRI). X-ray was
used to assess radiographic OA [joint space narrowing (JSN) and osteophytes]. Multiple
questionnaires were used to assess pain, function, history of knee joint injury/surgery and
physical activity.
The first study examined the cross-sectional association between history of knee injury and
knee structural damage assessed on MRI in middle-aged adults from the Offspring study and
in a random community based sample of older adults. In middle-aged adults, BML presence,
tibial bone area and meniscal extrusion presence were significantly higher in those with knee
injury, whereas in older adults, cartilage defect presence, cartilage volume, BML presence
and tibial bone area were significantly associated with knee injury. This was the first study to
look at the association between history of knee injury and knee joint structural changes
assessed on MRI and found that the association between knee injury and MRI-assessed
1
structural pathology in the knee joint is moderate and appears to be stronger in older adults
compared to middle-aged adults.
In the second study, a family history of knee joint replacement due to OA increased the risk
of radiographic OA (JSN and osteophytes) and medial tibial cartilage volume loss over 10
years compared to community acquired controls with no family history of OA. Most of these
changes were mediated by differences in baseline characteristics of offspring and controls
except for increase in medial JSN.
Third study looked at the natural history of BMLs in middle-aged adults and found that the
natural history of knee BMLs was unstable. BMLs were common in middle-aged adults at
baseline. 24% of these BMLs at baseline increased in size, 55% remained stable and 21%
decreased in size or resolved completely over 8 years. Change in BMLs was predicted by
BMI and strenuous physical activity. An increase in BML size or a new BML resulted in an
increase in pain especially in males and those with a family history of OA.
Fourth study looked at the natural history of meniscal tears. Only 22% of the participants had
a meniscal tear at baseline. Over 8 years, 16 % of the participants had an increase in severity
of meniscal tears while none improved. Change in meniscal tears shared common risk factors
with knee OA and was independently associated with worsening knee pain and structural
damage suggesting that meniscal tears are on the knee OA causal pathway and not just a
result of mechanical factors.
Fifth study looked at the natural history of cartilage defects. 44% of the participants had at
least one cartilage defect at any site at baseline. Most of these defects remained stable,
whereas 26% increased and 13% decreased in severity over 10 years. Cartilage defects
independently predicted cartilage volume loss in the lateral compartment only. Change in
cartilage defects on the other hand was associated with changes in BMI and structural
changes/symptoms mostly in the lateral compartment, suggesting a more crucial role of
cartilage defects in the development of lateral compartment knee OA.
Sixth study examined the correlation between changes in structural abnormalities assessed on
MRI and change in radiographic OA over 10 years. Change in JSN was correlated with
change in meniscal tears and, to a lesser extent, with meniscal extrusion and cartilage defects.
In this sample, change in JSN was a composite measure that did not reflect cartilage volume
2
loss prompting the review of the use of JSN as an outcome measure in chondro-protective
drug trials.
In conclusion, this series of related studies detail the natural history of knee structural
progression in middle-aged adults. Structural changes such as BMLs and cartilage defects
have the potential of reversibility in early disease and should be targeted in disease modifying
clinical trials. Meniscal tears and BMLs should be targeted in symptom modifying clinical
trials especially in those with a family history of OA. Lastly findings from this thesis suggest
that the use of JSN as an outcome measure in chondro-protective trial should be reviewed.
3
Chapter 1 Literature Review
Chapter One
Literature Review
Page 4
1.1 Osteoarthritis
Osteoarthritis (OA) is the most common joint disorder, and there is evidence that a majority
of individuals over the age of 65 have radiographic and/or clinical evidence of OA [1]. It is a
chronic joint disease and commonly involves weight-bearing joints such as the knees, hips, or
spine, with hands and neck also being frequently affected sites [2].
Early investigators tended to regard OA as an isolated degenerative disease that resulted from
wear and tear, and was an inevitable consequence of aging [3]. Over the past decade there has
been a significant shift in the conceptualization of OA etiology and pathogenesis. OA is now
considered as the end-point of a complex series of structural changes that result from a series
systemic risk factors over many years [4-8]. The old-fashioned wear and tear model has been
rejected in favour of a newer and more nuanced inflammatory/molecular model [9, 10]. The
synovial joint is now conceptualized as an organ, and OA represents failure of that organ [8].
At a molecular level, the disease occurs when the dynamic equilibrium between the
breakdown and repair of the synovial joint tissues is overwhelmed [11].
1.2 Knee OA
The knee joint is the most commonly affected weight bearing joint by OA [4, 12]. Nearly one
in two older adults are affected by knee OA by the age of 85 [12]. The knee joint is also the
site most affected by pain in older adults and most of this is attributed to OA in this age group
[12]. Despite the high disease burden, there are currently no registered disease-modifying
knee OA drugs available. Therefore, there is an urgent need for research to better understand
the disease process and develop cost-effective approaches to prevent or slow down the
progression of the disease.
The research conducted in this thesis focuses on knee OA and unless otherwise stated the
remainder of the literature review will discuss OA at this site.
Page 5
1.2.1 Burden of disease
As the most common form of joint disease, OA is associated with an extremely high
economic burden. This burden is largely attributable to the effects of disability, comorbid
disease, and the expense of treatment. Although typically associated with less severe effects
on quality of life and per capita expenditures than rheumatoid arthritis [13], OA is
nevertheless a more costly disease in economic terms because of its far higher prevalence and
lack of cost-effective conservative treatment options available to the patients [14-16]. The
global prevalence of radiographically confirmed symptomatic knee and hip OA in 2010 was
estimated to be 3.8% and 0.85%, respectively [17]. OA ranked 13th in the top 25 causes of
global years lived with disability and the 4th leading cause that showed an increase in the
years lived with disability from 1990 to 2013 [13].
1.9 million (or one in 12) Australians suffer from OA costing the health system $3.75 billion
and the economy around $22 billion annually [18]. The burden of OA is expected to increase
exponentially in coming decades due to an ageing and increasingly obese population, with
prevalence expected to reach three million Australians by 2032 [18]. 76.7% of the
expenditure on OA is due to admitted patient costs mainly attributed to knee and hip
arthroplasties [14]. In 2012-13 there were 103,763 hospitalisations with a principal diagnosis
of OA, an increase from 347 hospitalisations per 100,000 population in 2002-03 to 453 per
100,000 population in 2012-13 [14]. Over 85,000 knee and hip replacement procedures
(majority due to OA) were performed in Australian hospitals during 2012, each costing an
average of $15,000–$31,900 [14]. The number of joint replacements being performed is
increasing at a rate of 10% per annum and by 2018, it is expected that the number of joint
replacements will be double the number performed in 2012 [14].
Page 6
While the mortality rate for OA is low, there is also a cost in terms of burden of disease. The
pain and disability patients experience can lead to a loss of health and wellbeing, loss of
leisure time, and a decreased quality of life. This further contributes to the costs of OA
through the loss of production to the economy, increased absenteeism, reduced work capacity
and performance, and reduced labour force participation as a result of the related disease
morbidity.
1.2.2 Applied anatomy of the knee
The knee joint (Figure 1.1) is one of the most complex joints in the human body and is an
important joint for locomotion [19]. The knee joint acts as a hinge joint for locomotion
resulting in flexion-extension movement of the knee, with only a small degree of axial
rotation [20, 21]. During flexion-extension the articular surfaces of the femur roll (and glide)
over the tibial and patellar surfaces [21]. The knee joint is therefore conceptualised as two
joints—a tibiofemoral and a patellofemoral joint. Functionally the tibiofemoral joint is further
divided into the medial and lateral compartments. The medial tibiofemoral compartment is
under higher compressive forces due to the way the femur and tibia bones are aligned.
In nearly all circumstances, the knee works in axial compression under the action of gravity.
It must therefore reconcile two opposed requirements, namely mobility and stability. This
problem is resolved by an ingenious arrangement of soft tissue structures in and around the
knee joint [22]. The knee joint is encapsulated by the tendons of the leg muscles, including
the patellar tendon, which make the joint movement possible and provides dynamic stability
[22]. Anterior cruciate ligament (ACL), posterior cruciate ligament (PCL) and collateral
ligaments stabilise the femur and tibia within the knee joint [22]. Articular surfaces of all the
Page 7
bones are covered by fibrocartilage that help in reducing friction and wear-and-tear on the
surfaces of the underlying bones during movement [21]. There are two menisci in the space
between the femoral and tibial condyles [21, 23]. These are crescent-shaped lamellae, each
with an anterior and a posterior horn, and are triangular in cross section. Menisci aid in load
transmission between the femur and tibia bones, act as shock absorbers under compressive
forces and protect the articular cartilage from wear-and-tear [24]. Physiological synovial
fluid/effusion lubricates all these structures to decrease friction during movement as well
[25].
Figure 1.1. Knee joint anatomy. A) Anterior view B) superior view of the tibial plateau
(Reproduced with permission from Markis, E. A. et al. [26])
Page 8
1.2.3 Knee OA pathophysiology
The natural history of osteoarthritis of the knee is highly variable, with the disease improving
in some patients, remaining stable in others, and gradually worsening in others [11].
Traditionally knee OA was regarded as an isolated disease that resulted from articular
cartilage volume loss due to age related degenerative process [27]. However advancements in
imaging modalities [28, 29] and molecular analyses [30] over the last decade have shown that
knee OA is an active process, which can involve any part of the joint. The structural
alterations that form the OA disease process are markedly collinear [31], that is, as hyaline
articular cartilage becomes morphologically abnormal, other structural processes occur
including changes to the meniscus, bone marrow, sub-chondral bone, synovial lining,
ligaments and peri-articular muscles.
1.2.4 Symptoms
OA represents one of the most frequently occurring painful conditions [12, 17, 32]. Pain is
the most common symptom in OA and the usual reason for seeking medical advice [32].
According to the National Health Survey 2011-12 [33], people aged 18 and over with
osteoarthritis were 3.5 times more likely to report very severe pain (4.9%) compared with
those without OA (1.4%). Other signs and symptoms of the disease include joint stiffness,
tenderness, inflammation, crepitus, instability, and muscle weakness [34-36]. These
symptoms are initially felt during and after activity, but as the disease progresses it may occur
with minimal movement or even during rest [36]. Knee pain is associated with a considerable
reduction in functional ability [12, 35], which in turn strongly predicts future disability and
dependency [36].
Approximately 25 percent of persons 55 years of age or older have had knee pain on most
days in a month in the past year, and about half of them have radiographic evidence of knee
OA, a group considered to have symptomatic OA [33, 36]. The prevalence of symptomatic
knee OA and knee pain has increased substantially over the past 20 years, independent of age
and obesity [34]. OA is also the most common cause of chronic pain especially in older
adults [32]. This suggests that as the Australian population ages, the prevalence of OA-
Page 9
induced chronic pain will increase.
The determinants of pain and structural damage in OA are not well understood but are
believed to involve multiple interactive pathways [37]. OA pain is a mixed phenomenon
where nociceptive [38] and neuropathic [36, 39] mechanisms are involved in both the local
and central levels. OA pain perception is influenced by multiple environmental,
psychological, or constitutional factors, and OA pain intensity is not just determined by the
structural damage in the synovial joint [37]. Subjects with the same degree of structural
damage experience widely different levels of pain, a phenomenon that is poorly understood.
Whilst radiographic evidence of joint damage predisposes to joint pain, it is clear that the
relation of the severity of joint damage to the severity of the pain is not strong [8, 40].
However, using other imaging modalities such as magnetic resonance imaging (MRI),
numerous structural alterations have been related to knee pain. It remains unclear, which of
these local tissue factors predominate.
Constitutional factors that can predispose to symptoms including self-efficacy, pain

catastrophizing and the social context of the disease (social support, pain communication) are
all important considerations in understanding the pain experience [41]. The limitations
imposed by OA can be detrimental to a person's self-esteem and self-image and can lead to
negative emotional states, anxiety, depression and feelings of helplessness that can in turn
exacerbate the perception of symptoms [42]. There is also evidence that genetic factors may
also influence the pain perception pathways. Candidate genes have been identified which
could alter the processing of nociceptive pain associated with OA [43].
1.2.5 Risk factors
OA is a multifactorial disease and the development of the disease is dependent on interactions

between several factors. This process may be considered the product of interplay between
systemic and local factors [44]. Risk factors vary for different joints, for different stages of
disease, and for the development versus the progression of the disease.
Age remains one of the strongest risk factors for the development of knee OA [45]. Knee OA
can occur at any age but the relationship between age and the development of knee OA is
Page 10
non-linear. Population based studies have indicated a sharp increase in the incidence of knee
OA between the ages of 50 and 75 years [46] and a levelling off or a decline after the age of
80 years [47].
Obesity or being overweight is probably the most important risk factor for the development
and progression of not only knee OA but also at other sites such as hands and hip [45].
Mechanical forces exerted on the knee joint are a significant cause of OA and body mass
index (BMI) is one of the most important modifiable risk factors [48, 49]. Relationship
between obesity and the development of knee OA is mainly linear [48]. Obesity alone or in
patients with metabolic syndrome increases the risk of radiographic knee OA [50, 51]. One
study found that increasing from normal to overweight during adult life might give a slightly
higher risk of developing knee OA than being constantly overweight during adult life [49].
Another study found that among women at an elevated risk of OA due to high BMI, weight
loss decreased this risk substantially [52]. A meta-analysis showed that over-weight people
had higher odds (odds ratio (OR) 1.98 (95% CI 1.57-2.20) for developing knee OA and the
risk increased further in obese people (OR=2.66 (95% CI 2.15-3.28)) [49]. Moreover an
estimated 17.3-24.6% of new cases of knee pain are related to being overweight or obese
[45]. Another study estimated that 69% of the knee arthroplasties are attributed to obesity
[51].
Females are at a higher risk of developing knee OA (OR=1.68 (95% CI 1.37-2.07)) [45].
Prevalence of knee OA is roughly the same in males and females in younger adults but the
prevalence increases in females after menopause[53]. Twice as many females suffer from the
disease compared to males between the ages of 50-80 years [54]. These findings suggest a
role of sex hormones but the evidence supporting the role of specific sex hormones thus far
has been inconsistent [45].
History of joint injury is widely accepted as a contributory factor in the development of knee
OA. Major trauma to the knee has the potential for damage to any of the knee structures that
are important for joint homeostasis. Several epidemiological studies have shown this
association using radiography. In a prospective cohort study, Wilder et al. [55] showed that
individuals with a history of knee injury were 7.4 times more likely to develop knee ROA
than individuals who did not have a history of knee injury. A recent meta-analysis showed
Page 11
that history of knee injury was consistently associated with higher odds of developing knee
OA (OR= 2.83 (95% CI 1.91-4.19)) [45]. The same study also estimated that 5.1% of new
knee pain/OA patients could be attributed to a previous knee joint injury [51].
Systemic predisposition is also a potent risk factor for knee OA. Several studies have shown
that OA is often generalized and affects multiple joints. In a post-mortem bone study, Rogers
et al. [56] confirmed the hypothesis that OA is caused primarily by a systemic predisposition.
Hand OA, usually diagnosed clinically by the presence of Heberden's nodes, increases the
odds for the development of the knee OA (OR=1.30 (95% CI 0.90-1.87)) [45].
Other common risk factors mentioned in the literature include occupational risk factors,
physical activity and genetic factors. Occupational risk factors include work involving
kneeling/bending [57-59] and heavy lifting [58-60]. Similarly farmers and construction
workers are also at a higher risk of developing knee OA [61]. Evidence for physical activity
is rather inconsistent. Intense physical activities such as long-distance running have been
shown to increase the risk of developing knee OA [62-64]. Similarly Dore et al. [65] showed
that >10,000 steps/day are associated with worsening knee structures. However, Felson et al.
[66] and Øiestad et al. [67] did not see any association between objectively assessed physical
activity and progressive knee structural damage. There is strong evidence that genetic factors
play an important role in the development of OA of the hands and the spine [68, 69]. The
evidence is rather inconsistent for knee OA [68-72] and it has proven difficult to isolate
candidate genes [73].
1.2.6 Diagnosis
OA is diagnosed using a combination of clinical examination and imaging [74].
1.2.6.1 Clinical criteria
Knee OA is defined clinically using frequent knee symptoms. The American College of
Rheumatology (ACR) definition for symptomatic OA is most commonly used and defines
OA as “pain, aching, or stiffness in or around the knee on most days” for at least one month
Page 12
during the past 12 months [74]. This definition has been validated by several epidemiological
studies [75, 76]. This definition is however simplistic and does not take into account the
severity of pain. Several pain scores, such as the Visual Analog Score (VAS), the Knee injury
and Osteoarthritis Outcome Score (KOOS) [77] and the Western Ontario and McMasters
Universities Osteoarthritis Index (WOMAC) [78], have been developed for research
purpose’s and take into account the severity of knee pain and longitudinal changes in the
severity of the symptoms.
1.2.6.2 Imaging
Imaging studies have provided many insights, however much remains unknown and
uncertain. Imaging developments in OA are an important rate-limiting step to further
therapeutic development.
Radiographs remain the gold standard for diagnosing OA. Radiography enables the detection
of OA-associated bony features such as osteophytes, subchondral sclerosis, and cysts [79].
Radiographs can also determine joint space narrowing (JSN) [79] that is traditionally thought
to result from cartilage volume loss but may also reflect changes in meniscal integrity [31].
Although radiographs cannot directly assess any of the intra-articular soft tissue structures,
radiography is still the only method approved by Food and Drug Administration (FDA) for
monitoring disease progression in the Disease Modifying Osteoarthritis Drug (DMOAD)
Trials [80].
The severity of radiographic OA can be assessed with semi-quantitative scoring systems. The
Kellgren and Lawrence (KL) [81] grading system (Table 1.1) was the earliest proposed and
one of most widely used scoring system. KL grading system defines radiographic OA based
on the presence of a definite osteophyte (grade 2). However, KL grading has its limitations;
in particular, KL grade 3 includes all degrees of definite JSN, regardless of the extent.
Secondly it places too much emphasis on osteophytes.
The research conducted in this thesis used the atlas from the Osteoarthritis Research Society
International (OARSI) (Table 1.2), which was first published in 1995, by Altman et al. [82,
83]. The OARSI atlas provides grades for specific features of OA rather than global scores
Page 13
like the KL scheme. The atlas grades tibiofemoral JSN (Figure 1.2) and osteophytes (Figure
1.3) (on a 0-3 grade where 0=absent and 3=severe) separately for each compartment of the
knee (medial tibiofemoral, lateral tibiofemoral, and patellofemoral). The OARSI atlas
provides clinicians and researchers with a standardised semi-quantitative methodology for
radiographic features and has been validated in both cross-sectional and longitudinal studies.
Table 1.1. Kellgren and Lawrence (K/L) grading system
Grades Description
Grade 0: No osteoarthritis No osteoarthritis
Grade 1: Doubtful Doubtful narrowing of joint space and possible osteophytic

lipping
Grade 2: Mild Definite osteophytes and possible narrowing of joint space
Grade 3: Moderate Multiple osteophytes, definite narrowing of joint space and

some sclerosis and possible deformity of bone ends
Grade 4: Severe Large osteophyte, marked narrowing of joint space, severe
sclerosis and definite deformity of bone ends
Page 14
Table 1.2. Individual radiographic features measured by the OARSI atlas for tibiofemoral
osteoarthritis
Radiographic Features Score
Marginal osteophytes
Medial femoral condyle 0-3
Medial tibial plateau 0-3
Lateral femoral condyle 0-3
Lateral tibial plateau 0-3
Joint space narrowing
Medial compartment 0-3
Lateral compartment 0-3
Other
Medial tibial attrition Absent/present
Medial tibial sclerosis Absent/present
Lateral femoral sclerosis Absent/present
OARSI: Osteoarthritis Research Society International
Page 15
Figure 1.2. Radiographic joint space narrowing
Osteophytes
Figure 1.3. Radiographic osteophytes
Page 16
1.2.7 Radiography versus MRI
Radiographs are still the standard modality to diagnose OA and to monitor disease
progression in clinical trials [79]. Radiographs do have certain advantages [8, 79]. Firstly,
radiographs are cheaper compared to other imaging modalities and readily available.
Secondly radiographic scoring systems are standardised, which makes them easier to
implement in both clinical and large multicentre research settings.
Radiographs do have certain limitations that have seen the continued use of radiographs as an
outcome measure in OA research criticised. Radiographs have a low resolution and cannot
directly visualise the soft tissue structures that make up the knee joint such as cartilage,
meniscus and ligaments. Subject positioning and changes in the radio-anatomic alignment of
the knee joint in serial examinations influence the reproducibility of radiographs [31].
Radiography is also insensitive to early disease changes. Approximately 10% of the knee
articular cartilage is already lost before we observe any changes on the radiographs [84].
Similarly radiographs are also insensitive to small changes overtime [85]. Longitudinal
studies require a long follow-up period before we can observe any appreciable changes on
radiographs making the studies more time consuming and expensive. Lastly several studies
have shown that radiographic features such as JSN and osteophytes do not correlate well with
clinical symptoms. In population studies, there is a significant discordance between
radiographically diagnosed OA and knee pain and vice versa [8, 40, 86-88]. Although
radiographic evidence of joint damage predisposes to joint pain, it is clear that the relation of
the severity of joint damage to the severity of the pain is not strong. In a systematic review,
Bedson et al. [87] reported that the proportion of knee pain found to have ROA ranged from
15–76% and in those with ROA the proportion with pain ranged from 15–81%.
MRI has proven to be an important alternative imaging tool in OA research and has
revolutionised the understanding of OA pathology. The cross-sectional image display, spatial
resolution, and tissue contrast of MRI enables whole organ assessment of the knee joint [28].
MRI is ideally suited for imaging arthritic joints, as is it free of ionizing radiation and has the
ability to acquire morphological and biochemical data [28]. Advances in MRI technology
have improved its ability to detect bone marrow lesions, joint fluid changes, ligamentous and
Page 17
meniscal damage, osteophyte formation, cartilage morphology, as well as macromolecular

changes, which often precede morphological changes [8].
As mentioned earlier, there is a significant discordance between radiographically diagnosed

OA and knee symptoms. Radiographic disease predisposes to knee symptoms but the extent
of joint damage does not correlate well with the severity of knee pain [8]. Several population-
based studies using MRI have described associations between knee pain and knee structures
such cartilage defects [89, 90], BMLs [91-93], meniscal damage [75, 94] and
effusion/synovitis [7, 95]. Recent evidence has clearly shown that most of the soft tissue
changes in the knee joint OA causal pathway are collinear [31]. However it is not clear which
of these structures predominantly result in knee pain as most of the earlier studies did not
account for all the soft tissues that can result in symptoms. For example cartilage is an
aneural and an avascular structure and theoretically cartilage damage should not result in
symptoms unless the subchondral bone is denuded. However, several studies have shown an
association between cartilage defects/thickness loss and knee symptoms [89, 90]. These
positive associations are most likely due to the fact that these studies did not account for
global knee structural changes [96].
To develop new treatments to prevent the progression of OA, we need to understand what
subtle changes occur at different stages of the disease. Extensive use of MRI to study knee
OA pathophysiology has provided us with great insight into different soft tissues that can be
targeted. Structural abnormalities can be assessed either semi-quantitatively or quantitatively.
Several semi-quantitative scores such as Whole-Organ Magnetic Resonance Imaging Score
(WORMS) [97], Knee Osteoarthritis Scoring System (KOSS) [98] and Boston–Leeds
Osteoarthritis Knee Score (BLOKS) [92] have been proposed and have generally shown good
reliability and specificity. However, as with any ordinal scale, these semi-quantitative scores
lack sensitivity to change over time. Large-scale population based studies using MRI are
expensive and require long follow-up periods to have adequate sensitivity to detect change in
knee structures using ordinal scales. Most of the earlier studies using MRI to assess knee OA
were either cross-sectional or had short follow-up periods, which makes semi-quantitative
scales less sensitive to detect longitudinal progression.
Page 18
Several manual, semi-automated and automated image processing softwares have been
developed to quantitatively assess knee structural abnormalities. Quantitative assessment of
the knee articular cartilage is particularly important, as articular cartilage loss is often cited as
the structural hallmark of OA progression. Commonly used sequences for morphological MR
imaging include original and fast or three-dimensional (3D) variations of spin-echo (SE) and
gradient-recalled echo (GRE) and 3D dual-echo steady state (3D DESS). Cartilage
boundaries in different knee compartments (tibial, femoral and patellar) are usually
segmented on a slice-by-slice basis and then extrapolated to assess cartilage volume. The
measurement of cartilage volume from MRI has been shown to correlate well with the ex-
vivo assessments of cartilage volume [99, 100]. Quantitative assessment techniques for other
structural abnormalities such BML area, bone area, meniscal extrusion and knee effusion
area/volume have also shown to be reproducible and sensitive to change over time.
Currently no DMOADs are available to slow the progression or reverse structural damage
that results from OA. The biggest obstacle in developing an effective treatment modality is
the lack of suitable imaging modalities to monitor disease progression and serve as an end-
point in DMOAD trials. An ideal imaging modality should be able to visualize global knee
structural abnormalities, should be reproducible, should correlate well with symptoms, should
be able to detect earlier structural abnormalities before changes become irreversible and
should be sensitive to subtle changes over a short period of time. Despite the obvious
limitations, radiographs remain the current gold standard to diagnose knee OA and is the only
modality approved by Food and Drug Administration (FDA) to show efficacy of disease
modifying OA drugs in phase 3 clinical trials [79].
MRI has become a key imaging tool for OA research because of its ability to visualize
disease in structures not imaged by radiography, ability to monitor multiple tissue changes
simultaneously over several time points and the ability to detect physiologic changes within
joint tissues (eg, cartilage and menisci) before morphologic changes become apparent. Yet
there are several reasons why MRI is not the gold standard for diagnosing knee OA and
monitoring disease progression. Firstly there is a lack of clarity about the diagnostic
performance of different soft tissue abnormalities found on MRI scans as there is a lack of
long term studies looking at the predictive value of these early structural changes [101].
Secondly there is a lot of contradictory data available [101] looking at the association
Page 19
between structural changes and symptoms, as most of the earlier studies did not account for
global knee structural abnormalities. Thirdly there is a lack of standardisation due to the
absence of a MRI structural definition of OA [101]. Using a modified Delphi approach, a
panel of experts in the field have developed 11 propositions for a definition of knee OA on
MRI [101]. The goal of this exercise was to develop definitions of knee OA that can be more
formally tested in relation to their diagnostic performance in long-term studies. Lastly and
most importantly the majority of long-term studies have thus far focused on older aged
populations with established symptomatic disease. There is little long-term MRI data
available for knee structural changes especially in early disease.
1.2.8 Why study middle-aged adults?
Knee OA is a disease that mostly affects older adults and it made sense for earlier studies to
focus on older populations with advanced symptomatic disease. However once the disease is
established, there is little potential of any reversibility. It is important to detect early disease
changes in middle-aged groups with little or no radiographic changes and symptoms. This
would firstly allow us to study the natural history of soft tissue abnormalities and understand
how knee OA evolves. Studying the natural history of knee structural changes can help us
identify people who are fast-progressors, making recruitment for clinical trials more efficient
and treatments possibly more effective. Furthermore, studying middle-aged populations can
help us identify pathological abnormalities that trigger OA development and identify factors
that can possibly prevent, delay or slow down OA progression. We can also possibly identify
tissues with potential of regeneration before irreversible damage sets in. Targeting the right
populations and early structural abnormalities is imperative for the success of future
DMOAD trials.
1.3 Offspring Study
The Offspring study [102, 103] is a longitudinal controlled population-based study of middle-
aged adults (more detailed description of the Offspring study is provided in the Methods
Page 20
section). The Offspring study began in Southern Tasmania in June 2000. Half of the
participants were the adult offspring of patients who had a knee replacement performed for
idiopathic knee OA and the other half were age and sex matched controls without a family
history of knee OA. First follow-up was after two years and the second follow-up was after
ten years (this thesis mainly used the ten-year follow-up data). The Offspring study was one
of the first large scale population based longitudinal studies to use MRI data to investigate
knee OA, the first major MRI based knee OA study in middle aged adults and currently the
longest follow-up knee OA study with MRI data.
1.3.1 Evidence from earlier phases of Offspring study
Offspring/control differences
Cross-sectional baseline data showed that compared to controls, offspring were significantly
heavier, had a higher prevalence of knee pain, weaker lower limb muscles, a larger tibial
bone area and a higher prevalence of cartilage defects [104]. There was no significant
difference between the two groups for the prevalence of ROA and mean cartilage volume
[104]. Similarly cartilage volume, muscle strength, bone area and BMLs showed high
heritability in sib-pair analysis [70].
Pain/symptoms
Cross-sectional data from the baseline visit of the Offspring study showed that knee pain was
fairly common in middle-aged adults with a prevalence of 35% [105]. However the severity
of knee pain was low [105]. Knee pain was independently associated with non-full thickness
chondral defects (particularly femoral and patellar), osteophytes, urinary C-terminal cross-
linking telopeptide of type II collagen (CTX-II) and obesity [105]. Separate studies from the
first follow-up of the Offspring study showed that the prevalence of both BMLs [106] and
meniscal tears [94] were significantly associated with knee pain, however the magnitude of
the associations remained low.
Page 21
Cartilage defects and cartilage volume loss
Cartilage defects were very common with a prevalence of 44% [89]. Presence of cartilage
defects was associated with low cartilage volume (in a dose response manner) and with
urinary CTX-II [89]. Another important finding was that cartilage volume loss started about
the age of 40 in most people and the rate of loss was about 2% per annum and increased with
age, consistent with the natural history OA [107]. Cartilage defects at baseline also
independently predicted cartilage volume loss over 2 years [108]. Natural history of cartilage
defects was quite variable [89]. Approximately one-third of the cartilage defects increased in
severity but interestingly a similar number decreased in severity or completely resolved over
2 years. These findings suggest that OA structural damage is not irreversible especially in
younger adults.
BMLs
A cross-sectional study from first follow-up of the Offspring study was the first population-
based study to examine the risk factors for the development of BMLs [106]. BMLs were
fairly common in middle aged adults (24% in the medial tibiofemoral compartment and 14%
in the lateral tibiofemoral compartment). Besides genetic factors, presence of BMLs was
positively associated male gender, BMI, cartilage defects and ROA.
Meniscus
Meniscal tears are generally thought to result from mechanical factors such as knee
injury/surgery. A cross-sectional study [94] from the Offspring study was one of the earliest
studies to suggest that meniscal tears share common risk factors with OA. Meniscal tears
were not only associated with symptoms but also with cartilage defects, cartilage volume and
ROA. In summary, earlier visits of the Offspring study provided valuable cross-sectional and
short term MRI data. However long-term data was required to better understand the natural
history of knee structural changes, correlation between structural changes and symptoms and
validate MRI changes against long-term radiographic changes.
Page 22
1.4 Summary
About 8% of Australians are affected by OA. By 2050, it is projected that the prevalence of
OA will be 11% of the population. OA is a leading source of health expenditure on arthritis,
accounting just under half of total allocated expenditure on arthritis in 2007. There was a
54% rise in total knee replacements for OA from 2002-03 to 2011-12. There are no disease-
modifying treatments available for OA. There is an urgent need for identifying modifiable
risk factors for this disease and for understanding early structural changes. Identifying these
factors before knee structural damage becomes irreversible can delay or may even prevent the
development of knee OA later in life. The following Chapters describe long-term knee
structural natural history data in middle-aged adults and the correlation with frequent knee
symptoms. Ten year follow-up data from the Offspring study was used to explore these
questions. The specific research questions that directed this work are described in the
following Chapter.
Page 23
Chapter 2 Research Questions
Chapter Two
Research Questions
Page 24
Chapter 2 Research Questions
2.1 Research Questions
A population-based sample of middle-aged adults, with an offspring-control design, was

examined at baseline and approximately 2 and 10 years later:
1. What the cross-sectional association between history of knee injury and knee
structural damage assessed on MRI in middle-aged adults from the Offspring study
and how does that compare to a random community based sample of older adults.
2. Does a family history of knee joint replacement due to OA increase the risk of
radiographic OA progression and cartilage volume loss over 10 years?
3. What is the long-term natural history of BMLs in middle-aged adults and how are
these changes associated with symptoms and other knee structural changes?
4. What is the natural history of meniscal tears in middle-aged adults and how are these
changes associated with symptoms and structural changes?
5. What is the long-term natural history of cartilage defects, do cartilage defects predict
cartilage volume loss and how changes in cartilage defects are associated with
changes in symptoms and structure?
6. How do changes on MRI correlate with changes in radiographs and whether
radiographic JSN should be used as an end-point in chondro=protective drug trials?
2.2 Key Hypothesis
1. Family history of OA influences structural progression and symptoms associated with

knee OA
2. Studying the long-term natural history in middle-aged will help identify reversible
structural changes which can potentially be targeted in clinical drug trials.
Page 25
Chapter 3 Methodology
Chapter Three
Methodology
Page 26
3.1 Prelude
This thesis arose from analyses of the Offspring study population, and a number of outcome
factors, study factors, and covariates have been utilised. This chapter describes the Offspring
study population and its design, as well as the protocols for measurement of factors that are
common to multiple chapters in this thesis. Additional factors, which are unique to each
chapter, are described in more detail within the methodology section of each of the
subsequent chapters.
It should be noted that the following chapters are presented in the form in which they were
submitted to, or accepted by, peer-reviewed journals for publication. Thus, throughout these
chapters there are some differences in the description of methods, analyses, results, and
interpretations, due chiefly to requests from journal reviewers.
3.2 Study population and design
The work in this thesis was conducted as part of the Offspring study, a population-based
study with an Offspring-control design, aimed at identifying the environmental, genetic, and
biochemical factors associated with the development and progression of knee osteoarthritis
(OA). The Offspring study began in Southern Tasmania (primarily in the city of Hobart) in
June 2000. Matched sampling was used to recruit the study participants (mean age 45(26–61)
years; 58% females) [102]. Half of the participants were the adult offspring of patients who
had a knee replacement performed for idiopathic knee OA at any Hobart hospital from 1996
to 2000. The diagnosis was confirmed by reference to the medical records of the orthopaedic
surgeon and the original radiographs when possible. The other half were age and sex matched
controls, randomly selected from the population (using electoral rolls) with no history of knee
OA in either parent. Electoral rolls represent the most complete population information
available in Australia because voting in federal and state elections is compulsory. This thesis
includes data from visit-1 (2000-01), visit-2 (2002-03) and visit-3 (2010-11) of the study.
Participants were excluded if they had a contraindication to MRI (including metal sutures,
presence of shrapnel, iron filing in eye, or claustrophobia). Participants were also excluded if
Page 27
they had undergone a knee replacement surgery or did so after the commencement of the
study. Knee pain and knee injury were not a basis for exclusion.
Figure 3.1 provides an overview of participant recruitment and withdrawal during the study
period. Offspring study started with 372 participants (Offspring=186 and Controls=186) in
year 2000. 326 (88%) participants (Offspring= 162 and Controls= 164) were followed up
after approximately 2 years, whereas 220 (59%) participants (Offspring= 115 and Controls=
105) were followed up after approximately 10 years.
Figure 3.1. Flow chart Offspring study
Page 28
Table 3.1 summarises the baseline characteristics of those participants who completed the
follow-up (n=220) and those which did not (n=152). There was no significant difference in
age, sex, BMI, proportion of Offspring (%) and prevalence of ROA and knee pain between
the two groups.
The sample size used in the following chapters of this thesis varies depending on the
available data for each of the research questions.
lost to follow up
Characteristic Follow-up Loss Follow-up P-value

(n = 220) (n = 152)
Age (years) 45.3 ± 6.7 45.1 ± 7.2 0.806
Female (%) 58 59 0.749*
BMI (kg/m2) 27.2 ± 4.9 26.8 ± 4.3 0.499
Offspring (%) 52 47 0.891*
Radiographic OA (%) 18 15 0.486*
Knee pain present (%) 33 34 0.917*
Mean ± standard deviation except for percentages; *Determined by Chi square test, others by
t-test
Page 29
3.3 Anthropometrics
Weight was measured to the nearest 0.1 kg (with the subject’s shoes, socks, and bulky
clothing removed), with a single pair of electronic scales (Delta Model 707; Seca, Munich,
Germany) that were calibrated using a known weight at the beginning of each clinic session.
Height was measured to the nearest 0.1 cm (with shoes and socks removed) using a
stadiometer. Body mass index (BMI) was calculated as weight (kg)/height (m2).
3.4 Leg strength
Muscle strength was measured by dynamometry at the lower limb (involving both legs
simultaneously) at all three visits. This primarily involves the hip flexors and knee extensors.
The participants were instructed in each technique prior to testing, and each measure was
performed twice. The repeatability estimate (Cronbach’s alpha) was 0.9 [70]. The device was
calibrated by suspending known weights at regular intervals.
3.5 Knee joint injury and surgery
History of knee joint injury and surgery were assessed using a self-administered
questionnaire which included the following questions and was identical in both cohorts:
“Have you ever had a previous knee injury which resulted in non-weight bearing treatment
for 24 hours or more?”
“If yes, then which knee?”
“Please provide further details about the injury”
“Have you ever had a knee surgery?”
“Please provide further details about the surgery”
Page 30
3.6 Knee pain
Knee pain was assessed at visit-1 using an interviewer-administered questionnaire as

described previously [102]. All the participants were asked the following question:
Have you had knee pain for more than 24 hours in the last 12 months or daily pain on greater
than 30 days in the last year?
Knee pain was assessed by self-administered questionnaire using the Western Ontario and
McMaster Universities Osteoarthritis Index (WOMAC) [93] at visits 2 and 3. Five categories
of pain (walking on flat surface, going up or down stairs, at night, sitting or lying, and
standing upright) were assessed separately with a 10-point scale from 0 (no pain) to 10 (most
severe pain). Each category was summed to create a total pain score (range 0 to 50).
Furthermore, the five categories were clinically categorized into weight-bearing pain
(including walking on flat surface, going up or down stairs and standing) and non-weight-
bearing pain (including pain at night and sitting or lying).
3.7 Magnetic Resonance Imaging
MRI of the right knee was acquired at all three visits with a 1.5T whole-body magnetic
resonance unit (Picker, Cleveland, OH, USA) using a commercial transmit/receive extremity
coil [89, 104, 108]. The following image sequences were used: (1) a T1-weighted fat-
suppressed 3D gradient-recalled acquisition in the steady state, flip angle 55°, repetition time
58 msec, echo time 12 msec, field of view 16 cm, 60 partitions, 512×512–pixel matrix, slice
thickness of 1.5 mm without an interslice-gap (at all three visits); and (2) a T2-weighted fat
saturation 2D fast spin echo, flip angle 90°, repetition time 3067 ms, echo time 112 ms, field
of view 16 cm, 15 partitions, 256×256 matrix, slice thickness of 4 mm with an interslice gap
of 0.5–1.0 mm (at visit 2 and 3).
The same scanner (same model and machine) was used at all the three visits for both T1-
weighted fat-suppressed and T2-weighted fat saturation images.
Page 31
3.7.1 Cartilage volume
Tibial and patellar cartilage volume was assessed at all three time points, by a trained
observer on T1-weighted MR images, using Osiris (University of Geneva, Geneva,
Switzerland) software as previously described [104, 109]. The image data were transferred to
the workstation. The volumes of individual cartilage plates (medial tibial, lateral tibial and
patella) were isolated from the total volume by manually drawing disarticulation contours
around the cartilage boundaries on a section-by-section basis. These data were then
resampled by means of bilinear and cubic interpolation (area of 312 and 312 µm and 1.5mm
thickness, continuous sections) for the final 3D rendering using Osiris software. The
coefficient of variation (CV) for both studies ranged between 2.1–2.2% for intra-observer
repeatability [84]. Femoral cartilage volume was determined using Cartiscope (ArthroLab,
Montreal, Quebec, Canada), as previously described [110, 111]. First, using a semi-
automated image processing (segmentation), the whole cartilage geometry is extracted from
MR sagittal range images (Figure 3.2). For each sagittal image in the volume data set, semi-
automatic delineation is performed using an active-contour-segmentation technique. These
initial contour lines are then automatically adjusted by using a 2D/3D; active-contour process
(snake) to more closely fit the cartilage margins with sub-pixel accuracy, as already described
[111, 112]. The CV for both studies was approximately 2% for intra-observer and inter-scan
repeatability [111, 112].
Figure 3.2. Cartilage volume segmentation
Page 32
3.7.2 Cartilage defects
Cartilage defects were assessed on T1-weighted gradient echo MR images at all three visits.
Cartilage defects were graded at the medial tibial, medial femoral, lateral tibial, lateral
femoral and patellar sites on a 0-4 scale (grade-0=normal cartilage; grade-1=focal blistering
and intra-cartilaginous low-signal intensity area with an intact surface and bottom; grade-
2=irregularities on the surface or bottom and loss of thickness of less than 50%; grade-
3=deep ulceration with loss of thickness of more than 50%; grade-4=full-thickness chondral
wear with exposure of sub-chondral bone), as previously described [113] (Figure 3.3). A
cartilage defect also had to be present in at least two consecutive slices. If multiple defects
existed at one site, the highest grade was used. Intra-observer reliability (expressed as intra-
class correlation coefficient (ICC)) ranged from 0.89-0.90. Inter-observer reliability was
assessed in 50 MR images and yielded an ICC of 0.85-0.90.
Figure 3.3. Grade-2 cartilage defect
3.7.3 Bone marrow lesions
Subchondral BMLs were assessed at visit 2 and 3, using Osiris software (University of
Geneva, Geneva, Switzerland) and were defined as areas of increased signal adjacent to the
Page 33
subcortical bone at the medial tibial, medial femoral, lateral tibial, lateral femoral, superior
patella and inferior patella sites[93] (Figure 3.4). One trained observer scored the BMLs by
measuring the maximum area (cm2) of the lesion at both time points. The observer manually
selected the MRI slice with the greatest BML size. The BML with the highest score was used
if more than one lesion was present at the same site. MRIs at both time points were read
paired with the chronological order known to the observer and the observer blinded to clinical
status. Intraobserver repeatability was assessed in 40 subjects with at least a two-week
interval between the readings. The ICC was 0.97. Participants were given a BML score (cm2)
for each of the six sites (medial tibial, medial femoral, lateral tibial, lateral femoral, superior
patella and inferior patella sites).
Figure 3.4. Bone marrow lesion
3.7.4 Meniscal tears
Meniscal tears were assessed by a trained observer (musculoskeletal radiologist with several
years of experience) on T2-weighted fat saturated (side by side) MR images at visit-2 and 3
of the study as previously described [110] (Figure 3.5). The proportion of the menisci
Page 34
affected by a tear was scored separately (0-2 scale; 0=absence of a tear, 1=simple tear of
different types: longitudinal, oblique, radial or horizontal, 2=macerated tear signifying
loss>50% area of meniscal tissue) at the anterior, middle, and posterior horns. Anterior,
middle and posterior scores were summed to create medial and lateral meniscal tear scores.
The intra- and inter-observer correlation coefficient (expressed as ICC) ranged from 0.86 to
0.96 [111].
Meniscal tears were initially scored cross-sectionally on T-1 weighted MR images using a
different protocol at Visit-2 of the study. More details of the exact protocol are provided in
Chapter 4.
Figure 3.5. A). Normal meniscus. B) Meniscal tear
3.7.5 Meniscal extrusion
The extent of meniscal extrusion on the medial or lateral edges of the tibial femoral joint
space, not including the osteophytes, was evaluated at visit-2 and 3 for the anterior, body, and
posterior horns of the menisci on T1-weighted gradient echo MR images, as previously
described[111]. A score from 0 to 2 was used (0= no extrusion, 1= partial meniscal extrusion,
and 2= complete meniscal extrusion with no contact with the joint space). The scores of
anterior, body and posterior horns of medial or lateral menisci were summed to create a total
meniscal extrusion score for each of the medial and lateral tibiofemoral compartments which
Page 35
had a possible range from 0 to 6. The intra- and inter-observer correlation coefficient
(expressed as ICC) ranged from 0.85 to 0.92 for meniscal extrusion [110].
3.7.6 Tibial bone area
Knee tibial plateau bone area was assessed on T1-weighted MR images at visits 1 and 2, and
determined by means of image processing on an independent workstation using the software
program Osiris (University of Geneva) as previously described [104]. To transform the
images from the sagittal plane to the axial plane, the Analyse Software package developed by
the Mayo Clinic was employed. Medial and lateral tibial plateau bone area was determined
by creating an isotropic volume from the three input images closest to the knee joint (Figure
3.6). The bone area of the medial and lateral tibial plateau was then directly measured from
the reformatted axial images. The CV was 2.2–2.6% for intra- observer repeatability[104].
Figure 3.6. Medial and lateral tibial bone area
Page 36
3.7.7 Effusion
Effusion was assessed in the supra-patellar pouch on T2-weighted fat saturated MR images at
visits 2 and 3 on a 0-3 scale[114]. Grade-0 signified absence of fluid over the upper margin of
the patella in a sagittal image; Grade-1 signified some fluid above the upper margin of the
patella but the length of the fluid column shorter than that of the patella; Grade-2 signified a
fluid column above the upper margin of patella longer than the length of the patella; Grade-3
signified a fluid column above the upper margin of patella longer than the length of the
patella with a thickness of ≥ 1cm. Intra-observer reliability was assessed in 50 MR images
and yielded an ICC of 0.89-0.98. Pathological effusion was defined as any effusion score ≥2.
3.8 Radiology
A standing anteroposterior semiflexed x-ray of the right knee was taken in all subjects at
visits 1 and 3. The angle was kept to 10–15˚ by a purpose built goniometer. The tube to film
and tube to tibial plateau angle was 90˚. Daily quality assurance was performed on the
equipment. Radiographs were scored individually for osteophytes and joint space narrowing
(JSN), as described previously [84] (Figures 3.7, 3.8). Each of the following four features
was scored on a scale from 0 to 3 (0 = normal and 3 = severe): medial JSN, lateral JSN,
medial osteophytes (femoral and tibial combined) and lateral osteophytes (femoral and tibial
combined). Each score was arrived at by consensus with two readers simultaneously
assessing the radiograph with immediate reference to the Osteoarthritis Research Society
International (OARSI) atlas [82]. A non-zero score in either JSN or osteophytosis was
regarded as evidence of ROA. Reproducibility was assessed in 50 radiographs, two weeks
apart, and yielded an ICC of 0.99 for osteophytes and 0.98 for JSN.
Page 37
Figure 3.7. Joint space narrowing
Figure 3.8. Osteophyte
3.9 Summary of outcome factors, study factors and covariates
Table 3.2 summarises the variables used in each chapter of this thesis.
Page 38
Table 3.2. Summary of outcome factors, study factors, and covariates used in this thesis
Chapter Outcome factors Study factors Covariates
4 Cartilage volume, cartilage defects, BMLs, Knee Injury Age, sex, BMI, family history of OA
Meniscal tears/extrusion, bone area
5 Cartilage volume loss, changes in JSN and Family history of OA Age, sex, BMI, knee pain, bone area,
osteophytes cartilage defects and muscle strength
6 Change in BMLs Family history of OA, knee pain, BMI, Age, cartilage defects, meniscal tears
sex, physical activity
7 Change in meniscal tears Knee injury, Family history of OA, Age, sex
knee pain, cartilage volume, BMLs
8 Change in cartilage defects, cartilage volume Family history of OA, knee pain Age, sex, meniscal tears, meniscal
loss extrusion, BMLs
9 Changes in JSN and osteophytes Cartilage volume loss, changes in Age, sex, BMI, family history of OA
cartilage defects, meniscal tears and
extrusion
Page 39
3.10 Sample size and role of the candidate in the Offspring study
As the Offspring study was in progress before commencement of the PhD candidature formal
sample size calculations were not performed during the design of this thesis. Therefore,
participant numbers in the analyses reported in this thesis were limited to the numbers
recruited at baseline and follow-up, and to those who provided complete data for relevant
outcome and study factors. As such, sample sizes vary between chapters, and the reasons for
exclusion are described in each chapter. Nevertheless, it subsequently proved that sample
sizes were more than adequate to answer the thesis research questions as this thesis has
reported significant findings related to the research questions.
The candidate was involved in Offspring study data acquisition and collection, data
management, analysis and interpretation of data, initial manuscript preparation and
manuscript revision. Data acquisition was also completed prior to and during the candidature
by a number
of other Offspring study staff members and volunteers, including Graeme Jones, Changhai
Ding, Dawn Aitken, Flavia Cicuttini, Jean-Pierre Pelletier and Johanne Martel-Pelletier.
Several colleagues had also begun analyses using Offspring study data before candidature
was undertaken, and the candidate gratefully acknowledges their contribution.
3.11 Ethical considerations
All procedures followed were in accordance with the ethical standards of the responsible
committee on human experimentation (institutional and national) and with the Helsinki
Declaration of 1975, as revised in 2000. The Southern Tasmanian Health and Medical
Human Research Ethics Committee approved the protocol, and written informed consent was
obtained from all participants.
Page 40
3.12 Statistical analysis
T-tests and chi-squared tests were used to compare differences in means and proportions as
appropriate. A P value less than 0.05 (two-tailed) was considered statistically significant. A
more detailed description of statistical analyses performed is presented in their relevant
chapters. All statistical analyses were performed on Intercooled Stata V.12.0 for windows
(StataCorp LP).
Page 41
Chapter 4 Injury and MRI-Assessed Knee Structure
Chapter Four
History of Knee Injury and MRI-Assessed Knee

Structures in Middle- and Older-Aged Adults: A
Cross-Sectional Study
Page 42
4.1 Introduction
Knee osteoarthritis (OA) is a major public health problem and a major cause of pain and
disability in older people [115]. Although the pathogenesis and etiology of OA is not fully
understood, it is regarded as an active process involving the whole joint [116].
History of joint injury is widely accepted as a contributory factor in the development of knee
OA. There is some evidence to suggest that joint injury may lead to OA of other joints as
well but it is less consistent compared to knee joint [117, 118]. Major trauma to the knee has
the potential for damage to any of the knee structures which are important for joint
homeostasis [55]. Several epidemiological studies have shown this association using
radiography. Kellgren and Lawrence demonstrated the relation between joint injury and
radiographic osteoarthritis (ROA) of the knee as early as 1958 [119]. Since then several
cross-sectional [120] and case-control studies [121, 122] have confirmed the association
between knee injury and subsequent ROA. In a prospective cohort study, Wilder et al.
showed that individuals with a history of knee injury were 7.4 times more likely to
develop knee ROA than individuals who did not have a history of knee injury [55]. More
recently Toivanen et al. showed similar results in a 22 year follow-up study [123].
All the studies mentioned earlier have used radiography to assess OA. While radiography is
the current gold standard for the diagnosis of OA, magnetic resonance imaging (MRI) is
being increasingly used to study OA as it allows visualisation of the whole joint [124].
However, a lack of standard MRI criteria for OA has hampered its use in regular clinical
practice and research. In a recent systematic review, leading experts defined OA criteria on
MRI using a Delphi voting exercise [101]. Osteophytes, bone marrow lesions (BMLs),
cartilage defects and meniscal tears were included in the criteria to define OA using MRI.
Other structures such as cartilage loss and increased tibial bone area were not included in the
criteria but commonly result from the above factors and have also been shown to predict total
knee replacement (TKR) surgery [125, 126].
Despite the abundance of available literature showing the association between history of knee
injury and ROA, there are still some deficiencies in our understanding of the exact causal
relationship. Firstly, it’s not clear whether the injury to the knee joint or the subsequent
surgery leads to the development of secondary OA, as the available literature often refers to
Page 43
knee trauma without a clear distinction between the two [127, 128]. Secondly, it is not known
which structures comprising the knee joint, as determined by MRI, are affected by injury and
hence contribute towards the progression of the disease process.
The aim of this study, therefore, was to describe the cross-sectional association between
history of knee injury and knee structures using MRI in a population-based cohort of middle-
aged subjects and a randomly selected cohort of older subjects.
Page 44
4.2 Methods
4.2.1 Subjects
This study was conducted as part of the Offspring Study [113] and the Tasmanian Older
Adult Cohort (TASOAC) Study [129]; details for both studies have been published
previously.
The Offspring study is an ongoing population-based study. This study includes data from
Phase 1 of the Offspring study, which was carried out in southern Tasmania (primarily in the
capital city of Hobart), between June 2000 and December 2001. Matched sampling was used
to recruit the study participants (mean age 45 years, range 26–61; 58% females). Half of the
participants were the adult children of patients who had had a knee replacement performed
for idiopathic knee OA at any Hobart hospital between 1996 to 2000. Controls were
randomly selected and matched by age and sex.
The Tasmanian Older Adult Cohort (TASOAC) study is an ongoing, prospective, population-
based study aimed to examine OA progression. Men and women aged 50–80 years in 2002
were selected from the electoral roll in Southern Tasmania (population 229,000) using sex-
stratified simple random sampling without replacement (response rate 57%). This study
includes data from the first follow-up which was carried out after 2.6 years (mean age 63
years, age range 51-79, females 51 %) as the questionnaire at the baseline visit did not assess
a history of knee injury.
Participants were excluded from both studies if they had a contraindication to MRI (including
metal sutures, presence of shrapnel, iron filing in eye, or claustrophobia) or were
institutionalized. The studies were conducted in compliance with the Declaration of Helsinki
and was approved by the Southern Tasmanian Health and Medical Human Research Ethics
Committee. Written informed consent was obtained from all participants.
Page 45
4.2.2 Anthropometrics
Height was measured to the nearest 0.1 cm (with shoes and socks removed), using a
4.2.3 Knee joint injury and surgery
History of knee joint injury and surgery were assessed using a self-administered
questionnaire which included the following questions and was identical in both cohorts:
“Have you ever had a previous knee injury which resulted in non-weight bearing treatment
for 24 hours or more?”
“Please provide further details about the injury”
“Have you ever had a knee surgery?”
“Please provide further details about the surgery”

Only right knee injuries were included in the analysis as MRI scans were on the right knee.
4.2.4 MRI
MRI scans of the right knee were performed in both studies. The following protocol was used
in each study:
Page 46
4.2.5 TASOAC
All knees were imaged in the sagittal plane on a 1.5T whole-body magnetic resonance unit
(Picker, Cleveland, Ohio, USA). The following sequence and parameters were used: (1) a T1-
weighted fat saturation three–dimensional (3D) gradient-recalled acquisition in the steady
state, flip angle 30°, repetition time 31 ms, echo time 6.71 ms, field of view 16 cm, 60
partitions, 512×512–pixel matrix, slice thickness of 1.5 mm without an interslice gap; and (2)
a T2-weighted fat saturation 2D fast spin echo, flip angle 90°, repetition time 3067 ms, echo
time 112 ms, field of view 16 cm, 15 partitions, 228×256–pixel matrix, slice thickness of 4
mm with a interslice gap of 0.5–1.0 mm.
4.2.6 Offspring
All knees were imaged in the sagittal plane on a 1.5T whole-body MR unit (Signa Advantage
HiSpeed; GE Medical Systems, Milwaukee, WI) using a commercial transmit–receive
extremity coil. The following sequence and parameters were used: (1) a T1-weighted fat-
thickness of 1.5 mm without an interslice gap; and (2) a T2-weighted fat saturation 2D fast
spin echo, flip angle 90°, repetition time 3067 ms, echo time 112 ms, field of view 16 cm, 15
partitions, 256 × 256 matrix, slice thickness of 4 mm with an interslice gap of 0.5–1.0 mm.
Tibial cartilage volume was assessed by a trained observer on T1-weighted MR images using
Osiris (University of Geneva, Geneva, Switzerland) software as previously described [104].
The image data were transferred to the workstation. The volumes of individual cartilage
plates (medial tibial, lateral tibial and patella) were isolated from the total volume by
manually drawing disarticulation contours around the cartilage boundaries on a section by
section basis. These data were then resampled by means of bilinear and cubic interpolation
(area of 312 and 312 µm and 1.5mm thickness, continuous sections) for the final 3D
Page 47
rendering using Osiris software. The coefficient of variation (CV) for both studies ranged
between 2.1–2.2% for intra-observer repeatability [84, 104]. Femoral cartilage volume was
determined using Cartiscope (ArthroLab, Montreal, Quebec, Canada), as previously
described [110, 112, 130]. First, using a semi-automated image-processing (segmentation),
the whole cartilage geometry is extracted from MR sagittal range images. For each sagittal
image in the volume data set, semi-automatic delineation is performed using an active-
contour-segmentation technique. These initial contour lines are then automatically adjusted
by using a 2D/3D, active-contour process (snake) to more closely fit the cartilage margins
with sub-pixel accuracy, as already described [131]. The CV for both studies was
approximately 2% for intra-observer and inter-scan repeatability [130].
Cartilage defects were assessed on T1-weighted MR images at the medial tibial, medial
femoral, lateral tibial, and lateral femoral sites, as previously described [113] as follows:
grade 0=normal cartilage; grade 1=focal blistering and intracartilaginous low-signal intensity
area with an intact surface and base; grade 2=irregularities on the surface or base and loss of
thickness <50%; grade 3=deep ulceration with loss of thickness >50%; and grade 4=full-
thickness chondral wear with exposure of subchondral bone. For the purpose of analysis in
this study, cartilage defects were used as a dichotomous variable signifying presence or
absence of any cartilage defect at a given site. Intraobserver reliability (expressed as
intraclass correlation coefficient (ICC)) was 0.90 for the medial tibiofemoral compartment
and 0.89 for the lateral tibiofemoral compartment for both studies. Interobserver reliability
was assessed in 50 MR images and yielded an ICC of 0.90 for the medial tibiofemoral
compartment and 0.85 for the lateral tibiofemoral compartment [113].
BMLs were assessed on T2-weighted MR images and defined as areas of increased signal
adjacent to the subcortical bone at the medial tibial, medial femoral, lateral tibial, and lateral
femoral sites. For the purpose of analysis, BMLs were used as a dichotomous variable
Page 48
signifying the presence or absence of any BMLs regardless of the site. The ICC was 0.97 for
intra-observer repeatability in TASOAC [93] and ranged from 0.89 to 1.00 in the Offspring
study [106].
4.2.10 Tibial bone area
Knee tibial plateau bone area was assessed on T1-weighted MR images and determined by
means of image processing on an independent workstation using the software program Osiris
(University of Geneva) as previously described [84, 89, 132]. To transform the images from
the sagittal plane to the axial plane, the Analyse Software package developed by the Mayo
Clinic was employed. Medial and lateral tibial plateau bone area was determined by creating
an isotropic volume from the three input images closest to the knee joint. The bone area of
the medial and lateral tibial plateau was then directly measured from the reformatted axial
images. The CV was 2.2–2.6% for intra- observer repeatability [84].
4.2.11 Meniscal damage
Meniscal damage was assessed by a trained observer on T1-weighted MR images as

previously described [110]. The proportion of the menisci affected by a tear, partial or full
extrusion was scored separately (yes/no) at the anterior, middle, and posterior horns
(medially/laterally). Anterior, middle and posterior scores were summed to get medial and
lateral meniscal tear/extrusion scores. For the purpose of analysis, a dichotomous score was
used which signified the presence or absence of any tear or extrusion at specific sites. The
intra- and inter-observer correlation coefficient ranged from 0.86 to 0.96 for meniscal tear
and 0.85 to 0.92 for meniscal extrusion [111].
4.2.12 Data analysis
Descriptive statistics of characteristics of the sample were tabulated. Regression analyses

were used to examine the association between history of knee injury and each knee structure.
Cartilage volume and tibial bone area were analysed as continuous variables whereas
Page 49
cartilage defects, bone marrow lesions and meniscal pathology were analysed as
dichotomised variables. Both continuous measures were analysed using linear regression
analysis whereas the three dichotomised variables were analysed using log binomial
regression analysis. Mean values of continuous variables and percentages (with exact
numbers in the form of fractions) are provided for both injured and non-injured group in all
the tables. Differences of means (DM) and prevalence ratios (PR) were used to express
continuous and dichotomised variables respectively. Analyses were adjusted for age, sex,
height and weight. Further adjustment for knee surgery was also performed. Offspring-
control interactions in Offspring study and sex interactions in both studies were explored for
all the associations between history of knee injury and knee structures. Additional analysis
examined the association between history of knee surgery and each knee structure.
A p value less than 0.05 (two-tailed) was considered statistically significant. All statistical
analyses were performed on Intercooled Stata V.12.0 for windows (StataCorp LP).
Page 50
4.3 Results
A total of 802 participants (Offspring =372; TASOAC=430) were included in this study.
Table 4.1 describes the characteristics of the two study populations.
Table 4.1. Characteristics of the study participants
OFFSPRING TASOAC
No of participants 372 430
Female 58% 51%
Age 45 (26-61) 63 (51-79)
BMI 27.0 (4.7) 27.8 (4.8)
Any joint space narrowing 14% 59%
Any osteophytes 7% 10%
History of knee injury 19% 12%
History of knee surgery 6% 12%
Participants in TASOAC study were older compared to the Offspring study. TASOAC had
approximately the same proportion of male and female participants, whereas the Offspring
study had a higher proportion of female participants. BMI in both cohorts was in the
overweight range and as expected the older cohort had a higher percentage of participants
with JSN and osteophytes. Participants in the Offspring study, despite being younger, had a
higher proportion report a history of knee injury.
Table 4.2 describes the association between knee injury and cartilage volume. In the
Offspring study, participants with a history of knee injury had higher mean cartilage volume
compared to participants in the non-injured group. However, after adjustment for age, sex,
height and weight, the association did not persist. In TASOAC, participants in the injured
Page 51
group had a lower adjusted knee cartilage volume but the difference was statistically
significant at the lateral tibial and the total tibial sites only.
Table 4.2. Association between injury and cartilage volume in the knee
Cartilage Volume Injury No Injury Unadjusted Adjusted*

(site) Mean Vol (mm3) Mean Vol (mm3) DM (95%CI) DM (95%CI)
Offspring
Medial Tibial 2436 2184 +252 (108,396) +54 (-63,+170)
Lateral Tibial 2828 2553 +278 (106,450) +12 (-115,+138)
Total Tibial 5264 4734 +530 (231,829) +67 (-152,+286)
Medial Femoral 4892 4500 +391 (10,772) +79 (-211,+369)
Lateral Femoral 5142 4649 +492 (108,877) +125 (-125,+375)
Total Femoral 10034 9150 +884 (162,1605) +204 (-275,+683)
Total Knee 15205 13925 +1280 (228,2332) +298 (-391,+987)
TASOAC
Medial Tibial 2108 2149 -41 (-221,+139) -59 (-198,+79)
Lateral Tibial 2402 2627 -224 (-463,+15) -265 (-439,-92)
Total Tibial 4510 4776 -265(-657,+199) -325 (-600,-51)
Medial Femoral 3882 3857 +26 (-407,+459) -93 (-369,+183)
Lateral Femoral 4280 4229 +51 (-400,+502) -69 (-325,+188)
Total Femoral 8163 8086 +77 (-778,+932) -162 (-645,+322)
Total Knee 12663 12910 -246(-1493,+1000) -429(-1100,+241)

DM-Difference of means
CI-Confidence Interval
*Adjusted for age, sex, height and weight
Page 52
Table 4.3 describes the association between knee injury and cartilage defects. Overall, those
in the injured group had a higher percentage of cartilage defects compared to the non-injured
group. These differences were significant for the TASOAC participants at the medial tibial,
lateral tibial, lateral femoral and total cartilage sites after adjustment. Prevalence ratios for
Offspring study were all weaker and non-significant.
Table 4.3. Association between injury and cartilage defects in the knee
Cartilage Defects Injury No Injury Unadjusted Adjusted*

Absent/Present(site) % (n/N) % (n/N) PR (95%CI) PR (95%CI)
Offspring
Medial Tibial 19%(14/71) 12% (35/299) 1.7 (0.9,2.9) 1.6 (0.9,2.8)
Lateral Tibial 15% (11/72) 14% (43/299) 1.1 (0.6,2.0) 0.9 (0.5,1.7)
Medial Femoral 8% (6/72) 6% (18/299) 1.4 (0.6,3.4) 1.4 (0.5,3.5)
Lateral Femoral 6% (4/72) 6% (19/299) 0.9 (0.3,2.5) 0.8 (0.3,2.5)
Total Knee 31% (22/72) 27% (82/299) 1.1 (0.8,1.7) 1.0 (0.7,1.5)
TASOAC
Medial Tibial 41% (16/39) 19% (69/373) 2.2 (1.4,3.4) 2.3 (1.5,3.4)
Lateral Tibial 44% (17/39) 28% (104/373) 1.6 (1.1,2.3) 1.6 (1.1,2.4)
Medial Femoral 44% (17/39) 29% (109/373) 1.5 (1.0,2.2) 1.4 (0.9,2.1)
Lateral Femoral 31% (12/39) 14% (53/373) 2.2 (1.3,3.7) 2.1 (1.3,3.5)
Total Knee 64% (25/39) 50% (186/373) 1.3 (0.9,1.6) 1.3 (1.0,1.7)
PR-Prevalence Ratio
Page 53
Table 4.4 describes the association between knee injury and BMLs. In both cohorts those in
the injured group had a higher percentage of BMLs present compared to the non-injured
group and significant associations were seen at the medial tibial, medial femoral and total
sites in both unadjusted and adjusted analysis.
Table 4.4. Association between injury and bone marrow lesions in the knee
BMLs Injury No Injury Unadjusted Adjusted*

Absent/Present(site) % (n/N) % (n/N) PR (95%CI) PR (95%CI)
Offspring
Medial Tibial 33% (13/39) 16% (25/159) 2.1 (1.2,3.8) 2.1 (1.2,3.7)
Lateral Tibial 31% (12/39) 25% (40/161) 1.2 (0.7,2.1) 1.3 (0.8,2.3)
Medial Femoral 31% (12/39) 11% (17/160) 2.9 (1.5,5.6) 2.6 (1.3,5.4)
Lateral Femoral 21% (8/39) 14% (23/159) 1.4 (0.7,2.9) 1.5 (0.7,3.2)
Total Knee 72% (28/39) 49% (82/158) 1.5 (1.1,1.9) 1.6 (1.2,2.1)
TASOAC
Medial Tibial 33% (15/46) 18% (63/346) 1.9 (1.1,3.1) 1.8 (1.1,3.0)
Lateral Tibial 17% (8/46) 15% (51/346) 1.5 (0.8,3.0) 1.6 (0.8,3.1)
Medial Femoral 17% (8/46) 12% (40/346) 2.0 (1.0,3.9) 1.9 (1.0,3.5)
Lateral Femoral 30% (14/46) 17% (58/346) 1.7 (0.9,3.0) 1.7 (0.9,3.0)
Total Knee 52% (24/46) 41% (143/346) 1.4 (1.1,1.9) 1.4 (1.0,1.9)
PR-Prevalence Ratio
Page 54
Table 4.5 describes the association between knee injury and tibial bone area. In both cohorts,
those in the injured group had a higher tibial bone area after adjustment for age, sex, height
and weight. In the Offspring study, there was a significant association between the history of
knee injury and the tibial bone area at all sites in adjusted analysis. In TASOAC, there was a
significant association at medial and total tibial bone area sites in adjusted analysis.
Table 4.5. Association between injury and tibial bone area in the knee
Bone Area Injury No Injury Unadjusted Adjusted*

(site) Mean area(mm2) Mean area(mm2) DM(95%CI) DM(95%CI)
Offspring
Medial 1874 1711 +162(+91,+233) +49(+7,+91)
Lateral 1293 1178 +116(+63,+168) +37(+4,+71)
Total 3167 2889 +278(+261,+395) +86(+23,+149)
TASOAC
Medial 2137 2133 +3(-156,+163) +91(+4,+178)
Lateral 1226 1238 -12(-125,+101) +49(-11,+109)
Total 3362 3370 -8(-269,+251) +140(+19,+260)

DM-Difference of means
Page 55
Table 4.6 describes the association between meniscal pathology and knee injury. Meniscal
tears were very common in both cohorts and there was no difference in the prevalence of
tears in injured compared to non-injured groups. Table 4.6 presents results for medial
meniscal extrusions only as lateral meniscal extrusions were extremely rare in both cohorts.
There was a higher percentage of participants with medial meniscal extrusions in the injured
group in both cohorts; however, the difference was statistically significant for the Offspring
study only.
Table 4.6. Association between injury and meniscal pathology in the knee
Meniscal Injury No Injury Unadjusted Adjusted*

Pathology % (n/N) % (n/N) PR (95%CI) PR (95%CI)
Offspring
Medial Tear 68%(41/60) 58%(138/238) 1.2 (0.9, 1.4) 1.2 (0.9, 1.5)
Lateral Tear 48%(29/60) 52%(124/238) 0.9 (0.7, 1.2) 0.9 (0.7, 1.3)
Medial Extrusion 13%(8/60) 5%(13/238) 2.4 (1.1, 5.6) 2.7 (1.1, 6.8)
TASOAC
Medial Tear 100%(29/29) 99%(289/290) 1.0 (0.9, 1.1) 1.0 (0.9, 1.0)
Lateral Tear 100%(29/29) 99%(289/290) 1.0 (0.9, 1.1) 1.0 (0.9, 1.0)
Medial Extrusion 28%(8/29) 17%(50/290) 1.6 (0.8, 3.0) 1.6 (0.8, 3.0)
PR-Prevalence Ratio
Page 56
Further adjustment for history of knee surgery did not change the effect size for any of the
associations mentioned above.
There was no offspring-control or sex interaction for any of the above mentioned associations
in the Offspring cohort. There was no sex-interaction in the TASOAC cohort for all the above
mentioned associations except for the lateral tibial cartilage volume and lateral tibial cartilage
defects sites, where male participants with a history of knee injury were losing significantly
more cartilage volume and had more cartilage defects compared to female participants.
Additional analysis was performed to examine the association between history of knee
surgery and each knee structure. There were no significant associations between the history
of knee surgery and any knee structures in the Offspring study. In TASOAC, history of knee
surgery was independently associated with a higher prevalence of cartilage defects at all sites
(highest prevalence ratio at the lateral femoral site: PR=2.9(1.9-4.5)), higher lateral tibial
bone area (DM= +29 (+3, +56)) and a higher prevalence of medial (PR=1.1 (1.0, 1.1) and
lateral (PR=1.1 (1.0, 1.1) meniscal tears (after taking a history of knee injury, age, sex,
weight and height into account).
Page 57
4.4 Discussion
To the best of our knowledge, this is the first paper to describe the association between
history of knee injury and global knee structural damage using MRI. The prevalence of
BML’s and tibial bone area was higher in those with a history of knee injury and this was
seen in both cohorts. Medial meniscal extrusion presence was higher in those with a history
of knee injury in the middle-aged cohort only, whereas cartilage defects and cartilage volume
(lateral and total tibial only) were significantly associated with knee injury in the older-aged
cohort. Meniscal tears showed no significant associations in either cohort.
Bone marrow lesions are known to be a consequence of acute injury and trauma, especially
after fractures [133] and anterior cruciate ligament (ACL) rupture [134]. Very few studies
have looked at the long term risk of developing BMLs after a knee injury. In a recent study,
Frobell et al. [135] followed 61 patients for 2 years after acute ACL disruption who were
either treated with early/late ACL reconstruction or with conservative rehabilitation. They
found that lateral compartment BMLs sustained after ACL rupture, completely resolved after
2 years in almost all patients but 34 % patients developed new BMLs over the 2 years follow-
up period. In our study there were no significant differences in the lateral compartment
between the injured and the non-injured groups in both cohorts but we did see an
approximate doubling of prevalence in the medial compartment, a site which is of greater
relevance to OA in older adults [136]. Data from this study suggests that age may not play a
major role in the association between knee injury and BMLs as we saw very similar trends in
both cohorts at all sites.
Tibial plateau bone area is associated with knee OA with increases in bone area predicting
increased JSN, osteophyte development and cartilage loss on MRI [137]. To the best of our
knowledge, no study has examined the association between history of knee injury and tibial
bone area. Data from this study shows that history of knee injury is significantly associated
with increased tibial bone area and this may reflect an attempt by subchondral bone to repair.
Age may not play a crucial role in mediating this association as total tibial bone area was
significantly associated with injury in both cohorts, although the effect size was consistently
larger in the older cohort possibly reflecting the increased time since the injury.
Page 58
Several studies have examined the association between cartilage defects/loss and history of
knee surgery [138]; however, there is limited data showing the same association with knee
injury. Elsaid et al. have shown in a rabbit injury model that knee injury can result in loss of
boundary-lubricating ability of synovial fluid which can cause damage to the articular
cartilage matrix [139]. In a 7-10 year follow-up study, Crema et al. found that knee trauma,
either surgical or non-surgical, was associated with cartilage degenerative changes in only a
minority of patients and very few patients with a complete ACL tear at baseline showed
cartilage loss at follow-up [140]. We found significant associations between cartilage defects
and history of knee injury at almost all sites in the older cohort but this was not the case for
the middle-aged cohort. This appears consistent with the Crema et al. findings as their study
population had an average age of approximately 34 years. Very few patients in their study
showed cartilage loss after a complete ACL tear. Similarly, we did not find consistent
associations between history of knee injury and cartilage volume with the lateral and total
tibial cartilage volume sites associated with knee injury in the older cohort only. This
possibly reflects changes in the ability of cartilage to repair given that cartilage defects in the
middle-aged adults can improve with time [89] but such improvement in cartilage defect
grade is very rare in older adults [141]. Cartilage defects are associated with cartilage loss
[113], so it is plausible that the associations we saw for reduced cartilage volume were also
mediated by cartilage defects or that other structural changes result in cartilage loss in later
life.
There was a significant association between the history of knee injury and meniscal
extrusions in the middle-aged cohort only but no significant associations for meniscal tears in
either cohort. In a 30 month prospective study, Englund et al. found a strong association
between meniscal pathology (tears and extrusions combined) and knee injury OR=4.14 (95%
CI 2.06 to 8.31) [142].We analysed tears and extrusions separately as both cohorts had a very
high prevalence of meniscal tears which would have then affected the meniscal extrusion
findings. Age may not affect the associations for meniscal tears/extrusions as there was no
difference in the prevalence of tears between injured and non-injured groups in either cohort
and our data suggests that tears become more prevalent with age regardless of knee injury.
History of knee surgery seems to be more important in the older adults for knee structural
pathologies. We did not observe any significant associations in the middle-aged cohort but
Page 59
this could possibly be explained by the fact that only 6% of the middle-aged participants
reported undergoing any knee surgical procedure compared to 12% of the older participants.
Interestingly meniscal tears, despite having a very high prevalence, showed a modest but a
significant association with the history of knee surgery. Our data from the two cohorts
suggests that the prevalence of meniscal tears increases with age but knee surgery probably
does play a role as well. Cartilage defects were the only structural pathology which was
consistently associated with both knee injury and surgery in the older adults, again
highlighting the possibility of reparative potential in the younger age groups.
This study has strengths and limitations. A strength includes the use of MRI to assess knee
structure and the consistent method of defining knee injury across both cohorts. The MRI
readers who performed the scoring in this study have all undergone extensive training and
demonstrate significant expertise and experience in scoring MRI features. They have
consulted and been advised by radiologists specialized in musculoskeletal imaging.
Moreover, and as noted in the manuscript, the reproducibility is high for all the MRI
measures. Limitations include, firstly the cross-sectional design which does not give
information about causality although injury preceded the MRI in all subjects. Secondly, the
use of a questionnaire to assess injury can potentially lead to errors in recall especially in
older patients who might have had a knee injury several decades ago. The older cohort in this
study had a lower prevalence of history of knee injury, which could be due to recall bias but
could also be a result of random variation between the cohorts. Other studies investigating the
role of history of knee injury in the development of OA have used structural changes like
ACL damage, meniscal tears and bone marrow lesions as markers of knee injury[143, 144].
These structural changes, although objective, are changes because of injury and not the
primary injury itself. Furthermore these structural changes can be degenerative in nature and
can occur without knee injury in association with other structural changes as part of an active
OA process [93, 94]. Thirdly, the definition of knee injury varies considerably amongst
studies as there is no standard definition of knee injury. Some studies have used similar
definitions to ours including non-weight bearing as a criterion for a significant knee injury
however the duration of non-weight bearing seems variable across studies. Other studies have
used outcomes such as duration of pain after the injury as a criteria for a significant injury
[128]. Therefore, our definition may have affected our findings; nonetheless we have
Page 60
demonstrated biologically plausible associations in both cohorts. Fourthly, we did not have
any information about the time from injury which can potentially be an important factor when
assessing the severity of structural damage [145].
Conclusion
The association between knee injury and MRI-assessed structural pathology in the knee joint
is moderate and appears to be stronger in older adults compared to middle aged adults.
Page 61
Chapter 5 Role of Family History in Osteoarthritis Progression
Chapter Five
A Family History of Knee Joint Replacement

Increases the Progression of Knee Radiographic
Osteoarthritis and Medial Tibial Cartilage Volume
Loss Over 10 Years
Page 62
5.1 Introduction
Osteoarthritis (OA) is a slowly developing chronic disease that has a multifactorial origin
with the knee being the most commonly affected joint [115]. The pathogenesis of OA is not
fully understood but some of the factors which contribute towards the development of OA
include genetics, obesity, joint injury and occupational factors [68]. There is strong evidence
that genetic factors play an important role in radiographic OA (ROA) of the hands and the
spine [68, 69]. A cross-sectional study [70] using the present cohort showed a significant
genetic contribution to the severity but not prevalence of knee ROA but the evidence is
inconsistent for knee ROA [68-72]. This may reflect the difficulty to target specific genes. A
recent meta-analysis of 9 genome-wide association studies including 5636 knee OA patients
and 16972 controls, found that only 2 out of 199 published candidate OA genes had any
significant association with OA [146]. The inconsistency may be due to different study
designs [146], inherent measurement error associated with diagnosis of ROA, short follow-up
periods and varying levels of genetic susceptibility of different phenotypic components of
knee OA [147, 148].
Magnetic resonance imaging (MRI) is being increasingly used to study OA as it allows

visualisation of the whole joint [124]. It is possible that different structures comprising the
knee joint are under separate genetic influences. Twin studies have already shown high
heritability estimates for cartilage volume in all compartments of the knee joint [149].
Previous work using the present cohort has also shown high heritability estimates for tibial
and patellar cartilage volume [70] and a significant genetic contribution to medial tibial
cartilage loss over 2 years [150]. Along with cartilage volume loss, change in cartilage
defects, tibial bone area and quadriceps muscle strength were all shown to be under genetic
influence [150]. All these structural changes are thought to contribute towards the
progression of the disease, but a limitation in the design of these studies [70, 150] was the
lack of radiographs at two years as it was not expected to see any major changes on
radiographs in this time frame in a middle-aged population.
The aim of this population-based longitudinal study was therefore to describe the 10 year
change in knee ROA and cartilage volume loss between offspring having at least one parent
Page 63
with a total knee replacement for severe primary knee OA, and age- and sex-matched
controls with no family history of knee OA.
Page 64
5.2 Methods
5.2.1 Study subjects
This study was conducted as part of the Offspring study, which is an ongoing population-
based study. The Offspring study began in southern Tasmania (primarily in the city of
Hobart) in June 2000. Half of the participants were the adult offspring of patients who had a
knee replacement performed for idiopathic knee OA at any Hobart hospital from 1996 to
2000 [102]. The diagnosis was confirmed by reference to the medical records of the
orthopaedic surgeon and the original radiographs when possible. The other half were age and
sex matched controls, randomly selected from the population with no history of knee OA in
either parent. This study includes data from the baseline visit, 2 year and 10 year follow up.
The Southern Tasmanian Health and Medical Human Research Ethics Committee approved
the protocol, and written informed consent was obtained from all participants. Participants
were excluded if they had a contraindication to MRI (including metal sutures, presence of
shrapnel, iron filing in eye, or claustrophobia). Participants were also excluded if they had
undergone a knee replacement surgery or did so after the commencement of the study. Knee
pain and knee injury were not a basis for exclusion.
5.2.3 Knee pain
Knee pain was assessed using an interviewer administered questionnaire as described

previously [102]. All the participants were asked the following question:
Page 65
Have you had knee pain for more than 24 hours in the last 12 months or daily pain on greater
than 30 days in the last year?
5.2.4 Leg strength
Muscle strength was measured by dynamometry at the lower limb (involving both legs
simultaneously). This primarily involves the hip flexors and knee extensors. The participants
were instructed in each technique prior to testing, and each measure was performed twice.
The repeatability estimate (Cronbach’s alpha) was 0.91 [70]. The device was calibrated by
suspending known weights at regular intervals.
5.2.5 Magnetic resonance imaging
MRI of the right knee was performed as described previously [89, 104, 108]. Knees were
imaged in the sagittal plane on a 1.5-T whole-body magnetic resonance unit (Picker
International, USA) using a commercial transmit-receive extremity coil at the baseline visit, 2
year and 10 year follow up. The following image sequence was used: (1) a T1-weighted fat-
thickness of 1.5 mm without an interslice-gap (at all three visits); and (2) a T2-weighted fat
saturation 2D fast spin echo, flip angle 90°, repetition time 3067 ms, echo time 112 ms, field
of view 16 cm, 15 partitions, 256×256 matrix, slice thickness of 4 mm with an interslice gap
of 0.5–1.0 mm (at visit 2 and 3).
Cartilage volume:
Knee cartilage volume was evaluated at baseline and 10 years by a trained observer on T1-
weighted gradient echo MR images. Knee cartilage volume was determined by means of
image processing on an independent workstation at baseline and follow up. The volumes of
Page 66
individual cartilage plates (medial tibia and femora, and lateral tibia and femora) were
isolated from the total volume by manually drawing dis-articulation contours around the
cartilage boundaries on a section by section basis. These data were then resampled by means
of bilinear and cubic interpolation (area of 312 × 312 µm by 1.5 mm thickness, continuous
sections) for the final three-dimensional rendering to calculate the cartilage volume.
Tibial cartilage volume was assessed using Osiris (University of Geneva, Switzerland)
software as previously described [104, 109]. The coefficient of variation(CV) ranged from
2.1–2.2% for intra-observer repeatability [84]. Femoral cartilage volume was determined
using Cartiscope (ArthroLab, Montreal, Canada), as previously described [110-112]. The CV
was approximately 2% for intra-observer and inter-scan repeatability [111]. Total cartilage
volume was calculated as: tibial + femoral cartilage volume.
Change in cartilage volume was calculated as: follow-up total cartilage volume - baseline
total cartilage volume.
Readers were not blinded to the chronological sequence of the scans to reduce measurement
error.
Cartilage defects were assessed on T1-weighted gradient echo MR images at the medial
tibial, medial femoral, lateral tibial, and lateral femoral sites on a 0-4 scale, as previously
described [113]: grade 0=normal cartilage; grade 1=focal blistering and intra-cartilaginous
low-signal intensity area with an intact surface and base; grade 2=irregularities on the surface
or base and loss of thickness <50%; grade 3=deep ulceration with loss of thickness >50%;
and grade 4=full-thickness chondral wear with exposure of subchondral bone. Intraobserver
reliability (expressed as intraclass correlation coefficient(ICC)) ranged from 0.89-0.90.
Interobserver reliability was assessed in 50 MR images and yielded an ICC of 0.85-0.90
[113]
Page 67
5.2.7 Bone area
The following measures of bone size were determined: total patella bone volume, and medial
and lateral tibial plateau areas as described previously [104]. Contours were drawn around the
patella in images 1.5 mm apart on sagittal views. Total volume was calculated for the patella
due to its irregular shape, which made it difficult to identify a simpler, representative measure
of patella size. Medial and lateral tibial plateau area was determined by creating an isotropic
volume from the 3 input images closest to the joint after reformatting in the axial plane. The
areas of the medial and lateral tibial plateaus were then directly measured from these images.
The CV was 2.2% for the patella, 2.3% for the medial tibial plateau, and 2.4% for the lateral
tibial plateau [104].
Meniscal tears were assessed by a trained observer on T1-weighted gradient echo and T2-
weighted (side by side) MR images at visit-2 and 3 of the study as previously described
[151]. The proportion of the menisci affected by a tear was scored separately (0-2 scale;
0=absence of a tear, 1=simple tear of different types: longitudinal, oblique, radial or
horizontal, 2=complex tear signifying loss>50% area of meniscal tissue) at the anterior,
middle, and posterior horns. Anterior, middle and posterior scores were summed to create
medial and lateral meniscal tear scores. The intra- and inter-observer correlation coefficient
ranged from 0.86 to 0.96 [111]. Meniscal tears were measured at visits 2 and 3 of the
Offspring study, 2 and 10 years after the baseline visit.
Bone marrow lesions (BMLs) were assessed on fat suppressed T2-weighted MR images as
described previously [93]. BMLs were defined as areas of increased signal intensity in the
sub-chondral bone at the medial tibial, medial femoral, lateral tibial, lateral femoral, superior
patellar and inferior patellar sites. One trained observer scored the BMLs by measuring the
maximum area of the lesion in a specific compartment. The observer manually selected the
Page 68
MRI slice with the greatest BML size. The BML with the highest score was used if more than
one lesion was present at the same site. The ICC was 0.97. BMLs were measured at phase 2
of the Offspring study, 2 years after the baseline visit.
5.2.10 Radiology
baseline and 10 years. The angle was kept to 10–15˚ by a purpose built goniometer. The tube
to film and tube to tibial plateau angle was 90˚. Daily quality assurance was performed on the
equipment. Radiographs were scored individually for osteophytes and joint space narrowing
(JSN), as described previously [84]. Each of the following four features was scored on a scale
from 0 to 3 (0 = normal and 3 = severe): medial JSN, lateral JSN, medial osteophytes
(femoral and tibial combined) and lateral osteophytes (femoral and tibial combined). Each
score was arrived at by consensus with two readers (LC, AM) simultaneously assessing the
radiograph with immediate reference to the Osteoarthritis Research Society International
(OARSI) atlas [82]. A non-zero score in either JSN or osteophytosis was regarded as
evidence of ROA. Reproducibility was assessed in 50 radiographs, two weeks apart, and
yielded an ICC of 0.99 for osteophytes and 0.98 for JSN.
Change in ROA was calculated as: follow-up ROA score - baseline ROA score.
Readers were not blinded to the chronological sequence of the scans to reduce measurement
error.
5.2.11 Statistical analysis
This study was no longer paired as matching is no longer possible due to loss to follow up.
T-tests were used to describe the differences in baseline characteristics and ROA/cartilage
volume loss over 10 years between offspring and controls. Negative binomial and linear
regression were used to describe radiographic changes (expressed as difference in ratios (dr))
and cartilage loss (expressed as difference in means (dm)) respectively. Multivariable
Page 69
analyses were first adjusted for age, sex and the corresponding baseline measures (i.e.
baseline cartilage volume for cartilage loss). We then adjusted for the five baseline measures
which were significantly different between offspring and controls in the original whole
sample using conditional logistic regression (BMI, knee pain, cartilage defects, bone size and
leg strength) [102, 152] in order to examine potential mediators. Further analysis was done to
explore any sex interaction within offspring and control groups for ROA changes and
cartilage volume loss in the multivariable models.
A p-value less than 0.05 (two-tailed) was considered statistically significant. All analyses
were performed on Intercooled Stata V.12.0 for windows (StataCorp LP).
Page 70
5.3 Results
Of the 371 participants included in the Offspring study, 220 between the ages of 26 and 61
years were followed up for 10 years. None of the participants who were lost to follow-up
underwent a knee replacement surgery. Table 5.1 describes the baseline characteristics of
participants who were followed up (220) compared to participants who were lost to follow up
(151). There were no significant differences between the two groups except for a higher
lateral tibial bone area in the participants who were followed up.
Table 5.2 describes baseline characteristics of the offspring (n=115) and controls (n=105).
The mean age of both offspring and controls at baseline was approximately 45 years and both
groups had a higher proportion of female participants. Prevalence of ROA at baseline was
low in both groups without any significant differences between the two groups. Offspring had
a slightly but significantly higher BMI, higher lateral femoral cartilage volume, knee pain
prevalence and total cartilage defects score compared to controls.
Comparison between offspring and controls (Table 5.3) for radiographic score changes
revealed that offspring had a significantly greater increase in medial JSN, total medial
osteophytes, total lateral osteophytes, total osteophytes and total ROA scores. There was no
significant difference in lateral and total JSN scores. For cartilage volume loss (Table 5.3),
offspring had a significantly greater loss at the medial tibial site only. There was no
significant difference in cartilage volume loss at lateral tibial, medial femoral, lateral femoral
and patellar sites.
Multivariable comparison (Table 5.4) between offspring and controls for radiographic score
changes revealed that after adjustment for age, sex and the corresponding baseline measures,
offspring had a greater increase in medial JSN, total medial osteophytes, total lateral
osteophytes, total osteophytes and total ROA scores. However, after further adjustment for
the baseline factors, which were significantly different between offspring and controls, the
difference in ratios remained significantly greater only for medial JSN score. Further
adjustment for medial meniscal tears (measured at 2 year) had no effect; however adjustment
for medial (tibial + femoral) BMLs (measured at 2 years) changed the effect size by more
than 10% [(dr = +1.63 (+0.84, +3.03)]. For absolute cartilage volume loss (Table 5.4),
difference in means at the medial tibial site became non-significant (p=0.054) after adjusting
Page 71
for age, sex and corresponding baseline measure and remained so after further adjustment for
differences in baseline factors (p=0.055).
Table 5.1. Baseline characteristics of the participants who were followed-up and who were
lost to follow up
Characteristic Follow-up Loss Follow-up P-value

(n = 220) (n = 151)
Age (years) 45.3 ± 6.7 45.1 ± 7.2 0.806
Female (%) 58 59 0.749*
BMI (kg/m2) 27.2 ± 4.9 26.8 ± 4.3 0.499
Offspring (%) 52 47 0.891
Radiographic OA (%) 18 15 0.486*
Knee pain present (%) 33 34 0.917*

Medial tibial cartilage volume (mm3) 2234.1 ± 547.3 2230.8 ± 585.3 0.956
Lateral tibial cartilage volume (mm3) 2620.9 ± 671.3 2579.3 ± 680.9 0.561
Medial femoral cartilage volume (mm3) 4594.8 ± 1295.2 4541.4 ± 1145.1 0.734
3
Lateral femoral cartilage volume (mm ) 4753.6 ± 1268.3 4719.6 ± 1252.0 0.836
Patellar cartilage volume (mm3) 3480.2 ± 976.3 3430.3 ± 975.1 0.629
Medial tibial cartilage defects 1.2 ± 0.4 1.2 ± 0.4 0.697
Lateral tibial cartilage defects 1.2 ± 0.4 1.2 ± 0.4 0.948

Medial femoral cartilage defects 0.9 ± 0.5 1.0 ± 0.5 0.443
Lateral femoral cartilage defects 0.9 ± 0.5 0.9 ± 0.5 0.526
Patellar cartilage defects 1.2 ± 0.9 1.2 ± 1.1 0.987
Medial tibial bone area (cm2) 17.6 ± 2.8 17.1 ± 2.6 0.092
Lateral tibial bone area (cm2) 12.2 ± 2.1 11.7 ± 1.9 0.027
Patellar bone volume (cm3) 13.9 ± 3.3 13.5 ± 3.3 0.279
Mean ± standard deviation except for percentages; *Determined by Chi square test, others by t-
test
Page 72
Table 5.2. Baseline characteristics of the study participants
Offspring (N=115) Controls (N=105) P-Value
Age (years) 44.8 ±6.8 45.8 ±6.5 0.261
Female (%) 55% 60% 0.436
BMI (kg/m2)* 27.9 ±5.3 26.3 ±4.5 0.018
Any ROA (%) 18% 17% 0.894
Any medial JSN (%) 14% 14% 0.937
Any lateral JSN (%) 3% 4% 0.907
Any tibial osteophytes (%) 15% 8% 0.199
Any femoral osteophytes (%) 14% 4% 0.052
Medial tibial cartilage volume (mm3) 2271 ±46 2194 ±59 0.295
Lateral tibial cartilage volume (mm3) 2692 ±670 2544 ±668 0.104
Medial femoral cartilage volume(mm3) 4679 ±1174 4354 ±1181 0.055
Lateral femoral cartilage volume(mm3) 4859 ±1254 4437 ±1305 0.022
Patellar cartilage volume (mm3) 3534 ±949 3421 ±1006 0.393

Knee pain prevalence(%)* 45% 20% <0.001
Total tibial bone area (mm2)* 3017 ±428 2934 ±498 0.191
Patellar bone volume (mm3) 13970 ±3196 13770 ±3440 0.651
Mean total cartilage defects score* 4.4 ±1.3 4.0 ±1.2 0.039
Mean leg strength (kg)* 128 ±4.5 126 ±4.4 0.718
Any bone marrow lesion#^ 68% 60% 0.249
Any meniscal tear@^ 20% 23% 0.367
Where errors are shown, results are means ± SD

BMI (body mass index), ROA (radiographic osteoarthritis), JSN (Joint space narrowing)
Mean total cartilage defects score (mean of sums of medial tibial, medial femoral, lateral tibial and
lateral femoral cartilage defects)
*significantly different between offspring and controls in the whole baseline study population (using
conditional logistic regression)
#
Any bone marrow lesion= tibial, femoral and/or patella
@
Any meniscal tear=medial and/or lateral
^
Measured at phase 2 (two years after the baseline visit)
Page 73
Table 5.3. Comparison of radiographic changes and cartilage loss (absolute) between
offspring and controls
Outcome factor Offspring (N=115) Controls (N=105)
Radiographic score changes Mean score Mean score P-Value

±SD ±SD
Increase in medial JSN 0.32 ±0.56 0.17 ±0.39 0.019
Increase in lateral JSN 0.07 ±0.35 0.09 ±0.32 0.774
Increase in total JSN 0.39 ±0.70 0.25 ±0.52 0.113
Increase in total medial osteophytes 0.35 ±0.78 0.15 ±0.41 0.025
Increase in total lateral osteophytes 0.42 ±0.92 0.18 ±0.46 0.018
Increase in total osteophytes 0.77 ±1.44 0.34 ±0.71 0.007
Increase in total ROA score 1.15 ±1.90 0.59 ±0.87 0.007
Cartilage loss (absolute) Mean loss (mm3) Mean loss (mm3) P-Value
±SD ±SD
Medial tibial -610 ±327 -518 ±347 0.047
Lateral tibial -300 ±370 -297 ±412 0.953
Medial femoral -698 ±331 -697 ±380 0.981
Lateral femoral -701 ±337 -689 ±391 0.821
Patellar -778 ±633 -777 ±643 0.992

Total = tibial + femoral
Page 74
Table 5.4. Multivariable analyses of differences between offspring and controls in changes in
radiographic changes and cartilage loss (absolute)
Outcome factor Unadjusted Adjusteda Adjustedb
Radiographic changes Difference in ratios and 95% confidence interval

Increase in medial JSN +2.03 +2.04 +1.93
(+1.11, +3.51) (+1.12, +3.52) (+1.04,+3.51)
Increase in lateral JSN +0.82 +0.82 +0.53
(+0.31,+2.83) (+0.31,+2.83) (+0.21,+1.80)
Increase in total JSN +1.50 +1.51 +1.44
(+0.91,+2.60) (+0.93,+2.53) (+0.82,+2.32)
Increase in total medial +2.34 +2.32 +1.84
osteophytes (+1.11,+4.53) (+1.11,+4.57) (+0.93,+3.80)
Increase in total lateral +2.32 +2.51 +1.91
osteophytes (+1.22,+4.63) (+1.27,+5.11) (+0.92,+3.93)
Increase in total osteophytes +2.30 +2.36 +1.63
(+1.30,+4.03) (+1.33,+4.24) (+0.94,+2.92)
Increase in total ROA score +1.90 +1.81 +1.52
(+1.31,+3.04) (+1.21,+2.79) (+0.93,+2.23)
Cartilage loss (absolute) Difference in means (mm3) and 95% confidence interval
Medial tibial -91.52 -78.81 -79.13
(-181.61,+1.31) (-158.91,+1.23) (-161.92,+3.71)
Lateral tibial -3.00 +10.62 +35.41
(-107.90,+101.78) (-90.59,+112.02) (-69.33,+140.12)
Medial femoral -1.23 +30.87 +18.59
(-101.39,+98.72) (-56.11,+117.91) (-72.24,+109.41)
Lateral femoral -11.80 +30.72 +45.81
(-114.42,+90.80) (-60.42,+121.72) (-45.43,+136.91)
Patellar -0.90 +15.93 +80.20
(-171.57,+169.82) (-132.63,+164.54) (-67.33,+227.69)
a
Adjusted for age, sex and corresponding baseline measure
b
Adjusted for a + baseline differences between offspring and controls (BMI, knee pain, cartilage
defects score, tibial bone area and leg strength)
Total = tibial + femoral
There were no significant differences between the two groups for percentage per annum
cartilage loss at any site. Medial tibial region showed a higher percentage per annum loss in
the offspring group without reaching statistical significance in either the unadjusted [(dm = -
0.31 (-0.72, +0.03; p = 0.078)] or the fully adjusted model [(dm = -0.30 (-0.71, +0.01; p =
0.055)].
Page 75
5.4 Discussion
This is the first study to confirm that offspring of those with a knee replacement for OA have
a higher risk of worsening knee OA over 10 years. Despite no difference in ROA (which had
a low prevalence) at baseline between the offspring and controls, offspring experienced
greater increases in medial JSN and osteophytes at all sites. Offspring also had higher
absolute cartilage volume loss. The increases in osteophytes and cartilage volume loss were
largely mediated by differences between the offspring and controls at baseline (BMI, knee
pain, cartilage defects, bone size and leg strength) as the estimates were reduced by 18-30%
for osteophytes and 14% for absolute cartilage volume loss. Increase in medial JSN was
independent of these baseline differences and accounted for only 5% reduction in estimates.
Several studies have described the role of genetics in prevalent disease using radiographs [68,
153] but very few have examined the influence of genetic factors on disease over time and
none have done so in a younger population. Results from this study not only suggest that
offspring with a family history of knee OA are at a higher risk of worsening knee OA over 10
years but also highlight the structural and non-structural factors that mediate these changes.
The data shows that OA is not very common at age 45 in those with a predisposition to OA
but becomes more prevalent over a 10-year time frame compared to a control population.
This suggests that the genes responsible may express themselves later in life, possibly
through interaction with environmental factors such as BMI and muscle strength, as pointed
out by reduction in estimates after adjustment for baseline differences. Another possibility is
that the mechanisms counteracting the expression of these genes are more effective at a
younger age.
The data from this study also suggests that progression of both JSN and osteophytes are
under genetic influence. Previously only Zhai et al. [154] have shown high heritability
estimates for disease progression in the medial compartment of the knee over 7 years using a
twin study design. Our results are consistent with Zhai et al. for the progression of JSN only,
as they did not find any significant heritability estimates for osteophytes. These results point
to some interesting aspects of the role genes play in the progression of OA. Firstly, our data
suggests that both JSN and osteophytes are under genetic influences as suggested by higher
progression of JSN in offspring in the medial compartment and osteophytes at all sites.
Page 76
Previously, Uitterlinden et al. [155], have shown that two separate genes control the
expression of JSN and osteophytosis in a population-based sample of healthy older adults.
Interestingly, progression of osteophytes was mediated by baseline differences between the
two groups, whereas progression of medial JSN was independent of these differences. This
suggests that the gene responsible for progression of osteophytes possibly interacts with
environmental factors such as BMI and muscle strength to express its effect. The twin study
design is often criticized due to the assumption of similar shared environment between
monozygotic and dizygotic twins. Unlike twins, offspring and controls do not share the same
environment, which would explain why consistently higher estimates for progression of
osteophytes at all sites were observed.
Offspring also had a significantly higher absolute cartilage volume loss at medial tibial site
compared to controls over 10 years. As mentioned previously, the gene coding for COL2A1
has been shown to be associated with JSN [155]. COL2A1 is a structural protein found in
articular cartilage, which explains the similar trend shown by medial JSN and medial tibial
cartilage loss. Also similar to JSN, we saw the association only in the medial compartment.
The fact that we did not see any differences between the two groups for medial femoral
cartilage volume loss, raises a few questions: (i) it is possible that cartilage volume loss at
medial femoral and medial tibial sites are under separate genetic, structural or environment
influences (ii) cartilage volume loss at the medial femoral site contributes less to JSN or
happens later in life (iii) cartilage at these two sites varies in composition (iv) other co
pathologies such as meniscal tears or BMLs might be associated more strongly with tibial
compared to femoral cartilage volume loss (v) we used different methodologies to measure
cartilage volume at the two sites, which might have led to measurement error.
Previous work from the offspring study has shown the role of genetics for the development of
meniscal tears and BMLs. Ding et al. [94] showed that offspring had a significantly higher
prevalence for meniscal tears, whereas Zhai et al. [106]showed high heritability estimates for
both the prevalence and severity of BMLs in offspring group sibling pairs. Interestingly
adjusting for medial meniscal tears did not alter the effect size of difference in ratio for
change at medial JSN site, but adjusting for BMLs changed the effect size by more than 10%.
Moreover, neither explained a majority of the change. It should be noted that both of these
Page 77
structures were scored at the first follow up, two years after the baseline visit, as we only had
the T1-weighted fat suppressed MRI sequences at baseline.
Baseline differences mediating the higher risk of ROA progression and cartilage volume loss
is biologically plausible. High BMI is a known risk factor for both ROA progression and
cartilage volume loss [156]. Tibial bone area, reduced leg strength and cartilage defects are
not only risk factors for ROA progression [125, 157] but also had high heritability in sib-pair
analysis from the present cohort [70]. Interpretation of higher prevalence of knee pain in the
offspring is tricky as the assessment of knee pain is subjective and can be influenced by a
variety of factors such as recall bias due to family history of OA. Nevertheless, there is
evidence pointing to genetic contribution to expression of pain in knee OA. We have
previously shown high heritability of knee pain in a sib-pair study [70]. Furthermore,
polymorphisms in COMT and TRPV1 genes have been identified which could alter the
processing of nociceptive pain associated with OA [43]. A high prevalence of knee pain in
the offspring suggests that genetic factors may also lead to knee pain. However, adjustment
for knee pain did not change the results in the present study. Different baseline characteristics
in the offspring (including higher prevalence of MRI assessed structural abnormalities) could
also mean that onset of the disease process in the offspring occurs at a younger age.
One of the major strengths of our study is the long follow-up period. This study has the
longest follow-up period for any OA study using MRI. Another strength of this study is the
exploration of the structural and non-structural factors mediating ROA changes and cartilage
volume loss. However, this study has potential limitations as well. Over the ten years there
was a loss to follow-up of around 40%. Such a high number, although not ideal, is expected
in a long follow up period. Although we did not see any major differences in the main study
variables between participants who were followed-up and who were lost to follow-up but it
can still be a potential source of bias in the results shown in this study. Loss to follow-up also
meant that the initial paired design of the study was invalidated. Loss of pairing resulted in a
slight gender and age imbalance between offspring and controls. Nonetheless, all our
analyses were adjusted for age and sex and adjusting for these had little effect on the results.
Moreover, while we could adjust for meniscal tears and BMLs scored at 2 years we did not
have them at baseline possibly leading to greater measurement error. Lastly, tibial and
femoral cartilage volume were segmented using different methodology as was outlined in the
Page 78
manuscript. Separate readers performed the measurements, which resulted in differences in

how the scans were processed. Although both methods are almost equally sensitive at picking
up any change in cartilage volume [158], this difference can still be a source of potential bias.
Conclusion:
The offspring of subjects having a total knee replacement have greater worsening of ROA
(both JSN and osteophytes) and higher medial tibial cartilage volume loss over ten years.
Most of these changes are mediated by differences in baseline characteristics of offspring and
controls except for increase in medial JSN.
Note: Letter to the editor by Kuijer et al. regarding results in this Chapter and our reply are
attached in Appendix A.
Page 79
Chapter 6 Natural History of Bone Marrow Lesions
Chapter Six
The Clinical Significance, Natural History and

Predictors of Bone Marrow Lesion Change Over
Eight Years
Page 80
6.1 Introduction
Osteoarthritis (OA) is the most common joint disorder worldwide, and the knee is the most
common joint affected [159, 160]. Bone marrow lesions (BMLs) play a key role in the
pathogenesis of knee OA – they are associated with OA symptoms such as pain and function,
and predict cartilage loss and joint replacement surgery [91, 93, 161, 162]. In a recent Delphi
exercise that aimed to establish a definition for OA on magnetic resonance imaging (MRI),
BMLs were included as a key component of the diagnostic criteria [101].
The current literature on the natural history of BMLs is conflicting, with significant variation
depending on the study population. A study in a healthy population has shown that incident
BMLs developed in 14% of individuals over 2 years, and that knee pain was more likely to
develop in these participants [163]. The same study showed that nearly one-half of the BMLs
present at baseline completely resolved, while another study of middle-aged healthy women
over 2 years found similar results [164]. Studies in symptomatic OA populations generally
show a lower percentage of BMLs resolving, with one study reporting that less than 1% of
patients showed a BML decrease over 30 months [165]. Other studies have quoted higher
figures, with 10% of BMLs resolving over 2 years in a study by Kornaat and colleagues
[166]. Our group has previously reported that rates of incident BMLs were low (7%), with
about one-quarter of BMLs showing an increase or a decrease in size over 2.7 years in a
population-based cohort of older adults with and without OA [93]. The reasons behind these
variations are unclear; however, it is worth noting that no study has looked at the natural
history of BMLs beyond 3 years.
Pain is a key criterion for a clinical diagnosis of OA. A number of studies have reported an
association between BMLs and pain across a range of demographics and activities [167-171].
Furthermore, longitudinal studies have shown that increases in BML size and incident BMLs
are both associated with increasing knee pain over 2 to 3 years [76, 172]. However, some
studies have shown no association between BMLs and pain longitudinally [166, 173] or
cross-sectionally [86].
Given the role of BMLs in OA, there has been interest in the risk factors that lead to an
increased risk of developing BMLs. There is significant overlap with the major risk factors
for OA, and age and weight have been shown to be some of the strongest risk factors for
Page 81
BMLs [174, 175]. Physical activity, particularly doing over 10,000 steps per day, may
aggravate existing BMLs [65]. Recently, vascular risk factors have also been implicated due
to their effects on blood flow in the small vessels of subchondral bone [176]. Smoking,
increased serum glucose levels, serum cholesterol and triglyceride, fatty acid intake,
carbohydrate intake and changes in retinal microvasculature have all been associated with
BMLs [177-181]. There is also a significant genetic component [106].
The conflicting data on natural history and clinical significance may be attributable to the
differing methodology in many of these studies. These include differences in imaging
protocols, sample size, age, sites measured, and severity of OA in the study sample.
Importantly, few studies have followed the progression of BMLs beyond 3 years. The aims of
this study were to describe the natural history of BMLs over 8 years, to examine the
relationship between change in BML size and change in knee pain, and to examine factors
predicting change in BML size.
Page 82
6.2 Materials and methods
based study. The Offspring study began in southern Tasmania (primarily in the capital city of
Hobart) in June 2000. One-half of the participants were the adult offspring of patients who
had a knee replacement performed for idiopathic knee OA at any Hobart hospital from 1996
to 2000 [102]. The diagnosis was confirmed by reference to the medical records of the
orthopedic surgeon and the original radiograph when possible. Controls were age and sex
matched and were randomly selected from the population. Participants were excluded if they
had a contraindication to MRI (including metal sutures, presence of shrapnel, iron filing in
eye, or claustrophobia). This study includes data from the second and third visits at
approximately 2 and 10 years respectively, because T2-weighted MRI scans were not
performed at baseline.
the protocol, and written informed consent was obtained from all participants.
stadiometer. Body mass index (BMI) was calculated as weight (kg)/height (m2). Smoking was
assessed by questionnaire and categorized as current smoker, past smoker or never smoked
[109] . Physical activity was also assessed by a self-administered questionnaire that assessed
the amount of time spent in light and strenuous physical activity on a five-point scale [182].
Participants were asked about the number of days during the last 14 days spent doing at least
20 minutes of strenuous exercise (that is, bicycling, brisk walking, jogging, aerobics, and so
forth that was enough to raise your pulse rate or cause you to breathe faster) and light
Page 83
exercise (that is, walking, light housework, slow bicycling, and so forth that was not severe
enough to cause a pulse rate rising or breathing increase). The participants then chose a score
between 1 and 5, where score 1 represents no days, score 2 represents 1 or 2 days, score 3
represents 3 or 5 days, score 4 represents 6 or 8 days, and score 5 represents 9 days or more
of exercise.
6.2.3 Leg strength
Leg strength was measured by dynamometry at the lower limb, involving both legs
simultaneously. This primarily involves the hip flexors and knee extensors. Each measure
was performed twice, with instructions given prior to testing. The repeatability estimate
(Cronbach’s alpha) was 0.91. The device was calibrated by suspending known weights at
regular intervals [150].
6.2.4 Knee pain
McMaster Universities Osteoarthritis Index at both visits [78]. Five categories of pain
(walking on flat surface, going up or down stairs, at night, sitting or lying, and standing
upright) were assessed separately with a 10-point scale from 0 (no pain) to 9 (most severe
pain). Each score was then summed to create a total pain score (range 0 to 45).
6.2.5 Radiography
A standing anteroposterior semiflexed view of the right knee (at 15° flexion) was performed
in all participants at baseline and 10 years. Radiographs were scored individually for
osteophytes and joint space narrowing. Each of the following four features was scored on a
scale from 0 to 3 (0 = normal and 3 = severe): medial joint space narrowing, lateral joint
space narrowing, medial osteophytes (femoral and tibial combined), and lateral osteophytes
(femoral and tibial combined). Each score was arrived at by consensus with two readers
Page 84
simultaneously assessing the radiograph with immediate reference to the Osteoarthritis

Research Society International atlas [83]. A nonzero score in either joint space narrowing or
osteophytosis was regarded as evidence of radiographic osteoarthritis (ROA). Reproducibility
was assessed in 50 radiographs, 2 weeks apart, and yielded an intraclass correlation
coefficient of 0.99 for osteophytes and 0.98 for joint space narrowing.
An MRI scan of the right knee was performed on a 1.5 T whole-body magnetic resonance
unit (Picker, Cleveland, OH, USA) with the use of a commercial transmit–receive extremity
coil. Knees were imaged in the sagittal plane and the following image sequences were used:
visit two, a T2-weighted fat saturation two-dimensional fast spin echo (flip angle 90°;
repetition time 3,067 ms; echo time 112 ms; field of view 16 cm; 256 × 256 matrix; slice
thickness of 4 mm with a between-slices gap of 0.5 to 1.0 mm); and visit three, a T2-
weighted fat saturation two-dimensional fast spin echo (flip angle 90°; repetition time
3,067 ms; echo time 112 ms; field of view 16 cm; 256 × 256 matrix; slice thickness of 2 mm
with a between-slices gap of 0.5 mm).
Visit one only involved T1 MRI scans, which were not suitable for comparison of BMLs over
time. Subchondral BMLs were assessed using Osirix software (University of Geneva,
Geneva, Switzerland) and were defined as areas of increased signal adjacent to the
subcortical bone at the medial tibial, medial femoral, lateral tibial, lateral femoral, superior
patella, and inferior patella sites as described previously [93]. One trained observer scored the
BMLs by measuring the maximum area (cm2) of the lesion at both time points. The observer
manually selected the MRI slice with the greatest BML size. The BML with the largest size
was recorded if more than one lesion was present at the same site. MRIs at both time points
were read paired with the chronological order known to the observer but blinded to clinical
status. Participants were given a BML score (cm2) for each of the six sites (medial tibial,
medial femoral, lateral tibial, lateral femoral, superior patella, and inferior patella sites) as
well as a total BML score, which was the sum of the scores at each site. Change in BML size
was then calculated by subtracting the visit two BML size from the visit three BML size.
Page 85
Intraobserver repeatability was assessed in 40 subjects with at least a 2-week interval

between the readings. The intra-class correlation coefficient was 0.97.
To examine the natural history of BMLs, a significant change in BML size was defined as
any change above or below the least significant criterion (LSC). The LSC takes into account
measurement error and the correlation between BML measurements at baseline and follow-
up. The formula is as follows, where σ is the standard error of the mean and ρ is the serial
correlation:
The LSC was calculated for each of the six sites in the knee (11 mm2 for medial femoral,
17 mm2 for lateral femoral, 16 mm2 for medial tibial, 14 mm2 for lateral tibial, 15 mm2 for
superior patellar, and 13 mm2 for inferior patellar BMLs). This was then used to calculate the
number of BMLs increasing and decreasing in size, where an increase in BML size was
defined as any change greater than the LSC, and vice versa for a decrease in BML.
Meniscal damage was assessed by a trained observer on T1-weighted MRI scans as described
previously [110]. Each meniscus is divided into three segments (anterior horn, body and
posterior horn) for the assessment of both meniscal extrusions and tears.
Extrusion is defined as when meniscal tissue extends beyond the tibial margin, and complete
extrusion is defined as when the meniscus has no contact with the joint space. For extrusions,
each segment (anterior horn, body, and posterior horn) of both medial and lateral menisci
were scored on a scale from 0 to 2 (0 = no extrusion, 1 = partial meniscal extrusion,
2 = complete meniscal extrusion with no contact with the joint space). Each meniscus can
have a maximum score of 6 and a total knee score of 12 for extrusions.
A maximum score of 6 can be given for tears (0 = no damage, 1 = one of three meniscal
areas involved (anterior, middle, and posterior horns), 2 = two of three areas involved, 3 = all
three areas involved). This value was then scored for both medial and lateral menisci, giving
a total score of 6.
Page 86
These scores were summed to create a total meniscal pathology score, which had a possible
range from 0 to 18.
The characteristics of study participants were compared using an independent-samples t test

for continuous variables and a chi-squared test for categorical variables. Linear regression
was used to estimate the relationship between change in pain and change in BML size. This
was performed using total BML area (summed across all six sites) as well BML area at each
site specifically. Multivariable analyses were adjusted for age, sex, BMI, leg strength, and the
presence of ROA. Interactions between BML change and sex, and between BML change and
offspring–control status, were assessed from the coefficient and its standard error of product
terms was formed from the covariates for the study factors involved. Linear regression was
also used to examine potential factors predicting a change in BML size. Univariable analysis
was performed with a range of lifestyle and demographic factors (BMI, physical activity,
smoking status, ROA, and offspring–control status) and those that were significantly
associated were included in the multivariable model together with covariates for age and sex
to adjust for these factors.
Standard diagnostic checks of model adequacy were performed on all final models. Residuals
from all models were normally distributed, or approximately so, without evidence of
heteroskedasticity.
Values of P <0.05 were considered statistically significant. All statistical analysis was
performed on Intercooled Stata 12.0 for Windows (StataCorp LP).
Page 87
6.3 Results
6.3.1 Participant characteristics
The participants in this study were 198 subjects with complete MRI measures at the 2-year
and 10-year visits (52.7% of those studied at baseline). There were no significant differences
in sex, BMI, age, height, weight, frequency of ROA, and pain at baseline between those lost
to follow up (n = 178) and the participants in our study (n = 198) (data not shown).
Table 6.1 presents the characteristics of the study sample. BMLs were present in 64%
(127/198) of the sample at visit two, with an average BML size of 0.63 cm2. Of note, those
with BMLs reported higher levels of pain at visit two and had a greater change in BML size.
There were no significant differences in sex, age, BMI, offspring status, physical activity
levels, height, weight, or change in pain, although those with BMLs tended to have a higher
proportion with ROA compared with those with no BMLs. The mean (standard deviation) of
physical activity in our cohort was 2.61 (1.33) for strenuous and 4.09 (1.11) for light activity
respectively (data not shown).
6.3.2 Natural history
The 127 participants with a BML had a total of 229 BMLs present at visit two (58 had a
BML at one site, 45 had a BML at two sites, 17 had a BML at three sites, five had a BML at
four sites and two had a BML at five sites). Figure 6.1 describes the natural history of these
BMLs. Roughly one-quarter of BMLs increased (n = 55) or decreased (n = 49) in size, whilst
the remainder remained unchanged in size (n = 125) based on a change less than the LSC. Of
those without BMLs at baseline (n = 71), slightly over one-half developed one or more
incident BMLs (n = 37) over the 8 years. There was no significant difference in natural
history between offspring and controls (data not shown).
Page 88
Table 6.1. Participant characteristics
BML present BML not

P value
(n = 127) present (n = 71)
Females (%) 41% (51/127) 47% (33/71) 0.4
Age (years) 47.6 (6.5) 47.2 (6.0) 0.65
Offspring (%) 58% (74/127) 49% (35/71) 0.22
BMI (kg/m2) 28.0 (5.8) 27.0 (4.2) 0.24
Height (cm) 168.5 (9.2) 169.1 (8.9) 0.7
Weight (kg) 79.6 (18.1) 77.6 (15.1) 0.43
Leg strength (kg) 114.9 (47.4) 118.3 (44.3) 0.64
Active smokers (%) 17.3 15.5 0.77
Light activity (per unit change) 4.1 (0.09) 4.00 (0.15) 0.43
Strenuous activity (per unit change) 2.65 (0.12) 2.52 (0.15) 0.51
Radiographic osteoarthritis (%) 20% (26/127) 10 (7/71) 0.06
Pain score (0 to 29) 3.7 (5.9) 1.3 (3.1) <0.01
Change in pain score (-25 to 44) 2.5 (8.1) 2.2 (4.2) 0.84
Total BML area (0 to 4.10 cm2) 0.63 (0.80) – –
MF BML area (0 to 1.98 cm2) 0.37 (0.50) – –
LF BML area (0 to 2.54 cm2) 0.64 (0.75) – –
MT BML area (0 to 1.83 cm2) 0.33 (0.36) – –
LT BML area (0 to 1.99 cm2) 0.26 (0.31) – –
SP BML area (0 to 1.70 cm2) 0.38 (0.34) – –
IP BML area (0 to 1.06 cm2) 0.23 (0.27) – –
Change in BML area (-2.09 to 4.46 cm2) 0.61 (1.03) 0.28 (0.46) 0.01
Values were taken from visit two of the study, except for radiographic osteoarthritis that was taken from visit
one. Change refers to the difference in values between visit two and visit three. Values represent mean (standard
deviation) unless percentages. Light and strenuous activity were rated on a five-point scale: 1 = no days, 2 = 1 or
2 days, 3 = 3 to 5 days, 4 = 6 to 8 days, and 5 = 9 days or more spent doing at least 20 minutes of the respective
level of activity in the past 14 days. BMI, body mass index; BML, bone marrow lesion; IP, inferior patella; LF,
lateral femoral; LT, lateral tibial; MF, medial femoral; MT, medial tibial; SP, superior patella. Bold data
indicate P <0.05.
Page 89
Figure 6.1. Natural history of bone marrow lesions
Page 90
6.3.3 Pain
Table 6.2 presents associations between change in total BML size and change in pain. Every
1 cm2 increase in BML size resulted in a 1.53 (95% confidence interval = 0.37, 2.70) unit
increase in pain score after adjustment for age, sex, BMI, leg strength, and the presence of
ROA. Adjusting for baseline joint space narrowing, osteophytes, meniscal extrusion, and
meniscal tears did not significantly affect our findings (<10% change in the beta coefficient
following further adjustment). A significant offspring–control interaction was present
(P = 0.08), with change in BML size more strongly associated with change in pain among
offspring than controls. Furthermore, this association was stronger in males compared with
females among both the offspring group and the whole sample.
Table 6.3 presents associations between site-specific change in BML size and change in pain.
Changes in medial and lateral tibial BMLs were significantly associated with change in pain,
with the association stronger among offspring at the medial tibial site (P value for interaction
was <0.01). After adjustment for age, sex, BMI, leg strength, and ROA, change in lateral
tibial BMLs and change in pain was no longer significantly associated. No significant
association was found between change in BML size and change in pain at other sites.
For those with no pain or BMLs present at baseline (n = 42), the development of a BML was
significantly associated with an increase in pain after adjustment for age, sex, BMI, leg
strength, and ROA in offspring and controls combined with a change in pain score of 3.60
(95% confidence interval = 1.14 to 6.05) points per 1 cm2 change in BML size (not shown in
Table 6.3).
6.3.4 Factors affecting bone marrow lesion change
Table 6.4 presents predictors of BML change. BMI and strenuous activity were deleteriously
associated with change in BML size. These associations remained statistically significant
after adjusting for age and sex and each other. Smoking status, the presence of ROA, light
activity, offspring–control status, and leg strength was not associated with change in BML
size.
Page 91
Table 6.2. Relationship between change in WOMAC and change in BML size
Univariable Multivariablea Femalesa Malesa
Total 1.74 (0.65, 2.84) 1.53 (0.37, 2.70) 0.63 (-0.93, 2.20) 2.53 (0.76, 4.30)
Offspring 2.68 (1.22, 4.13) 2.50 (0.96, 4.05) 2.25 (0.10, 4.41) 3.06 (0.88, 5.24)
Controls -0.05 (-1.72, 1.62) -0.39 (-2.23, 1.44) -0.77 (-3.51, 1.98) 1.44 (-1.25, 4.14)
Data presented as beta coefficient (95% confidence interval). Values are the change in WOMAC pain score per
cm2 change in BML size. An interaction between offspring–control status and change in BML size (P = 0.01) as
well as sex and change in BML size (P = 0.08) on change in pain was present. BML, bone marrow lesion;
WOMAC, Western Ontario and McMasters Universities Arthritis Index. Bold data indicate P <0.05. aAdjusted
for age, sex, body mass index, leg strength, and radiographic osteoarthritis.
Table 6.3. Relationship between change in WOMAC and site-specific change in BML size
Univariable Multivariablea
Medial tibial 2.96 (0.59, 5.34) 3.67 (0.89, 6.45)

Controls -1.39 (-4.18, 1.40) -3.38 (-7.15, 0.39)
Offspring 8.99 (5.30, 12.68) 8.98 (5.22, 12.73)

Lateral tibial 2.37 (0.08, 4.66) 1.98 (-0.36, 4.32)
Medial femoral 2.11 (-1.14,5.36) 1.63 (-1.64, 4.90)
Lateral femoral -0.09 (-2.41, 2.24) -0.69 (-3.11, 1.73)
Superior patella 0.74 (-2.73, 4.21) 0.61 (-2.85, 4.08)

Inferior patella 3.95 (-0.59, 8.50) 4.30 (-0.26, 8.86)
Data presented as beta coefficient (95% confidence interval). Values are the change in WOMAC pain score per
cm2 change in BML size. An interaction between offspring and control status was present at the medial tibial
site (P <0.01). BML, bone marrow lesion; WOMAC, Western Ontario and McMasters Universities Arthritis
Index. Bold data indicate P <0.05. aAdjusted for age, sex, body mass index, leg strength, and radiographic
osteoarthritis.
Page 92
Table 6.4. Predictors of change in bone marrow lesion size
Univariable Multivariablea
mass index (kg/m2) 0.03 (0.00, 0.05) 0.03 (0.01, 0.05)

Strenuous activity (per unit change) 0.13 (0.03, 0.22) 0.14 (0.04, 0.23)
Current smoker (yes/no) 0.04 (-0.13, 0.20) –
Ever smoker (yes/no) 0.06 (-0.20, 0.31) –

Radiographic osteoarthritis (yes/no) 0.28 (-0.06, 0.61) –
Light activity (per unit change) 0.02 (-0.10, 0.13) –
Offspring status (yes/no) 0.13 (-0.11, 0.38) –
Leg strength (kg) 0.00 (0.00, 0.00)
Data presented as beta coefficient (95% confidence interval). Values are the change in bone marrow lesion size
(cm2) per unit change in covariates. All factors are from visit two, except for radiographic osteoarthritis that was
collected at visit one. Strenuous and light activity were assessed on a five-point scale: 1 = no days, 2 = 1 or
2 days, 3 = 3 to 5 days, 4 = 6 to 8 days and 5 = 9 days or more. Bold data indicate P <0.05. aAdjusted for age,
sex, body mass index, and strenuous activity.
Page 93
6.4 Discussion
This population-based study of middle-aged adults has investigated the natural history of
BMLs over 8 years and the association between change in BML size and change in pain.
Incident BMLs were common; roughly one-half of those without BMLs at visit two
developed new BMLs by visit three. Of the BMLs present at visit two, 55% remained stable
while 24% increased and 21% decreased in size. An increase in BML size or new BML
resulted in a significant increase in knee pain, especially for male offspring. BMI and
strenuous activity independently predicted change in BML size.
This is the first study to report the natural history of BMLs over an extended period of time.
Many of the previous studies have been conducted over a much shorter timeframe. Davies-
Tuck and colleagues reported a much higher proportion of BMLs improving, with 46% of
BMLs resolving completely in a healthy, pain-free population over 2 years [163]. In a
symptomatic population with ROA, less than 1% of BMLs resolved or reduced in size [165].
The conflicting data may be a reflection of different study populations as well as different
grading systems used for the assessment of BMLs. In this study, patellar BMLs were also
assessed, which may explain the high percentage of participants with BMLs compared with
other studies that did not assess patella BMLs [93, 163, 164]. A previous study by our group
in a population-based sample of older adults that employed the same quantitative BML
methodology found very similar results to our current study where approximately one-quarter
of BMLs both increased and decreased in size [93].
In our study, incident BMLs in subjects without BMLs at visit two were also high, most
probably due to the period of follow-up. Other studies have reported much lower figures,
ranging from 9 to 14% in a healthy population over 2 years [163, 164] and 20% in a cohort
with symptomatic knee OA over 30 months [165]. Of clinical importance, the development
of an incident BML was significantly associated with the development of pain in those who
were pain free at visit two. This association has been corroborated by a prior study looking at
healthy populations [163] and a cohort consisting of subjects with OA or at high risk of OA
[76]. This observation further lends weight to the argument that BML development may be a
major contributor to incident knee pain.
Page 94
A 1 cm2 increase in BML size resulted in a 2.5 unit increase in knee pain in those with a
family history of OA, whereas no association was seen in controls. Previous findings from
this cohort have shown that offspring with a family history of knee replacement were more
likely to have a greater BMI, more knee pain, and less muscle strength cross-sectionally
compared with matched controls [102]. However, there have been few studies examining the
genetic factors influencing BMLs. Zhai and colleagues reported that BMLs have a significant
genetic component in this cohort [106], but they did not investigate whether there was a
genetic component to the role that BMLs play in pain. BMLs are perhaps more likely to
cause pain in genetically susceptible individuals given that genes can discriminate those with
OA and pain from those with OA without pain [183]. Alternatively, BML pathology may be
different in those with a family history of OA and MRI is somewhat nonspecific with regard
to the underlying pathology.
The association between change in BMLs and change in pain was also stronger in males
compared with females. Few studies have reported on sex difference in BMLs. Davies-Tuck
and colleagues reported that sex was not associated with the presence, development or
persistence of BMLs [163], whilst Dore and colleagues found that males were more likely to
have BMLs and have a BML increase over time [93]. In our cohort, males had significantly
larger BMLs at both visit two and visit three. When we adjusted our model for BML size at
visit two and visit three, however, the sex difference persisted, suggesting that there may be a
difference in the way BMLs mediate pain between sexes.
When we examined site-specific BMLs and their association with pain, we found that tibial
BMLs were associated with change in pain but not patellar or femoral BMLs. In particular,
medial tibial BMLs were strongly associated with change in pain for offspring. To the best of
our knowledge, no study has looked at site-specific associations between BMLs and pain.
Studies have shown that BMLs can lead to increased bone mineral density locally [168, 184]
and greater cartilage loss at the same site [162]. A local effect would thus be consistent with
the existing literature.
Higher BMI and strenuous activity were found to predict BML change. Obesity is a strong
risk factor for OA [49], and prior studies have also reported a cross-sectional association
between BMI and BML prevalence and severity [185]. However, a 36-month follow-up
Page 95
found no association between BML progression and BMI [185]. Similarly, high-intensity
physical activity has been shown to increase the risk of OA [186, 187]. However these
findings need to be balanced against the strong evidence demonstrating that physical activity
improves symptoms and physical function in OA [188]. A recent longitudinal study by our
group reported that physical activity measured by steps per day was deleteriously associated
with BML change [65]. Whilst we found a significant association between strenuous physical
activity and BML change, we did not find an association between light physical activity and
BML change, suggesting intensity of activity may be important. Therefore, whilst physical
activity may be good for symptoms, excessive physical activity may be detrimental to knee
structure. Randomized controlled trials evaluating the effect of physical activity on a
sensitive measure of knee structure, such as MRI, are needed to gain a better understanding
of this relationship.
Our study has several potential limitations. Firstly, slightly different MRI protocols were used
at visit two and visit three. Due to the long follow-up period, the protocol at visit three had
slightly different parameters – namely a smaller slice thickness. This means that a greater
number of small BMLs might have been picked up at visit three; the rate of incident BMLs
over 8 years may therefore not be as high as our study indicates. There may also have been an
overestimate of BML change, which would mean that the true magnitude of the association
between BML change and change in pain is stronger.
Secondly, due to the long follow-up period, a significant proportion of our subjects were lost
to follow-up. However, there were no significant differences in pain scores and demographics
between those lost to follow-up and participants in this study, suggesting this bias was not
systematic.
Thirdly, BML area was measured by taking the slice with the greatest BML size at a
particular site. This is a surrogate measure of volume and may overestimate shallow, flat
lesions. However, this method of BML measurement has proven to be sensitive to change in
a recent clinical trial [189].
Fourthly, the interslice gap was 0.5 to 1.0 mm, which means that small BMLs might have
been missed if they lay completely within the interslice gap, which seems unlikely.
Page 96
Fifthly, we used a subjective measure of physical activity to assess the amount of light and
strenuous activity that participants undertook. We also did not differentiate between different
modes of physical activity such as weight-bearing and nonweight-bearing exercise, and did
not specifically ask about strenuous incidental physical activity (for example, occupational or
household activity). It is also important to note that only current physical activity data were
captured over a 14-day period as opposed to a longitudinal measure of physical activity.
However, the significant correlation between strenuous activity and BML change is
consistent with pedometer-derived physical activity.
Sixthly, we did not assess analgesia such as paracetamol or nonsteroidal anti-inflammatory

drugs or knee malalignment, which could have been potential confounders.
Lastly, ROA was assessed at baseline and not at visit two. This difference in timing may
result in a slight underestimate in the prevalence of ROA at visit two, which may influence
the borderline results.
Conclusion
In this midlife cohort, the proportion of BMLs increasing in size was similar to those
decreasing in size, with the majority remaining stable. Change in BMLs can be predicted by
lifestyle factors, namely BMI and strenuous activity. An increase in BML size or a new BML
resulted in an increase in pain especially in males and those with a family history of OA.
Page 97
Chapter 7 Natural History of Meniscal Tears
Chapter Seven
Natural History and Clinical significance of

Meniscal Tears over 8 Years in a Mid-Life Cohort
Page 98
7.1 Background
Loss of meniscal function due to tears is a potent risk factor for knee osteoarthritis (OA) and
may be one of the earliest changes in the OA causal pathway [24]. Meniscal tears share
common risk factors with knee OA [94, 142] and explain more of the variation in joint space
narrowing (JSN) than cartilage volume [190]. Cross-sectional studies using magnetic
resonance imaging (MRI) have also shown that damage to menisci in the form of tears is
paralleled by other structural abnormalities such as lower cartilage volume [94] and an
increased severity of cartilage defects [94] and bone marrow lesions (BMLs) [191].
Although meniscal tears are a common finding in people with asymptomatic disease [192], it
is a potential source of pain associated with OA. The periphery of menisci have nociceptive
innervation [193, 194] and it is reasonable to hypothesise that meniscal tears that extend to
this area can cause pain. However longitudinal studies, conducted over 15-24 months, have
shown conflicting results thus far [75, 195]. It is uncertain if change in meniscal tears is
directly associated with worsening pain [195] or if both meniscal damage and pain are a
result of OA through intermediate pathologies (such as BMLs and effusion) rather than a
direct link between the two [75].
Furthermore, there is limited longitudinal data on the natural history of meniscal tears. It is
not clear how meniscal tears change over a long period of time and how change in meniscal
tears is associated with global knee structural changes. The aim of this study was to describe
the natural history of meniscal tears over 8 years, the predictors of change in meniscal tears
and the association between change in meniscal tears and change in knee pain and structures.
Page 99
7.2 Methods
This study was conducted as part of the Offspring study, a population-based study that began
in Southern Tasmania in June 2000. Matched sampling was used to recruit the study
participants (mean-age 47 (28–63) years; 57% females). Half of the participants were the
adult offspring of patients (only one parent) who had a knee replacement performed for
idiopathic knee OA at any Hobart hospital from 1996 to 2000 [102]. The diagnosis was
confirmed by reference to the medical records of the orthopaedic surgeons and the original
radiographs when possible. The other half were age and sex matched controls, randomly
selected from the population with no history of knee OA in either parent. This study includes
data from the first (visit-2) and second (visit-3) follow-up visits at approximately two and ten
years respectively, as we did not have the correct MRI sequence to score meniscal tears at
baseline.
obtained from all participants.
Participants were excluded if they had a contraindication to MRI (including metal sutures,
presence of shrapnel, iron filing in eye, or claustrophobia). Participants were also excluded if
they had undergone a knee replacement surgery or did so after the commencement of the
study. Knee pain and knee injury were not a basis for exclusion.
7.2.1 Knee pain

McMaster Universities Osteoarthritis Index (WOMAC) at both visits [93]. Five categories of
pain (walking on flat surface, going up or down stairs, at night, sitting or lying, and standing
upright) were assessed separately with a 10-point scale from 0 (no pain) to 10 (most severe
pain). Each category was summed to create a total pain score (range 0 to 50). Furthermore,
the five categories were clinically categorized into weight-bearing pain (including walking on
Page 100
flat surface, going up or down stairs and standing) and non-weight-bearing pain (including
pain at night and sitting or lying).
7.2.2 Knee joint injury
History of knee joint injury was assessed using a self-administered questionnaire [196] which
included the following questions:
• “Have you ever had a previous knee injury which resulted in non-weight bearing
treatment for 24 hours or more?”
• “If yes, then which knee?”
• “Please provide further details about the injury”
MRI of the right knee was performed as described previously [89, 104]. Knees were imaged
in the sagittal plane on a 1.5-T whole-body magnetic resonance unit (Picker International,
USA) using a commercial transmit-receive extremity coil. The following image sequence was
used: (1) a T1-weighted fat-suppressed 3D gradient-recalled acquisition in the steady state,
flip angle 55°, repetition time 58 msec, echo time 12 msec, field of view 16 cm, 60 partitions,
512×512–pixel matrix, slice thickness of 1.5 mm without an interslice-gap; and (2) a T2-
weighted fat saturation 2D fast spin echo, flip angle 90°, repetition time 3067 ms, echo time
112 ms, field of view 16 cm, 15 partitions, 256×256 matrix, slice thickness of 4 mm with an
interslice gap of 0.5–1.0 mm.
Meniscal tears were assessed by a trained observer (musculoskeletal radiologist with several
years of experience) on T2-weighted fat saturated (side by side) MR images at visit-2 and 3
of the study as previously described [110]. The proportion of the menisci affected by a tear
was scored separately (0-2 scale; 0=absence of a tear, 1=simple tear of different types:
longitudinal, oblique, radial or horizontal, 2=macerated tear signifying loss>50% area of
Page 101
meniscal tissue) at the anterior, middle, and posterior horns. Anterior, middle and posterior
scores were summed to create medial and lateral meniscal tear scores. The intra- and inter-
observer correlation coefficient (expressed as intraclass correlation coefficient (ICC)) ranged
from 0.86- 0.96 [111].
space, not including the osteophytes, was evaluated at visit-2 and 3 for the anterior, body, and
posterior horns of the menisci on T1-weighted gradient echo MR images, as previously
described [111]. A score from 0 to 2 was used (0 = no extrusion, 1 = partial meniscal
extrusion, and 2 = complete meniscal extrusion with no contact with the joint space). The
scores of anterior, body and posterior horns of medial or lateral menisci were summed to
create a total meniscal extrusion score for each of the medial and lateral tibiofemoral
compartments which had a possible range from 0 to 6. The intra- and inter-observer
correlation coefficient ranged from 0.85 to 0.92 for meniscal extrusion [110]. All knees were
evaluated for the presence of meniscal extrusion regardless of whether they had a meniscal
tear or not.
Tibial and femoral cartilage volume was assessed on T1-weighted gradient echo MR images
using Osiris (University of Geneva, Switzerland) and Cartiscope (ArthroLab, Montreal,
Canada) software respectively at visit-2 and 3, as previously described [104, 111]. The
coefficient of variation (CV) for intra-observer repeatability ranged from 2.0–2.2% for both
tibial and femoral cartilage volume measurements [84, 130]. Total cartilage volume was
calculated as: tibial + femoral cartilage volume.
Page 102
Cartilage defects were assessed on T1-weighted gradient echo MR images on a 0-4 scale
(0=normal; 1=focal blistering/signal changes; 2=<50% thickness loss; 3=>50% thickness
loss; 4=full thickness defect) at visit-2 and 3, as previously described [113]. Intraobserver
reliability ranged from ICC of 0.89-0.90 [113]. Interobserver reliability was assessed in 50
MR images and yielded an ICC of 0.85-0.90 [113].
BMLs were assessed on T2-weighted fat saturated MR images at visit-2 and 3 and were
defined as areas of increased signal adjacent to the subchondral bone [93]. One trained
observer scored the BMLs by measuring the maximum area of the lesion in a specific
compartment. The observer manually selected the MRI slice with the greatest BML size. The
BML with the highest score was used if more than one lesion was present at the same site.
The ICC for intra-observer reliability, assessed on 40 MR images, was 0.97.
7.2.9 Effusion
visit-2 and 3 on a 0-3 scale [114]. Grade-0 signified absence of fluid over the upper margin of
the patella in a sagittal image; Grade-1 signified some fluid above the upper margin of the
patella but the length of the fluid column shorter than that of the patella; Grade-2 signified a
fluid column above the upper margin of patella longer than the length of the patella; Grade-3
signified a fluid column above the upper margin of patella longer than the length of the
patella with a thickness of ≥ 1cm. Intra-observer reliability was assessed in 50 MR images
and yielded an ICC of 0.89-0.98. Pathological effusion was defined as any effusion score ≥2.
7.2.10 Radiography
A standing anteroposterior semiflexed view of the right knee (at 15° flexion) was performed
in all participants at baseline and 10 years. Radiographs were scored individually for
Page 103
osteophytes and joint space narrowing, as described previously [84]. Each of the following
four features was scored on a scale from 0 to 3 (0 = normal and 3 = severe): medial joint
space narrowing (JSN), lateral JSN, medial osteophytes (femoral and tibial combined) and
lateral osteophytes (femoral and tibial combined). Each score was arrived at by consensus
with two readers simultaneously assessing the radiograph with immediate reference to the
Osteoarthritis Research Society International (OARSI) atlas [82]. A non-zero score in either
joint space narrowing or osteophytosis was regarded as evidence of radiographic
osteoarthritis (ROA). Reproducibility was assessed in 50 radiographs, two weeks apart, and
yielded an ICC of 0.99 for osteophytes and 0.98 for JSN.
Readers for all the scans were either musculoskeletal radiologists with several years of
experience in OA research or health professionals trained by musculoskeletal radiologists.
Readers were not blinded to the chronological sequence of the radiographs and MRI scans.
Change in all MRI structures and leg strength was calculated as: Visit-3 score – Visit-2 score
T-test and Chi-square tests were used to describe the baseline characteristics of the
participants with or without any change in mean meniscal tear score. T-test was further used
to compare change in meniscal score between offspring and control groups. Poisson
regression analysis was used to examine the predictors of change in meniscal tears and the
association between change in meniscal tears and change in meniscal extrusion. Linear
regression analysis was used to describe the association between change in meniscal tears and
change in pain, cartilage volume loss and change in BMLs. Multivariable analyses were
adjusted for demographics, body mass index (BMI), offspring-control status and knee
structures (global knee structural factors known to be associated with the presence of
meniscal tears or knee pain). Further analysis was performed to explore any offspring-control
interaction in the multivariable models for all the above mentioned associations.
A P-value of less than 0.05 (two-tailed) was considered statistically significant. All statistical
analyses were performed on Intercooled Stata 12.0 for windows (StataCorp LP).
Page 104
7.3 Results
A total of 198 subjects (57% female, mean age 47 years) had complete MRI measures at
visit-2 and 3. There were no significant differences in baseline characteristics between those
lost to follow-up (n=133) and the participants in our study in terms of age, sex, BMI and
ROA (data not shown).
7.3.1 Natural History
Figure 7.1A describes the prevalence of meniscal tears at visit-2. 22% of the participants
(44/198) had at least one meniscal tear at any site. 41/44 participants had at least one
meniscal tear at any of the three meniscal sites (anterior, body or posterior) in the medial
compartment, whereas only 3 participants had at least one meniscal tear in the lateral
compartment. None of the participants had a meniscal tear in both compartments.
41 participants with medial meniscal tears had 55 meniscal tears in total at all sites. 29/41
participants had a single meniscal tear at any site (anterior, body or posterior), 10/41
participants had a meniscal tear at 2 sites and 2/41 participants had a meniscal tear at all 3
sites. Medial posterior was the most commonly affected site (27/55), followed by medial
body (21/55) and medial anterior sites (7/55) (Figure 7.1B). 37/55 meniscal tears were simple
tears, whereas 18/55 were macerated tears.
3 participants with lateral meniscal tears had 8 meniscal tears in total at all sites. 1/3
participant had a meniscal tear at 2 sites and 2/3 participants had meniscal tears at all 3 sites.
Lateral posterior was the most commonly affected site (4/8), followed by lateral body (3/8)
and lateral anterior sites (1/8) (Figure 7.1B). 5/8 meniscal tears were simple tears, whereas
3/8 were macerated tears.
The majority of participant’s menisci (84%) remained stable over 8 years. 16% of the
participants (31/198) showed an increase in mean meniscal score – including incident tears
(14/31) and increase in the severity of existing tears (17/31). Most of these changes affected
the medial meniscus (87% (27/31)).
Page 105
Figure 7.1. Prevalence and natural history of meniscal tears. A) Prevalence of meniscal tears
at visit 2, B) Site-specific distribution of meniscal tears at visit 2
Page 106
Most of the participants showed an increase at the posterior meniscal site (15/31), followed
by body (12/31) and anterior (4/31) sites. None of the participants with a meniscal tear at
visit-2 showed an improvement in meniscal tear score over 8 years.
Table 7.1 describes the (visit-2) characteristics of participants with and without any increase
in mean meniscal tear score over 8 years. Participants with any increase in mean meniscal
score were significantly older, heavier, had a higher percentage of offspring, prevalence of
ROA, total femoral cartilage volume, total mean cartilage defect score, tibial bone area and
prevalence of supra-patellar effusion compared to participants without any increase in mean
meniscal score. Participants with any increase in mean meniscal tear score also had a higher
percentage of male participants, worse pain score and a higher prevalence of BMLs but these
differences did not reach statistical significance.
The majority of meniscal tear change occurred in the offspring group and this was significant
at the total medial, total posterior and the total knee sites in comparison to the control group
(all p<0.05).
Table 7.1. Characteristics (at visit-2) of participants with and without any change (incident
tears and increase in score) in tears over 8 years
Any change (n= 31) No change (n= 167) p-value

Age (years) 50.06 ± 6.35 47.37 ± 6.49 0.046
Male (%)* 57 39 0.069
BMI (kg/m2) 29.51 ± 7.10 26.77 ± 4.38 0.008
Offspring (%)* 72 45 0.008
Any ROA (%)*^ 33 18 0.046
WOMAC pain (mean) 4.77 ± 7.14 2.63 ± 4.71 0.051
Total tibial cartilage vol (mm3) 4868.43 ± 1012.85 4500.62 ± 1062.66 0.093
Total femoral cartilage vol (mm3) 9562.64 ± 2377.24 8531.51 ± 2269.82 0.047
Total cartilage defects (mean) 5.24 ± 2.04 3.80 ± 1.43 <0.001
Total tibial bone area (mm2) 3273.17 ± 473.89 3079.38 ± 473.01 0.049
Any bone marrow lesion (%)* 59 50 0.380
Any pathological effusion (%)* 55 34 0.028
Mean ± standard deviation except for percentages; *Determined by Chi square test, others by t-test
^Assessed at the baseline visit; the rest assessed at visit-2
Bold font denotes statistically significant (p=<0.05) results
Page 107
7.3.2 Predictors of change
Table 7.2 describes predictors of change in total knee meniscal tears over 8 years. Age at
visit-2, BMI, history of knee injury, cartilage defects, BMLs, JSN and osteophytes
significantly predicted change in meniscal tears in unadjusted analysis. Only BMI and
osteophytes independently predicted change in meniscal tears in the fully adjusted model.
BMI showed a significant association in all compartments including anterior, body and
posterior meniscal sub-groups whereas osteophytes predicted change in only total anterior
and posterior tears (data not shown).
Table 7.2. Predictors of change in total knee meniscal tears over 8 years
Change in total knee meniscal tears over 8 years
Unadjusted Adjusteda
Risk ratio (95%CI) Risk ratio (95%CI)
1.06 1.05
Age (1.02, 1.11) (0.98, 1.21)
1.09 1.11
BMI (1.03, 1.15) (1.04, 1.17)
2.16 1.91
Knee Injury (1.08, 6.01) (0.93, 3.92)
1.26 0.77
Cartilage defects (1.05, 1.52) (0.54, 1.09)
1.57 0.87
BMLs (1.06, 2.32) (0.33, 2.29)
3.17 2.11
JSN (1.41, 7.16) (0.74, 6.03)
1.79 1.78
Osteophytes (1.29, 2.47) (1.17, 2.71)
a= adjusted for age/BMI/knee injury, offspring-control status, cartilage defects at visit-2, BMLs at
visit-2 and/or ROA at visit-1.
(No significant offspring-control interaction for any of the above mentioned associations)
7.3.3 Pain
30/44 participants who had a meniscal tear reported knee pain at baseline.
Table 7.3 describes the association between change in meniscal tears and change in pain over
8 years. Increases in total knee meniscal tears was independently associated with increases in
Page 108
total knee pain, pain on each individual WOMAC sub-scale and in weight bearing and non-
weight bearing pain over 8 years in the whole population. There was also a significant
offspring-control interaction at all sites with offspring showing significantly greater increases
in pain per unit increase in meniscal tears compared to controls.
Table 7.3. Association between change in meniscal tears and change in pain over 8 years
Change in pain over 8 years

Change in total knee Unadjusted Adjusteda
meniscal tears β (95%CI) β (95%CI)
Whole group +2.87 (+1.84, +3.90) +2.81 (+1.40, +4.22)
-- Offspring +3.73 (+2.56, +4.89) +2.84 (+1.22,+4.46)
-- Controls -0.48 (-2.72, +1.75) -0.92 (-4.20, +2.36)
Change in pain subscales over 8 years
Change in pain while lying in bed
Whole group +0.89 (+0.64, +1.14) +0.82 (+0.46, +1.18)
Change in pain while sitting
Whole group +0.45 (+0.22, +0.67) +0.35 (+0.04, +0.67)
Change in pain while standing
Whole group +0.55 (+0.31, +0.80) +0.62 (+0.31, +0.94)
Change in pain while walking on flat surface
Whole group +0.56 (+0.35, +0.77) +0.49 (+0.20, +0.78)
Change in pain while climbing stairs
Whole group +0.33 (+0.02, +0.65) +0.59 (+0.15, +1.02)
Change in pain in non-weight bearing
Whole group +1.34 (+0.90, +1.78) +1.18 (+0.56, +1.80)
Change in pain in weight bearing
Whole group +1.49 (+0.82, +2.16) +1.66 (+0.75, +2.58)
a= Adjusted for age, sex, BMI, offspring-control status, change in BMLs, change in cartilage defects,
change in meniscal extrusion, change in effusion, history of knee injury and ROA at visit-1.
(Note: Significant offspring-control interaction at all sites and sub-scales for the association
between change in meniscal tears and change in pain)
Page 109
7.3.4 Structural changes
Table 7.4 describes the association between change in meniscal tears and knee structures on
MRI over 8 years. Change in meniscal tears was independently associated with cartilage
volume loss in the medial compartment only, increases in medial, lateral and total
tibiofemoral BML area and with a higher risk of change in medial meniscal extrusion.
There was no significant association between change in meniscal tears and change in
cartilage defects at any site in the fully adjusted model.
Only two participants underwent knee surgery between baseline and visit-3 and on both
occasions the surgery was not a menisectomy or a joint replacement. Further adjustment for
knee surgery did not change the effect size considerably for any of the associations described
earlier (data not shown).
Table 7.4. Association between change in meniscal tears and knee structures on MRI over 8
years
β (95%CI) β (95%CI) Risk ratio(95%CI)

Adjusteda Adjusteda Adjusteda
Change in tears (site) Cartilage volume loss Change in BMLs Change in meniscal
extrusion
Total tibiofemoral Total tibiofemoral Total knee
Total knee -52 (-208, +102) +0.41 (+0.29, +0.52) N/A
Medial tibiofemoral Medial tibiofemoral Medial meniscus
Total medial -176 (-302, -49) +0.33 (+0.22, +0.43) 1.53 (1.14, 2.03)
Lateral tibiofemoral Lateral tibiofemoral Lateral meniscus
Total lateral +143 (-731, +1018) +0.26 (+0.10, +0.41) N/A
a= adjusted for age, sex, bmi, offspring-control status, cartilage volume loss, change in BMLs,
cartilage defects and meniscal extrusion, and ROA at visit-1
(No significant offspring-control interaction at any site for the association between change in
meniscal tears and change in BMLs)
Note: Not enough change in lateral meniscal extrusion for analysis due to lack of power
Page 110
7.4 Discussion
This study documents the natural history of meniscal tears over 8 years. In this midlife cohort
meniscal tears were common with 22% of the participants suffering from at least one. 16% of
the participants showed an increase in severity and none improved over 8 years. BMI and
osteophytes independently predicted an increase in meniscal tears over 8 years. Change in
meniscal tears was independently associated with an increase in knee pain severity, with
offspring showing a greater increase in the severity of pain per unit change in meniscal tears
compared to the control group. Change in meniscal tears was independently associated with
cartilage volume loss, change in BMLs and meniscal extrusion over 8 years.
Majority of the meniscal tears (55/63) at visit-2 affected the medial meniscus. Medial
posterior site showed the highest prevalence followed my medial body sites. Previous studies
by Englund et al. [192] in older adults and by K. A Beattie et al. [197] in middle-aged adults
showed a similar distribution in cross-sectional studies as well. Although the majority of the
menisci remained stable over the course of 8 years, 16% showed an increase in severity over
time. Again medial posterior was the most commonly affected site for both incident meniscal
tears and worsening meniscal tear grades. Of note, none of the meniscal tears improved over
the course of the study, unlike other knee structures such as BMLs [198] and cartilage defects
[89] as previously shown in this cohort. Previously Dillon et al. [199] followed 22 patients
with 27 intra-meniscal lesions with signal intensity changes on MRI but no tears on
arthroscopy. After 27 months only 2 completely disappeared. Similarly Boegard et al. [200],
followed 47 patients and found that only 2 meniscal tears out of 54 improved and none
disappeared over 2 years. Meniscal tears, unlike other knee structures, do not seem to have
the capacity to regenerate or improve over time. Slight discrepancies in the above mentioned
studies could be due different populations, a longer follow-up period resulting in less
measurement error in the present study and a possibly a more severe disease process in the
offspring sub-group.
High BMI was the most consistent independent risk factor for increase in meniscal tear
severity. A previous cross-sectional study from the present cohort showed that a higher BMI
is positively associated with prevalent meniscal tears [94]. Our findings are consistent with
Baker et al. [201] but differ from Englund et al. [142], who found a significant association
Page 111
between BMI and meniscal extrusion but not tears. A recent meta-analysis examining risk
factors for meniscal tears concluded that a high BMI is a moderate risk factor for developing
meniscal tears along with occupational and recreational joint loading [202]. Osteophytes at
visit-1 also predicted worsening of meniscal tears. Osteophytes are thought to be an early
instigating factor in the OA causal pathway and their true prevalence is under estimated on
radiographs [203]. Beattie et al. [197] showed, using peripheral MRI, that many peripheral
osteophytes are missed by standard radiographs and their presence corresponds with
degenerative meniscal changes at the same site. Presence of osteophytes in our study also
showed a significant association with change in meniscal tears at the peripheries (anterior and
posterior) and not at the meniscal body site. Interestingly, history of knee injury was not
independently associated with meniscal tear increase. Previously, Englund et al. [142] have
shown that history of knee injury is a strong risk factor for developing meniscal tears but they
did not adjust for potential confounders. Similarly, we found a significant association
between knee injury and meniscal tears in unadjusted analysis but this association did not
persist in the fully adjusted model. These findings suggest that the changes in meniscal tears
are not due to mechanical factors only and are mainly a part of an active osteoarthritic
process.
Previously in this cohort, we showed a cross-sectional association between presence of

meniscal tears and increased pain [94]. In a longitudinal study, Zanetti et al. [195] found that
asymptomatic participants with a meniscal tear are more likely to develop knee pain than
participants without one. Englund et al. [75] on the other hand concluded that any association
between meniscal damage and knee pain seems to be present because both pain and meniscal
damage are related to OA and not because of a direct link between the two. Our study is the
first study to show an independent longitudinal association between increasing severity of
meniscal tears and worsening pain, including pain on all individual WOMAC sub-scales, as
well as both weight bearing and non-weight bearing pain. Previous studies have also
suggested that meniscal tears appear to cause symptoms only when macerated tears extrude
and damage collateral ligaments or when bone marrow abnormalities are present [204].
Results in this study were independent of change in meniscal extrusion and BMLs as well as
localised inflammation as assessed by knee effusion, suggesting meniscal tears may be one of
the most important knee structures in relation to pain.
Page 112
Every unit increase in meniscal tears in the offspring group resulted in a greater increase in
pain compared to the controls. Previously in this cohort, we found similar differences
between the two groups when looking at the association between change in BMLs and pain
[198]. A possible explanation could be the differences in the pain perception pathways of the
two groups. Of note, polymorphisms in COMT and TRPV1 genes have recently been
identified which could alter the processing of nociceptive pain associated with OA [43].
Another possible explanation could be that meniscal pathology in the offspring is
morphologically different but this could not be differentiated on MRI.
Biomechanical studies have shown that the function of the meniscus is to reduce contact
stress by enlarging the contact surface and shock absorption [205]. Meniscal function can be
either lost due to meniscal tears or meniscal extrusion. Meniscal tears, especially macerated
tears, are a possible risk factor for meniscal extrusion [206] and findings from this study
confirm this. Loss of meniscal function can potentially damage articular cartilage and sub-
chondral bone. Cross-sectional studies have shown that prevalent meniscal tears are
associated with decreased cartilage volume [94] and BMLs [207]. Chang et al. [208] showed
that meniscal tears are longitudinally associated with site specific cartilage loss. Findings in
this study are in agreement with the latter study, as we found that meniscal tear increases
were associated with medial cartilage loss independent of other knee structural changes. The
present study is also the first to show a longitudinal association between increase in meniscal
damage and increase in BML size. Menisci aid in load distribution and BMLs have been
shown to be a consequence of abnormal loading within the knee joint [209], which explains
the association between the increasing severity of these structural abnormalities. High BMI
and osteophytes are possibly the early instigating factors that predict increasing severity of
meniscal tears and then change in meniscal tears is associated with other structural changes
such as meniscal extrusion, cartilage volume loss and BMLs.
A strength of our study is that it has the longest follow-up period of any OA cohort using
MRI. A limitation of our study is a significant loss to follow up. Loss to follow-up can be a
potential source of bias, however re-analysis of the data using inverse probability weighting
did not change any of the results, indicating robust results. This cohort also has a wide age
range (28-63 years old) as the inclusion criteria did not specify any specific age range.
However, all the results described in this study were adjusted for age. Another limitation was
Page 113
the absence of radiographs at visit-2 of the study, as we did not anticipate any major changes
on radiographs due to young mean-age of the cohort with a low osteoarthritis disease burden
and a short follow-up period of 2 years. Our study demonstrated an independent association
between change in meniscal tears and worsening knee pain. However, the changes in the knee
OA are remarkably collinear. Although we did adjust for other co-pathologies accounting for
knee pain, an ideal design would be a long-term study with global knee structural assessment
at multiple time points and a case-cross over design. Furthermore, we did not analyse
different types of simple tears (longitudinal, oblique, radial or horizontal) separately due to a
low number of individual lesions and hence insufficient power for analysis.
Conclusion
Change in meniscal tears shares common risk factors with knee OA and is independently
associated with worsening knee pain and structural damage suggesting that meniscal tears are
on the knee OA causal pathway.
Page 114
Chapter 8 Natural History of Cartilage Defects
Chapter Eight
Natural History and Clinical Significance of

Cartilage Defects over 10 Years
Page 115
8.1 Introduction
Knee cartilage defects are a common finding in young healthy adults and in people with early
osteoarthritis (OA) [89, 210] when the prevalence of other co-pathology is lower [211].
Initially they were thought to result from knee trauma only but recent evidence suggests
otherwise [196]. Studies using MRI to study OA have suggested that cartilage defects can
result as part of an active OA process [210, 212]. Cartilage degeneration in the form of
defects is an early instigating factor in OA cascade and is thought to precede cartilage volume
loss [113, 141] and is associated with radiographic osteoarthritis (ROA) [113, 212].
Our understanding of the pathophysiology of cartilage defects is however incomplete. Short

term data from the present cohort over 2 years [89] showed that prevalent cartilage defects
and change in defects predicted site specific cartilage volume loss. Another study, using a
similar methodology, found a similar association at the patellar site only [213]. Secondly, the
association between cartilage defects and pain or function is controversial as cartilage is an
aneural structure. A study by Baum et al. [90] suggested that only prevalent focal cartilage
lesions are significantly associated with knee pain and found no such associations with
meniscal pathologies, bone marrow lesions (BMLs), effusion and ligamentous lesions. A
recent systematic review [96] looking at the association between knee abnormalities and knee
pain found no significant association between cartilage defects and pain. Lastly, studies have
mostly focused on the medial tibiofemoral compartment. However, the prevalence of
cartilage defects is relatively higher in the lateral compartment compared to the other
abnormalities such as meniscal pathology [212] and BMLs [214, 215], and might play a more
crucial role in lateral compartment OA progression.
Studies looking at the natural history of cartilage defects thus far have been short-term with
no more than two study time points and did not account for global knee structural
pathologies. So the aim of this study was to describe the natural history, predictors and
structural/ symptomatic correlates of cartilage defects in a midlife cohort over 10 years.
Page 116
8.2 Methods
This study was conducted as part of the Offspring study [102], a population-based study that
began in Southern Tasmania in June 2000. Matched sampling was used to recruit the study
participants. Half of the participants were the adult offspring of patients who had a knee
replacement performed for idiopathic knee OA at any Hobart hospital from 1996 to 2000
[152]. The diagnosis was confirmed by reference to the medical records of the orthopaedic
surgeon and the original radiographs when possible. The other half were age and sex matched
controls, randomly selected from the population with no history of knee OA in either parent.
This study includes data from the baseline visit, 2 year and 10-year follow-up visits.
the protocol, and written informed consent was obtained from all participants. Participants
were excluded if they had a contraindication to MRI, underwent knee replacement surgery or
did or after the commencement of the study. Knee pain and knee injury were not a basis for
exclusion.
Weight was measured to the nearest 0.1 kg (Seca Delta Model 707). Height was measured to
the nearest 0.1 cm using a stadiometer. Body mass index (BMI) was calculated (kg/m2).
• Knee joint injury and surgery

• History of knee joint injury was assessed at all three visits using a self-administered
questionnaire [196] which included the following questions:
• Only right knee injuries were included in the analysis as MRI scans were on the right
knee.
Page 117
8.2.3 Knee pain
McMaster Universities Osteoarthritis Index (WOMAC) [93] at visit-2 and 3. Five categories
of pain (walking on flat surface, going up or down stairs, at night, sitting or lying, and
standing upright) were assessed separately with a 10-point scale from 0 (no pain) to 10 (most
severe pain). Each category was summed to create a total pain score (range 0 to 50).
Furthermore, the five categories were clinically categorized into weight-bearing pain
(including walking on flat surface, going up or down stairs and standing) and non-weight-
bearing pain (including pain at night and sitting or lying).
International, USA) using a commercial transmit-receive extremity coil. The following image
sequence was used: (1) a T1-weighted fat-suppressed 3D gradient-recalled acquisition in the
steady state, flip angle 55°, repetition time 58 ms, echo time 12 ms, field of view 16 cm, 60
partitions, 512×512–pixel matrix, slice thickness of 1.5 mm without an inter-slice-gap (at all
three visits); and (2) a T2-weighted fat saturation 2D fast spin echo, flip angle 90°, repetition
time 3067 ms, echo time 112 ms, field of view 16 cm, 15 partitions, 256×256 matrix, slice
thickness of 4 mm with an inter-slice gap of 0.5–1.0 mm (at visit 2 and 3 only).
Cartilage defects were assessed on T1-weighted gradient echo MR images at the baseline,
visit-2 and visit-3. Cartilage defects were graded at the medial tibial, medial femoral, lateral
tibial, lateral femoral and patellar sites on a 0-4 scale (grade-0=normal cartilage; grade-
1=focal blistering and intra-cartilaginous low-signal intensity area with an intact surface and
bottom; grade-2=irregularities on the surface or bottom and loss of thickness of less than
50%; grade-3=deep ulceration with loss of thickness of more than 50%; grade-4=full-
Page 118
thickness chondral wear with exposure of sub-chondral bone), as previously described. A

cartilage defect also had to be present in at least two consecutive slices. If multiple defects
existed at one site, the highest grade was used. Intra-observer reliability (expressed as intra-
class correlation coefficient (ICC)) ranged from 0.89-0.90. Inter-observer reliability was
assessed in 50 MR images and yielded an ICC of 0.85-0.90.
Tibial and patellar cartilage volumes were assessed on T1-weighted gradient echo MR
images using Osiris (University of Geneva, Switzerland) software as previously described
[104, 109]. Femoral cartilage volume was determined using Cartiscope (ArthroLab,
Montreal, Canada), as previously described [110, 111].
Absolute cartilage volume loss was calculated as: follow-up total cartilage volume - baseline
Meniscal tears were assessed on T2-weighted fat saturated MR images at visit-2 and 3 of the
study as previously described [110]. The proportion of the menisci affected by a tear was
scored separately (0-2 scale; 0=absence of a tear, 1=simple tears of different types
(longitudinal, oblique, radial or horizontal) signifying loss<50% area of meniscal tissue,
2=complex tear signifying loss>50% area of meniscal tissue) at the anterior, middle, and
posterior horns. Anterior, middle and posterior scores were summed to create medial and
lateral meniscal tear scores.
space, not including the osteophytes, was evaluated at baseline, visit-2 and visit-3 for the
anterior, body, and posterior horns of the menisci, as previously described [111]. A score
Page 119
from 0 to 2 was used (0=no extrusion, 1=partial meniscal extrusion, and 2=complete meniscal
extrusion with no contact with the joint space). The scores of anterior, body and posterior
horns of medial or lateral menisci were summed to create a total meniscal extrusion score for
each of the medial and lateral tibio-femoral compartment [110].
BMLs were assessed on T2-weighted fat saturated MR images at visit-2 and 3 at medial
tibial, lateral tibial, medial femoral, lateral femoral and patellar sites. BMLs were defined as
areas of increased signal adjacent to the subcortical bone and were measured as the maximum
area of the lesion in a specific compartment, as described previously [198].
8.2.10 Effusion
visit-2 and 3 on a 0-3 scale as previously described [217]. Pathological effusion was defined
as any effusion score ≥2.
8.2.11 Radiography
A standing antero-posterior semi-flexed view of the right knee (at 15° flexion) was performed
in all participants at baseline and visit-3. Radiographs were scored individually for
osteophytes and JSN on a scale of 0–3 [84], according to the OARSI guidelines [82].
Radiographs were read as paired by two readers simultaneously. The presence of ROA was
defined as any score ≥1 for JSN or osteophytes.
For all the structures described above, medial and lateral tibio-femoral scores were added
together to generate total tibio-femoral scores. Similarly, total tibio-femoral and patellar
scores were added to generate total knee scores.
Page 120
experience in OA research or health professionals trained by musculoskeletal radiologists.
Readers were not blinded to the chronological sequence of the radiographs and MRI scans.
Change in all MRI and radiographic structures were calculated as: Visit-3 score –
Baseline/Visit-2 score.
Change in total knee cartilage defects was dichotomised as any increase or no increase
(decrease in severity or a stable score) in the severity of the mean cartilage defects score over
ten years to describe the baseline characteristics of the study participants.
T-tests and Chi-square tests were used to describe the baseline characteristics of the
participants with or without any increase in mean cartilage defects score.
Descriptive analyses were used to describe the site-specific natural history of cartilage defects
over 10 years. T-tests were used to describe the differences between the offspring and control
groups for site-specific change in mean cartilage defects and the difference in the number of
incident or new cartilage defects at each site.
Box-plots were used to describe the association between compartment specific severity of
cartilage defects at baseline and cartilage volume loss over 10 years.
Linear regression analysis was used to examine the predictors of change in cartilage defects
and to describe the association between change in cartilage defects and change in WOMAC
pain scale and other structural pathologies assessed on MRI. Linear regression using mixed
methods analysis was used to examine the association between change in cartilage defects
and cartilage volume loss as both were assessed at all three time points.
A P-value of less than 0.05 (two-tailed) was considered statistically significant. All statistical
analyses were performed on Intercooled Stata 12.0 for windows (StataCorp LP).
Page 121
8.3 Results
Of the 372 participants included in the Offspring study, 220 between the ages of 26 and 61
years were followed-up for 10 years. 5 participants were excluded from the present study due
to missing data. There were no significant differences in the baseline characteristics between
those lost to follow-up (n=152) and the participants in our study in terms of age, sex, BMI,
baseline prevalence of cartilage defects and radiographic OA (data not shown).
Figure 8.1A describes the site-specific prevalence of cartilage defects at baseline. 44% of the
participants had at least one cartilage defect at any site at baseline. The patellar site (25.1%)
had the highest prevalence followed by medial tibial (14.4%), medial femoral (13%), lateral
tibial (10.4%) and lateral femoral (8.7%) sites. Figure 8.1B describes the site-specific change
in cartilage defects over 10 years. Most of these defects remained stable over 10 years. 26%
increased in severity with medial tibial (30%) being the most commonly effected site. 13% of
the cartilage defects present at baseline decreased in severity over 10 years with patellar site
having the highest percentage (17%) of defects that decreased in severity.
Offspring showed a higher increase in the severity of the cartilage defects compared to the
controls, however these differences were statistically significant only at the medial tibial and
the medial femoral sites (data not shown). There was a similar trend for incident (new
cartilage defects not present at baseline) cartilage defects but the difference was statistically
significant only at the medial femoral site (data not shown).
Table 8.1 describes the baseline characteristics of participants with and without any increase
in mean cartilage defects score over 10 years. Participants with any increase in cartilage
defects severity had a significantly higher prevalence of JSN and BMLs compared to
participants with no increase.
Figure 8.2 describes the association between compartment specific cartilage defects at
baseline and absolute cartilage volume loss over 10 years. Per unit increase in severity of
lateral tibio-femoral but not medial cartilage defects predicted cartilage volume loss over 10
years.
Table 8.2 describes the predictors of change in cartilage defects. Both the presence and
severity (per grade) of JSN and osteophytes and severity (per grade) of supra-patellar
Page 122
effusion predicted change in tibio-femoral and total knee cartilage defects in unadjusted
analyses but these associations only persisted for JSN in the fully adjusted model. Only the
severity (per unit area) of BMLs independently predicted change in tibio-femoral cartilage
defects. There were no significant associations between meniscal pathologies and change in
cartilage defects.
Figure 8.1. Natural history of cartilage defects. A) Site specific prevalence of cartilage
defects at baseline. B) Site-specific change in cartilage defects over 10 years
Page 123
Table 8.1. Characteristics of participants with and without any increase in mean cartilage
defects score over 10 years
Any increase No increase p-value

(n=146) (n=69)
Age (years) 45.1 ± 6.7 45.5 ± 6.9 0.691
Male (%)* 42 % 41 % 0.794
BMI (kg/m2) 27.1 ± 4.9 27.5 ± 5.1 0.599
Offspring (%)* 56 % 45 % 0.123
Ever smoked (%)* 42 % 46 % 0.525
Knee Injury (%)* 19 % 17 % 0.753
Any JSN (%)* 20 % 10 % 0.045
Any osteophytes (%)* 13 % 9% 0.357
WOMAC (mean) 2.8 ± 4.9 2.9 ± 5.7 0.818
Total knee cartilage volume (mm3) 17739.1 ± 4330.8 16824.3 ± 3563.4 0.151
Any meniscal tear (%)* 23 % 15 % 0.218
Any meniscal extrusion (%)* 9% 6% 0.431
Any effusion (%)* 42 % 29 % 0.079
Any bone marrow lesion (%)* 70 % 53 % 0.021
Mean ± standard deviation except for percentages; *Determined by Chi square test, others by
t-test
Page 124
Figure 8.2. Association between cartilage defects at baseline and cartilage volume loss over
10 years. A) Medial tibio-femoral compartment. B) Lateral tibio-femoral compartment. C)
Patellar compartment
Page 125
Table 8.2 Predictors of change in cartilage defects
Unadjusted Adjusteda Unadjusted Adjusteda

β (95%CI) β (95%CI) β (95%CI) β (95%CI)
Change in cartilage defects
Total tibiofemoral Total knee

-0.09 -0.38 +0.01 -0.39
Age (-0.58, +0.41) (-0.94, +0.17) (-0.56, +0.58) (-1.02, +0.25)
-0.26 -0.03 -0.27 -0.24
Sex (Female) (-0.76, +0.23) (-0.82, +0.76) (-0.85, +0.31) (-0.85, +0.96)
+0.02 +0.01 +0.02 +0.01
BMI (-0.03, +0.07) (-0.04, +0.06) (-0.03, +0.08) (-0.05, +0.07)
+0.14 -0.25 +0.41 -0.07
Knee Injury (-0.49, +0.77) (-0.95, +0.45) (-0.32, +1.14) (-0.87, +0.74)
+0.95 (+0.16, +0.83 +1.05 +1.13
Any JSN +1.31) (+0.32, +1.60) (+0.31, +1.80) (+0.10, +2.18)
+0.79 +0.86 +0.89 +1.04
Severity of JSN (+0.21, +1.38) (+0.02, +1.70) (+0.19, +1.58) (+0.10, +1.99)
+0.79 +0.47 +0.91 +0.42
Any osteophytes (+0.04, +1.53) (-0.74, +1.69) (+0.03, +1.79) (-0.96, +1.81)
Severity of +0.56 +0.12 +0.62 +0.03
osteophytes (+0.21, +0.91) (-0.51, +0.75) (+0.21, +1.03) (-0.69, +0.74)
Any meniscal +1.40 +0.72 +1.54 +0.87
extrusion (+0.50, +2.30) (-0.40, +1.83) (+0.50, +2.58) (-0.40, +2.14)
Severity of meniscal +0.66 +0.12 +0.79 +0.22
extrusion (+0.18, +1.15) (-0.49, +0.74) (+0.22, +1.36) (-0.48, +0.91)
+0.21 +0.18 +0.32 +0.34
Any meniscal tear* (-0.36, +0.78) (-0.47, +0.83) (-0.30, +0.95) (-0.37, +1.04)
Severity of meniscal +0.19 +0.03 +0.22 +0.05
tears* (-0.02, +0.41) (-0.27, +0.34) (-0.02, +0.46) (-0.29, +0.38)
+0.42 +0.49 +0.59 +0.38
Any BMLs* (-0.05, +0.89) (-0.01, +0.99) (+0.05, +1.14) (-0.20, +0.96)
+0.67 +0.64 +0.33 +0.36
Severity of BMLs* (+0.13, +0.78) (+0.10, +1.20) (-0.07, +0.72) (-0.09, +0.85)
+0.16 -0.18 +0.14 -0.34
Any effusion* (-0.31, +0.62) (-0.70, +0.34) (-0.38, +0.66) (-0.91, +0.22)
Severity of +0.25 +0.02 +0.40 +0.10
effusion* (+0.06, +0.43) (-0.17, +0.21) (+0.15, +0.64) (-0.16, +0.35)
aAdjusted for age, sex, bmi, offspring-control status, meniscal tears, meniscal extrusion, cartilage volume, bone
marrow lesions, pathological effusion and radiographic osteoarthritis
*Predicted change in cartilage defects over 8 years (between visits 2 and 3)
Page 126
Table 8.3 describes the association between change in cartilage defects and structural changes
assessed on MRI. Change in cartilage defects was independently associated with absolute
cartilage volume loss (using mixed method analysis) at the lateral tibio-femoral, total tibio-
femoral and total knee sites. There was an independent association between change in BMLs
and change in cartilage defects in the lateral tibio-femoral compartment only. There were no
significant associations between change in cartilage defects and change in meniscal
pathologies or effusion.
Table 8.4 describes the association between change in cartilage defects and change in
WOMAC (pain) over 8 years. Change in cartilage defects was significantly associated with
change in pain at the lateral and total tibio-femoral compartments in the unadjusted analysis
but these associations did not persist in the fully adjusted model.
Page 127
Table 8.3. Association between change in cartilage defects and structural changes assessed on MRI
Change in meniscal Change in meniscal Change in Change in

Cartilage volume loss extrusion tears bone marrow lesions pathological effusion
Change in cartilage Adjusted Adjusted Adjusted Adjusted Adjusted

defects (site) β (95%CI) β (95%CI) β (95%CI) β (95%CI) β (95%CI)
-33.70 -0.01 -0.09 -0.02 +0.23

Medial tibiofemoral (-88.49, +21.09) (-0.40, +0.38) (-0.41, +0.23) (-0.08, +0.04) (-0.10, +0.57)
-77.67 -0.21 +0.07 -0.05

Lateral tibiofemoral (-137.51, -17.84) Not enough change (-0.50, +0.09) (+0.01, +0.13) (-0.26, +0.17)
-34.64 -0.06 +0.16 -0.01 -0.09

Patellar (-90.90, +21.59) (-0.26, +0.15) (-0.03, +0.29) (-0.09, +0.06) (-0.27, +0.10)
-152.52 +0.13 -0.22 +0.06 +0.24

Total tibiofemoral (-210.60, -94.44) (-0.30, +0.56) (-0.52, +0.08) (-0.01, +0.13) (-0.15, +0.62)
-123.87 +0.08 -0.02 +0.05 +0.15
Total knee (-193.24, -54.50) (-0.39, +0.55) (-0.35, +0.31) (-0.02, +0.12) (-0.28, +0.58)
aAdjusted for age, sex, bmi, offspring-control status, cartilage volume loss, change in meniscal tears, change in meniscal extrusion, change in BMLs, change in supra-patellar
effusion and radiographic osteoarthritis at baseline
Page 128
Table 8.4. Association between change in cartilage defects and change in pain
Unadjusted Adjusted
β (95%CI) β (95%CI)
Change in cartilage defects
(site) Change in WOMAC (pain) over 8 years
Medial tibiofemoral +0.36 (-0.31, +1.05) +0.48 (-0.35, +1.32)
Lateral tibiofemoral +1.21 (+0.17, +2.25) +1.17 (-0.18, +2.52)
Patellar -0.34 (-1.65, +0.86) -0.95 (-2.70, +0.75)
Total tibiofemoral +0.61 (+0.05, +1.18) +0.55 (-0.20, +1.30)
Total knee +0.43 (-0.08, +0.94) +0.24 (-0.45, +0.93)
aAdjusted for age, sex, bmi, offspring-control status, change in bone marrow lesion, change in meniscal tears,
change in meniscal extrusion, change in supra-patellar effusion and radiographic osteoarthritis at baseline
Page 129
8.4 Discussion
This study documents the natural history of cartilage defects over 10 years with data available
at 3 time points. In this midlife cohort cartilage defects were common with 44% of the
participants suffering from at least one at the baseline visit. Most of these defects remained
stable, whereas 26% increased and 13% decreased in severity over 10 years. Severity of
cartilage defects at baseline predicted cartilage volume loss only in the lateral tibio-femoral
compartment. Presence and severity JSN and severity of BMLs and family history of knee
OA predicted an increase in cartilage defects over 10 years. Change in cartilage defects in
turn was associated with changes in BMLs and cartilage volume loss. There was no
independent association between change in cartilage defects and increase in pain severity
over 10 years.
A minority of the cartilage defects present at baseline increased in severity over 10 years in
this cohort. In previous longitudinal MRI studies, progression rates of cartilage defects based
on semi-quantitative scoring have been reported to vary from 17% to 68% [89, 200, 210, 218-
220]. This wide range of progression rates is due to a number of factors. Amin et al. [218]
reported progression of cartilage damage in 46% and 22% of knees in the medial and lateral
compartments, respectively, over 2.5 years in symptomatic subjects. Similarly, Davies-Tuck
et al. [219] reported worsening of cartilage defects of 32-68% at different joint sites over 2
years. However, both of these cohorts had a higher prevalence of ROA compared to the
Offspring study. The former cohort consisted of participants of whom 72% had a (Kellgren
and Lawrence) KL grade ≥ 2 whereas the latter was a convenience sample of subjects with
ROA. Several studies [218-220] have shown the progression of cartilage defects increases
with more severe radiographic disease, which explains a lower rate of progression in this
study despite a considerably longer follow-up period. In another study Cibere et al. [210]
showed progression rate of 22.7% over 3 years despite the fact that their cohort was a bit
older and had a slightly higher prevalence of ROA compared to the Offspring study.
However, they used a more conservative definition of cartilage defects progression that can
explain a slightly lower progression rate in their study.
Regression of cartilage defects is a controversial topic. In this cohort of young healthy

middle-aged adults 13% of the cartilage defects decreased in severity over 10 years.
Page 130
Experimental studies in animals have shown that cartilage has significant capacity to self-
repair especially for small sized hyaline or fibrocartilage defects [221, 222]. We have
previously shown that regression of cartilage defects is rare in older adults [141], using the
same assessment protocol, indicating that fibrocartilage loses the ability to self-repair as we
age. This could reflect declining mitotic and synthetic activity in chondrocytes that occurs
with age in cartilage [223], with fewer cartilage defects regressing over time due to less self-
repair in older adults. Similarly, some recent studies in middle-aged cohorts, using different
assessment protocols, have shown that cartilage defects can regress over time [219, 224].
Cartilage defects are traditionally thought to precede cartilage volume loss, however not
many studies have actually shown this relationship. Cross-sectional studies have shown that
prevalent cartilage defects are negatively associated with cartilage volume and positively
associated with cartilage breakdown products [113]. Two-year data from the Offspring study
[108] and another longitudinal study [225] conducted in 86 middle-aged men and women in
Melbourne showed that baseline cartilage defects predict cartilage volume loss in the medial
tibiofemoral compartment. Data from this study is somewhat consistent with these findings as
cartilage defects independently predicted cartilage volume loss in the lateral compartment.
Earlier studies did not account for both the BMLs and meniscal tears when looking at the
association between cartilage defects and cartilage volume loss. 8-year longitudinal data from
the Offspring study has shown that BMLs [198] and meniscal tears [226] are more common
in the medial compartment and are associated with cartilage volume loss. After adjustment
for these structural co-pathologies, we did not see any independent association between
cartilage defects and cartilage volume loss in the medial compartment. Similarly, we have
previously shown that baseline cartilage defects predict cartilage volume loss in a randomly
selected sample of older adults from the community in all three compartments but the
associations in the medial compartment did not persist after adjustment for BMLs [141].
However this study could not account for meniscal pathologies as all subjects had them. The
longitudinal relationship between change in cartilage defects and other structural co-
pathologies showed a similar trend. Change in cartilage defects was independently associated
with cartilage volume loss, using mixed methods over three time-points, in the lateral
compartment but not the medial and patellar compartments. Similarly change in cartilage
defects was associated with change in BMLs in the lateral compartment only. Cartilage
Page 131
defects are relatively more common in the lateral compartment compared to other co-
pathologies and possibly play a more crucial role in the development of the overall disease
process in the lateral compartment.
Several studies have shown a positive association between cartilage defects and knee pain,
despite the fact that cartilage is an aneural structure. In a cross-sectional sub-sample of 126
middle-aged adults from Osteoarthritis Initiative (OAI) database, Baum et al. [90] found that
elevated T-2 cartilage signals are associated with findings of pain in the early phase of OA,
whereas among morphologic knee abnormalities only knee cartilage lesions are significantly
associated with knee pain. Similarly, Javaid MK et al. [227] found that cartilage defects with
a score >2 can significantly discriminate between painful and non-painful knees. However, in
a recent systematic review Yusuf E et al. [96] found that the knee pain in OA is associated
with BMLs and effusion/synovitis but not cartilage defects. Our data suggests the same as
well. Change in cartilage defects was associated with increase in knee pain over 8 years in the
univariable analysis but once we accounted for BMLs and effusion/synovitis there was no
significant association between the two. Subchondral BMLs and effusion/synovitis that result
from cartilage defects, or knee structural damage in general, explain most of the pain
resulting from OA.
Our study has several strengths and weaknesses. The biggest strength of our study is the
longest knee MRI data currently available with cartilage defects natural history data available
at three time points. Secondly we accounted for global knee structural pathologies in our
results. A weakness of our study was that, unlike cartilage defects, the data from T-2
weighted MR images and WOMAC pain scale data was available at only 2 time points.
Secondly this data is from matched sample of middle-aged adults and is not representative of
general population especially older adults with more established knee OA.
Conclusion
Data from this midlife cohort suggests that cartilage defects are on the OA causal pathway for
structure and symptoms especially in the lateral compartment. Unlike meniscal pathology,
cartilage defects have the capacity to heal over time thus may be amenable to intervention.
Page 132
Chapter 9 Correlation between MRI and X-Ray Knee Structures
Chapter Nine
Correlation Between Changes in Global Knee

Structures Assessed on MRI and Radiographic
Osteoarthritis Changes over Ten Years in a Mid-
Life Cohort
Page 133
9.1 Introduction
Knee osteoarthritis (OA) is one of the most common causes of disability in adults over the
age of 60 [17]. Despite the increasing prevalence of OA worldwide, few effective treatments
are available. A major impediment to the development of effective treatment options that
would halt or delay the progression of the disease is the lack of a sensitive and reproducible
outcome for clinical trials. Currently, clinical examination and measurement of joint space
narrowing (JSN) and osteophytes on plain radiographs are the gold standard for diagnosing
knee OA [79]. Increase in JSN (or change in joint space width (JSW)) is the only method
approved by regulators to monitor progression in disease-modifying OA drug (DMOAD) and
chondro-protective trials [80]. However, the reproducibility of this method is influenced by
subject positioning and changes in the radio-anatomic alignment of the joint in serial
examinations[31]. Furthermore, radiography is insensitive to early disease structural changes
as an estimated 10% reduction in cartilage loss occurs before radiographic OA is detected
[84]. It is also not very sensitive to change as it takes several years to detect progression of
radiographic OA [84, 85].
JSN is traditionally considered a surrogate for cartilage volume and changes in joint space are
attributed to cartilage loss. Extensive use of magnetic resonance imaging (MRI) over the last
decade has shown that OA is a disease of the whole joint [212]. Structural changes, apart
from cartilage volume loss, such as cartilage defects, meniscal tears and meniscal extrusion
are also associated with JSN in cross-sectional studies [116] but there have been no long term
studies examining change in these global knee measures and their independent contributions
to changes in JSN (or osteophytes).
The aim of this study was to describe the correlation between change in structural
abnormalities assessed on MRI and change in radiographs over 10 years in a midlife cohort.
Page 134
9.2 Patients and Methods
9.2.1 Study Population
based study [102]. The Offspring study began in southern Tasmania (primarily in the city of
Hobart) in June 2000 (Figure 9.1). Matched sampling was used to recruit the study
participants (mean age 45(26–61) years; 58% females). Half of the participants were the adult
offspring of patients who had a knee replacement performed for idiopathic knee OA at any
Hobart hospital from 1996 to 2000 [228]. The diagnosis was confirmed by reference to the
medical records of the orthopaedic surgeon and the original radiographs when possible. The
other half were age and sex matched controls, randomly selected from the population with no
history of knee OA in either parent. This study includes data from visit-1 (2000-01), visit-2
(2002-03) and visit-3 (2010-11) of the study.
obtained from all participants. Participants were excluded if they had a contraindication to
MRI (including metal sutures, presence of shrapnel, iron filing in eye, or claustrophobia).
Participants were also excluded if they had undergone a knee replacement surgery or did so
after the commencement of the study. Knee pain and knee injury were not a basis for
exclusion.
Page 135
9.2.3 Knee joint injury and surgery
History of knee joint injury and surgery were assessed at all three visits using a self-
administered questionnaire [196] which included the following questions :
• “Have you ever had a knee surgery?”
• “Please provide further details about the surgery”
Only right knee injuries were included in the analysis as MRI scans were on the right knee.
9.2.4 Imaging
MRI
International, USA) using a commercial transmit-receive extremity coil at the baseline visit, 2
year and 10 year follow up visits. The following image sequence was used: (1) a T1-weighted
fat-suppressed 3D gradient-recalled acquisition in the steady state, flip angle 55°, repetition
time 58 msec, echo time 12 msec, field of view 16 cm, 60 partitions, 512×512–pixel matrix,
slice thickness of 1.5 mm without an interslice-gap (at all three visits); and (2) a T2-weighted
fat saturation 2D fast spin echo, flip angle 90°, repetition time 3067 ms, echo time 112 ms,
field of view 16 cm, 15 partitions, 256×256 matrix, slice thickness of 4 mm with an interslice
gap of 0.5–1.0 mm (at visit 2 and 3 only).
Page 136
Cartilage volume
Knee cartilage volume was evaluated at baseline and 10 years by one trained observer on T1-
weighted gradient echo MR images. Knee cartilage volume was determined by means of
image processing on an independent workstation at baseline and follow up. The volumes of
individual cartilage plates (medial tibia and femora, and lateral tibia and femora) were
isolated from the total volume by manually drawing dis-articulation contours around the
cartilage boundaries on a section by section basis. These data were then resampled by means
of bilinear and cubic interpolation (area of 312 × 312 µm by 1.5 mm thickness, continuous
sections) for the final three-dimensional rendering to calculate the cartilage volume.
Tibial cartilage volume was assessed using Osiris (University of Geneva, Switzerland)
software as previously described [104, 109]. The coefficient of variation(CV) ranged from
2.1–2.2% for intra-observer repeatability [84]. Femoral cartilage volume was determined
using Cartiscope (ArthroLab, Montreal, Canada), as previously described [110-112]. The CV
was approximately 2% for intra-observer and inter-scan repeatability [111]. Total cartilage
volume was calculated as: tibial + femoral cartilage volume.
Independent readers assessed tibial and femoral cartilage volume at baseline and 10 years.
Change in cartilage volume was calculated as: follow-up total cartilage volume - baseline
Cartilage defects
Cartilage defects were assessed on T1-weighted gradient echo MR images by one trained
observer at the medial tibial, medial femoral, lateral tibial, and lateral femoral sites on a 0-4
scale, as previously described [108]: grade 0=normal cartilage; grade 1=focal blistering and
intra-cartilaginous low-signal intensity area with an intact surface and base; grade
2=irregularities on the surface or base and loss of thickness <50%; grade 3=deep ulceration
with loss of thickness >50%; and grade 4=full-thickness chondral wear with exposure of
subchondral bone. The presence of any cartilage defect was defined as any score ≥2.
Intraobserver reliability (expressed as intraclass correlation coefficient(ICC)) ranged from
Page 137
0.89-0.90 [108]. Interobserver reliability was assessed in 50 MR images and yielded an ICC
of 0.85-0.90 [108].
Meniscal tears
Meniscal tears were assessed by one trained observer on T1-weighted gradient echo and T2-
weighted (side by side) MR images at visit-2 and 3 of the study as previously described
[110]. The proportion of the menisci affected by a tear was scored separately (0-2 scale;
0=absence of a tear, 1=simple tears of different types: longitudinal, oblique, radial or
horizontal, 2=complex or macerated tears signifying loss>50% area of meniscal tissue) at the
anterior, middle, and posterior horns. The presence of any meniscal tear was defined as any
score ≥1. Anterior, middle and posterior scores were summed to create medial and lateral
meniscal tear scores. The intra- and inter-observer correlation coefficient ranged from 0.86 to
0.96 [111]. Meniscal tears were measured at visits 2 and 3 of the Offspring study, 2 and 10
years after the baseline visit.
Meniscal extrusion
space, not including the osteophytes, was evaluated by one trained observer at baseline and at
10 years for the anterior, body, and posterior horns of the menisci, as previously described
[111]. A score from 0 to 2 was used (0 = no extrusion, 1 = partial meniscal extrusion, and 2 =
complete meniscal extrusion with no contact with the joint space). The presence of any
meniscal extrusion was defined as any score ≥1. The scores of anterior, body and posterior
horns of medial or lateral menisci were summed to create a total meniscal extrusion score for
each of the medial and lateral tibiofemoral compartments which had a possible range from 0
to 6 [111].
Page 138
9.2.5 Radiology
baseline and 10 years. The angle was kept to 10–15˚ by a purpose built goniometer. The tube
to film and tube to tibial plateau angle was 90˚. Daily quality assurance was performed on the
equipment. Radiographs were scored individually for osteophytes and JSN, as described
previously [84]. Each of the following four features was scored on a scale from 0 to 3 (0 =
normal and 3 = severe): medial JSN, lateral JSN, medial osteophytes (femoral and tibial
combined) and lateral osteophytes (femoral and tibial combined). Each score was arrived at
by consensus with two readers simultaneously assessing the radiograph with immediate
reference to the Osteoarthritis Research Society International (OARSI) atlas [82]. A non-zero
score in either JSN or osteophytosis was regarded as evidence of radiographic OA.
Reproducibility was assessed in 50 radiographs, two weeks apart, and yielded an ICC of 0.99
for osteophytes and 0.98 for JSN [216].
Change in radiographic OA was calculated as: follow-up radiographic OA score - baseline

radiographic OA score. Any increase in radiographic score over 10 years was defined as an
increase (≥1).
experience in OA research or health professionals trained by musculoskeletal radiologists. All
the scans were read by independent readers. All MRI structures were assessed independently
as well with readers not aware of the severity of other structural abnormalities. Baseline and
follow-up scans were read as paired. Readers were not blinded to the chronological sequence
of the radiographs and MRI scans.
T-test and Chi square test were used to describe the characteristics of the study participants
who had no change or any change in radiographic OA over 10 years.
Spearman ranked correlation analyses were used to examine the correlations between
structural changes on MRI and radiographs. Absolute change in all MRI and radiographic
Page 139
structures was calculated as: Visit-3 score – Baseline/Visit-2 score. Only absolute changes,
regardless of magnitude and direction, were used to describe correlations between MRI and
radiographic changes. Multivariable analyses were adjusted for age, sex, BMI, offspring-
control status, change in cartilage volume, cartilage defects, meniscal tears and meniscal
extrusion, and change in radiographic JSN/osteophytes.
Further analysis, using logistic regression, was performed to examine the association between
no increase or any increase in the severity of MRI and radiographic structures. Dichotomised
variables were used to describe no or any increase in the severity of MRI and radiographic
structures (0-1 scale; 0=stable score or a decrease in score, 1=increase in score). Cartilage
volume loss was dichotomised using mean cartilage volume loss over 10 years.
Interaction terms were calculated for all the above-mentioned correlations to see if the
strength of correlation was significantly different between the offspring and controls. Only
medial tibiofemoral compartment results are presented as the lateral compartment had limited
change and the resultant analysis lacked sufficient power.
A p-value less than 0.05 (two-tailed) was considered statistically significant. All analyses
were performed on Intercooled Stata V.12.0 for windows (StataCorp LP).
Page 140
9.3 Results
Figure 9.1 is flow-chart describing the study population of the offspring study. A total of
219/372 participants (57% female, mean age 45) completed the study. 8 participants had
missing radiographic or MRI data and were not included in the final analysis. There was no
significant difference in baseline characteristics between those who completed the study and
those lost to follow-up (follow-up vs loss to follow-up: age (years) 45.3 vs 45.1, p=0.80; sex
(female%) 57 vs 59, p=0.74; BMI (kg/m2) 27.2 vs 26.8, p=0.49; radiographic OA (%) 18 vs
15, p=0.48.
Figure 9.1. Flow chart of the Offspring study
Page 141
32% (67/211) of the participants showed an increase in the medial radiographic OA severity
with a mean increase in score of 0.49. All 211 participants, regardless of an increase or not in
medial radiographic OA score, were included in the final correlation analysis. Mean absolute
cartilage volume loss in the medial tibiofemoral compartment was -1284 mm3 (19% lower
than baseline).
Table 9.1 describes the characteristics of the study participants. First characteristics of the
whole study population are presented and then the characteristics of the participants split by
no or any change in medial radiographic OA over 10 years. Participants with any change had
a significantly higher prevalence of radiographic OA and meniscal tears at baseline and a
significantly higher percentage of participants who had any change in meniscal tears. There
was no other significant difference in either demographic or structural characteristics between
the two groups.
Table 9.2 describes the correlation between structural changes on MRI and radiographs.
Change in meniscal tears showed the strongest correlation with change in both JSN and
osteophytes in the fully adjusted model. Change in meniscal extrusion and cartilage defects
also showed significant correlations with change in JSN but not with change in osteophytes.
Cartilage loss showed weak and non-significant correlations with either change in JSN or
osteophytes. There were no significant differences between the offspring and controls for any
of the above mentioned correlations.
Further analysis was performed to describe the association between any increase, regardless
of magnitude, in MRI and radiographic structures. Any increase in the severity of meniscal
tears (OR=+3.6 (95%CI +1.7, +7.6), p=0.001) and extrusion (OR=+2.3 (+1.1, +4.6),
p=0.019), but not cartilage defects or volume, was significantly associated with any increase
in JSN. Only increase in meniscal tears was associated with any increase in osteophytes
(OR=+2.8 (+1.3, +6.1), p=0.008).
Only two participants underwent knee surgery between baseline and visit-3 and on both
occasions the surgery was not a meniscectomy or a joint replacement. Further adjustment for
knee surgery and incident knee injury did not change the effect size considerably for any of
the associations described earlier (data not shown).
Page 142
Table 9.1. Characteristics of the study participants
Whole No change in Any change in

population radiographic radiographic P-Value
OA (n=144) OA (n=67)
Age (years) 45.2 45.4 44.9 0.635
Female sex (%)* 57 58 55 0.742
BMI (kg/m2) 27.1 27.4 26.7 0.338
Offspring (%)* 51 50 54 0.614

Any radiographic OA
at baseline (%)* 18 10 22 0.031
Any medial meniscal
tear at baseline (%)* 20 16 30 0.031
Any change in medial
meniscal tears (%)* 13 8 23 0.005
Any medial meniscal
extrusion at baseline (%)* 8 8 6 0.529
meniscal extrusion (%)* 11 10 13 0.498
Medial cartilage volume at
baseline (mm3)^ 6766 6733 6887 0.558
Medial cartilage volume
loss (mm3)^ -1283 -1276 -1287 0.899
Any medial cartilage
defects at baseline (%)^* 25 28 20 0.219
cartilage defects (%)^* 50 48 53 0.554
*Determined by Chi square test, all others by t-test, ^Sum of medial tibial and femoral
Page 143
Table 9.2. Correlation between structural changes on MRI over 8-10 years and radiographic
changes over 10 years
Unadjusted Adjusteda Adjustedb
ρ P-value ρ P-value ρ P-value
Change in medial meniscal tears over 8 years

Change in medial
JSN over 10 years +0.42 <0.01 +0.38 <0.01 +0.37 <0.01
Change in medial
osteophytes over 10 years +0.34 <0.01 +0.34 <0.01 +0.31 <0.01
Change in medial meniscal extrusion over 10 years

Change in medial
JSN over 10 years +0.30 <0.01 +0.27 <0.01 +0.22 <0.01
Change in medial
osteophytes over 10 years +0.14 0.03 +0.14 0.04 +0.02 0.83
Medial cartilage volume loss over 10 years

Change in medial
JSN over 10 years -0.18 0.01 -0.11 0.14 -0.01 0.97
Change in medial
osteophytes over 10 years -0.14 0.06 -0.17 0.02 -0.12 0.14
Change in medial cartilage defects over 10 years

Change in medial
JSN over 10 years +0.12 0.09 +0.11 0.12 +0.16 0.04
Change in medial
osteophytes over 10 years +0.02 0.74 +0.01 0.94 +0.02 0.79
a = adjusted for age, sex, BMI, offspring-control status

b = a + for changes in all MRI and radiographic factors in the table
Page 144
9.4 Discussion
This study documents the individual contributions of change in knee structures assessed on
MRI and change in both JSN and osteophytes. Surprisingly, given JSN is currently being
used as a surrogate measure for cartilage volume loss, change in meniscal tears showed a
moderate correlation with both change in JSN and osteophytes. Change in meniscal extrusion
and cartilage defects showed a slightly weaker correlation with JSN only, whereas direct
measurement of cartilage volume loss showed no significant independent association with
radiographic change (despite both worsening over the study timeframe).
Conventional radiographs are an unreliable method of evaluating articular cartilage loss in

patients with early OA, as initial JSN is secondary to meniscal pathologies rather than
thinning of articular cartilage in most cases [229]. Joint space is shared by both articular
cartilage and meniscus. Damage to the meniscus either in the form of degenerative
tears/extrusion [31] or mechanical removal [230, 231] will result in a decrease in the joint
space. Previous cross-sectional studies have shown that the meniscus accounts for more of
the variation in JSN than cartilage, however most of these changes are usually attributed to
meniscal extrusion rather than tears [229]. Longitudinal studies[31, 111], conducted over 24-
30 months, have shown that change in meniscal extrusion has a stronger association with
JSN/JSW compared to tears and cartilage defects. In contrast, in our study, change in
meniscal tears showed the strongest correlation with both JSN and osteophytes over 10 years
and the two meniscal pathologies explained most of the change in JSN attributable to MRI
structures, with cartilage defects showing only a weak but a significant correlation. There are
a few possible explanations for stronger association of tears compared to extrusion. A recent
study has shown that meniscal extrusion usually has an immediate effect on JSN, which does
not progress much over time [111]. Furthermore, worsening of a meniscal tear, especially
from a simple to a complex tear, will require a long follow-up period to observe these
changes on MRI. This is perhaps why earlier studies with shorter follow-up periods failed to
observe this association.
JSN is traditionally considered a surrogate marker for cartilage volume loss. Longitudinal
studies, performed over shorter follow-up periods, have shown either moderate [232] or weak
[111, 233] correlations between cartilage volume loss and change in JSN. We saw a similar
Page 145
correlation between cartilage volume loss and JSN in unadjusted analysis but after adjustment
for other knee structures there was virtually no correlation between the two over 10 years. A
recent study by Crema et al. [212] showed that cartilage defects are the most common
structural abnormality in people with worsening JSN. Our findings are in agreement with
Crema et al. [212] but despite being more common than meniscal pathologies, change in
cartilage defects only showed a modest but significant correlation with JSN. Our findings
question the use of JSN as an outcome in chondro-protective drug trials as JSN is a poor
reflection of cartilage volume loss in particular. Previous studies have shown that cartilage
volume loss is a more sensitive and specific marker of cartilage integrity than JSN and
possibly the best outcome measure for drug trials looking to preserve cartilage[111].
Moreover, cartilage defects have an unstable natural history as they can improve without any
intervention depending on the age of the patient especially in people with less advanced
disease [89, 141]. This is a possible explanation of weaker associations we observed in our
study compared to previously available literature.
In terms of registration, radiographs are still the gold standard for OA trials [79]. The
European and American regulatory agencies require evidence of not only an effect on JSN
but also on pain and function as an end-point in DMOAD trials [234]. However, our data
shows that change in JSN is mainly a reflection of change in meniscus, with cartilage defects
showing only a weak, yet independent, correlation. Moreover, recent work from this cohort
has shown that meniscal tear changes are strongly associated with cartilage volume loss, pain
and function [211]. These findings suggest that meniscal tears are on the OA causal pathway
and chondro-protective/DMOAD trials should account for the severity of meniscal tears to
see the true effect of a certain drug on cartilage volume loss or changes in JSN.
We believe our findings are generalizable to a middle aged population. Although our cohort
had a low prevalence of radiographic OA to begin with, we did not see any significant
differences between participants with or without radiographic OA for any of the above-
mentioned correlations. Similarly, we did not find any significant differences between the
offspring and the controls either. Furthermore, our group has previously shown that change in
meniscal extrusion has a stronger association with change in JSN compared to cartilage
volume loss in a randomly selected older population as well [235]. Nevertheless, this finding
merits replication in a sample purely with radiographic OA.
Page 146
Our study has strengths and limitations. This study has the longest follow-up period of any
OA study with global knee structural changes assessed on both MRI and radiographs.
Radiographs and MRI scans for individual structures were assessed by independent readers,
minimising the possibility of bias. A limitation of our study was a significant loss to follow-
up over 10 years. Loss to follow-up can be a potential source of bias, however re-analysis of
the data using inverse probability weighting did not change any of the results, indicating
robust results. Another limitation of our study was the absence of meniscal tear scoring at the
baseline visit, as we did not have the correct MRI sequence to score tears. However, the
natural history data of meniscal tears in this cohort showed that none of the tears improved
over 8 years, suggesting that meniscal tears present at visit-2 were likely to be present at
baseline. This suggests that the association between radiographic and meniscal tear changes
would have either remained the same over 8 (between visit-2 and 3) and 10 (visit-1 and 3)
years or would have increased slightly if we had meniscal tears measurement at baseline,
making the correlations we described even stronger [236, 237]. Lastly, tibial and femoral
cartilage volume were segmented using different methodology as was outlined in the
manuscript. Separate readers performed the measurements, which resulted in differences in
how the scans were processed. Although both methods are almost equally sensitive at picking
up any change in cartilage volume [158], this difference can still be a source of potential bias.
Conclusion
Change in JSN is correlated with change in meniscal tears and, to a lesser extent, with
meniscal extrusion and cartilage defects. In this sample change in JSN is a composite
measure that does not reflect change in cartilage volume. These data, suggest that the use of
JSN as an outcome measure in chondro-protective drug trials should be reviewed.
Page 147
Chapter 10 Summary and Future Direction
Chapter Ten
Summary and Future Direction
Page 148
10.1 Summary
Osteoarthritis (OA) is the most common joint disorder in the world and in Western
populations is one of the most frequent causes of pain, loss of function, and disability in
adults [238]. The knee joint is the most commonly affected weight bearing joint by OA.
Nearly one in two older adults are affected by knee OA by the age of 85 [33]. There is
currently no cost-effective disease modifying osteoarthritic drugs available in the market.
Most researchers have focused on the older adults with more established disease with little or
no potential of reversibility. Hence it has been difficult to identify tissues with potential of
reversibility that can be targeted in clinical trials. The aim of this thesis was to look at the
long term natural history of knee structural changes in middle-aged adults, identify structures
with potential for reversibility, and to examine the role of family history of disease in disease
progression and predictors of early disease structural changes.
Chapter 4 examined the cross-sectional association between history of knee injury and knee
structural damage assessed on MRI in middle-aged adults from the Offspring study and in a
random community based sample of older adults. In middle-aged adults, BML presence,
tibial bone area and meniscal extrusion presence were significantly higher in those with knee
injury, whereas in older adults, cartilage defect presence, cartilage volume, BML presence
and tibial bone area were significantly associated with knee injury. This was the first study to
look at the association between history of knee injury and knee joint structural changes
assessed on MRI and found that the association between knee injury and MRI-assessed
structural pathology in the knee joint is moderate and appears to be stronger in older adults
compared to middle-aged adults. However due to the cross-sectional design of the study,
causation cannot be drawn from this data. A longitudinal study is warranted to further
understand the role of knee injury and structural damage in different age groups.
Chapter 5 examined the role of family history of knee joint replacement due to OA and the
risk of radiographic OA and cartilage volume loss over 10 years. Family history of knee OA
increased the risk of radiographic OA (JSN and osteophytes) and medial tibial cartilage
volume loss over 10 years compared to community-acquired controls with no family history
of OA. Most of these changes were mediated by differences in baseline characteristics of
offspring and controls except for increases in medial JSN.
Page 149
Chapter 6 looked at the natural history of BMLs in middle-aged adults and found that the
natural history of knee BMLs was unstable. BMLs were common in middle-aged adults at
baseline. 24% of these BMLs at baseline increased in size, 55% remained stable and 21%
decreased in size or resolved completely over 8 years. Change in BMLs was predicted by
BMI and strenuous physical activity. An increase in BML size or a new BML resulted in an
increase in pain especially in males and those with a family history of OA.
Chapter 7 examined the natural history of meniscal tears. Only 22% of the participants had a
meniscal tear at baseline. Over 8 years, 16 % of the participants had an increase in severity of
meniscal tears while none improved. Change in meniscal tears shared common risk factors
with knee OA and was independently associated with worsening knee pain and structural
damage suggesting that meniscal tears are on the knee OA causal pathway and not just a
result of mechanical factors.
Chapter 8 looked at the natural history of cartilage defects. 44% of the participants had at
least one cartilage defect at any site at baseline. Most of these defects remained stable,
whereas 26% increased and 13% decreased in severity over 10 years. Cartilage defects
independently predicted cartilage volume loss in the lateral compartment only. Change in
cartilage defects on the other hand was associated with changes in BMLs and cartilage
volume loss mostly in the lateral compartment, suggesting a more crucial role of cartilage
defects in the development of lateral compartment knee OA. There was no independent
association between change in cartilage defects and increase in pain after adjustment for
BMLs, meniscal tears and effusion.
Chapter 9 examined the correlation between changes in structural abnormalities assessed on

MRI and change in radiographic OA over 10 years. Change in JSN was correlated with
change in meniscal tears and, to a lesser extent, with meniscal extrusion and cartilage defects.
In this sample, change in JSN was a composite measure that did not reflect cartilage volume
loss prompting the review of the use of JSN as an outcome measure in chondro-protective
drug trials.
In conclusion, this series of related studies detail the natural history of knee structural
progression in middle-aged adults. Structural changes such as BMLs and cartilage defects
have the potential of reversibility in early disease and should be targeted in disease modifying
Page 150
clinical trials. Meniscal tears and BMLs should be targeted in symptom modifying clinical
trials especially in those with a family history of OA. Lastly findings from this thesis suggest
that the use of JSN as an outcome measure in chondro-protective trial should be reviewed.
10.2 Future Direction
This thesis has presented several novel findings from a unique 10-year follow-up study of
middle-aged adults. This thesis, and the Offspring study as a whole, has provided great
insights into early knee OA structural changes. The most important aspect of this thesis is the
detailed long-term knee structural natural history data, predictors of these changes and the
structural changes/symptoms associated with these changes.
Several studies have shown the association between history of knee injury and radiographic
OA [55, 123]. Chapter 4 was the first study that examined the association between history of
knee injury and global knee structural pathologies. This study provided great insight as to
how a knee injury early in life can predispose to structural damage in middle-aged and older
adults. It suggests that history of knee injury can help identify people who are more
predisposed to certain structural changes in different age groups. It may be possible to use
knee injury history to identify fast progressors who can be targeted in clinical trials in order
to slow disease progression and even prevent disease onset. Recently our group has finished
data collection of a small longitudinal study in young Tasmanian Aussie Rules Footballers
[239]. This study will examine the association between history of knee injury and structural
and symptomatic changes over a full season and also compare it to the contralateral knee.
Australian rules Footballers experience a high rate of injury and given the link between injury
and future risk of OA, it is important to gain a better understanding of knee structural changes
in athletes at high risk of injury. Data from this thesis highlighted that a previous history of
knee injury was associated with MRI abnormalities in both middle and older adults. Given
the clinical importance of knee injury, athletes who have had an injury or present with early
structural changes on MRI may be a lucrative population to test early OA interventions.
There may be scope to introduce therapeutic interventions into rehabilitation programs
following joint injury that may have the potential to prevent knee OA in the future. Our group
Page 151
is already in talks with Australian Football League (AFL) to carry out a larger study in
professional footballers.
Chapter 5 and the subsequent chapters looking at the natural history of knee structures
showed the crucial role family history of OA plays in the progression of disease. Having a
family history of knee OA increases the risk of both JSN and cartilage loss over time. This
highlights that this sub-group represents a potential group of patients who may particularly
benefit from early OA interventions. It may be possible to target them for clinical trials
targeting both structure and symptoms. Additionally, the data from this thesis has shown that
the offspring not only had a higher prevalence of pain but also showed greater worsening
over time. Data from chapters 6 and 7 suggests that having a family history of knee OA can
also affect the pain perception pathway. Offspring participants had a higher increase in pain
on the WOMAC scale per unit increase in severity of structural abnormalities including
BMLs and meniscal tears. As the WOMAC pain scale is a subjective measure of pain, future
studies could potentially use more objective measures of pain such as functional MRI scans
to gain a better understanding of the influence of genes on pain perception. The relationship
between reported pain intensity and the peripheral stimulus that evokes it depends on many
factors such as the level of arousal, anxiety, depression, attention and expectation or
anticipation. Pharmacological therapies such as anti-depressants and gabapentin can possibly
target these pathways in the future for symptomatic relief.
Chapter 6 and 8 showed that BMLs and cartilage defects respectively have an unstable
natural history and can regress in severity or completely heal. This suggests that these
structures can be potentially targeted in disease modifying clinical trials. BMLs are especially
an attractive target. Firstly, unlike cartilage defects, BMLs do not lose the ability to regress in
older adults. Secondly changes BMLs are independently associated with changes in
symptoms. Our group recently conducted a randomised controlled trial (RCT) examining the
effectiveness of zoledronic acid (ZA) on knee pain and knee BMLs [189]. This study was a
single centre double blind placebo controlled randomised trial of intravenous (IV) ZA (5 mg)
vs placebo in 59 adults aged 50–80 years with knee pain (>40 mm on a visual analog score
(VAS)) and a knee BML. We found that a single infusion of IV ZA was effective in reducing
pain intensity and BML size compared to placebo after six months. A larger multi-centre trial
is underway to further investigate the potential of this treatment. As discussed above,
Page 152
treatments such as this one may be effective in younger populations who have had a knee
injury, such as athletes, and are at an increased risk of developing future OA.
JSN is traditionally considered a surrogate marker for cartilage volume loss. Increase in JSN
is the only method approved by regulators to register participants, monitor progression in
disease-modifying OA drug (DMOAD) trials and is generally used as an end-point in
chondro-protective trials. Data from chapter 9 suggests that JSN is a composite measure that
reflects changes in meniscus mainly and to a lesser extent articular cartilage. Therefore it is
not surprising that chondro-protective trials have thus far shown disappointing results using
JSN as an outcome. Our findings form part of a large evidence base that is currently being
used to demonstrate to the FDA that JSN is no longer an appropriate outcome in trials
wanting to show a chondro-protective effect. Applications have been submitted to the FDA to
have other structural outcomes measured on MRI (e.g. BMLs) be accepted as both an
indication for therapy and an outcome measure in future trials. The success of such
applications will hopefully be publically available soon.
In conclusion, data from this thesis detailed the long-term natural history of global knee
structural changes in middle-aged adults. Data from this thesis identified the role of knee
injury and structural knee damage in middle-aged and older adults. This data also emphasized
the role of family history of knee OA in structural and symptomatic progression of the
disease. BMLs and cartilage defects were identified as structures that have the potential to
regress and can be potentially targeted in clinical trials. Lastly data from this thesis has
questioned the use of JSN as an end-point in chondro-protective trials.
Page 153
References
1. Goldring, S.R. and M.B. Goldring, Clinical aspects, pathology and pathophysiology of
osteoarthritis. J Musculoskelet Neuronal Interact, 2006. 6(4): p. 376-8.
2. Litwic, A., et al., Epidemiology and burden of osteoarthritis. Br Med Bull, 2013. 105: p. 185-99.
3. Felson, D.T., et al., The incidence and natural history of knee osteoarthritis in the elderly. The
Framingham Osteoarthritis Study. Arthritis Rheum, 1995. 38(10): p. 1500-5.
4. Felson, D.T., Clinical practice. Osteoarthritis of the knee. N Engl J Med, 2006. 354(8): p. 841-8.
5. Buckland-Wright, C., Subchondral bone changes in hand and knee osteoarthritis detected by
radiography. Osteoarthritis Cartilage, 2004. 12 Suppl A: p. S10-9.
6. Burr, D.B., Anatomy and physiology of the mineralized tissues: role in the pathogenesis of
osteoarthrosis. Osteoarthritis Cartilage, 2004. 12 Suppl A: p. S20-30.
7. Hill, C.L., et al., Knee effusions, popliteal cysts, and synovial thickening: association with knee
pain in osteoarthritis. J Rheumatol, 2001. 28(6): p. 1330-7.
8. Hunter, D.J., Imaging insights on the epidemiology and pathophysiology of osteoarthritis. Rheum
Dis Clin North Am, 2009. 35(3): p. 447-63.
9. Burr, D.B. and M.B. Schaffler, The involvement of subchondral mineralized tissues in
osteoarthrosis: quantitative microscopic evidence. Microsc Res Tech, 1997. 37(4): p. 343-57.
10. Pelletier, J.P., J. Martel-Pelletier, and S.B. Abramson, Osteoarthritis, an inflammatory disease:
potential implication for the selection of new therapeutic targets. Arthritis Rheum, 2001. 44(6): p.
1237-47.
11. Mandelbaum, B. and D. Waddell, Etiology and pathophysiology of osteoarthritis. Orthopedics,

2005. 28(2 Suppl): p. s207-14.
12. Murphy, L., et al., Lifetime risk of symptomatic knee osteoarthritis. Arthritis Rheum, 2008. 59(9):
p. 1207-13.
13. Global, regional, and national incidence, prevalence, and years lived with disability for 301 acute
and chronic diseases and injuries in 188 countries, 1990-2013: a systematic analysis for the Global
Burden of Disease Study 2013. Lancet, 2015. 386(9995): p. 743-800.
14. Health-care expenditure on arthritis and other musculoskeletal conditions 2008–09. Canberra:
AIHW, 2014(Arthritis series no. 20. Cat. no. PHE 177).
15. Stan, G., H. Orban, and C. Orban, Cost Effectiveness Analysis of Knee Osteoarthritis Treatment.
Chirurgia (Bucur), 2015. 110(4): p. 368-74.
16. Pinto, D., et al., Cost-effectiveness of nonpharmacologic, nonsurgical interventions for hip and/or
knee osteoarthritis: systematic review. Value Health, 2012. 15(1): p. 1-12.
17. Cross, M., et al., The global burden of hip and knee osteoarthritis: estimates from the global
burden of disease 2010 study. Ann Rheum Dis, 2014. 73(7): p. 1323-30.
Page 154
18. Time to Move: Osteoarthritis. Arthritis Australia, 2014 March.
19. Blackburn, T.A. and E. Craig, Knee anatomy: a brief review. Phys Ther, 1980. 60(12): p. 1556-60.
20. Wagner, M., [Functional anatomy of the knee joint]. Orthopade, 1987. 16(2): p. 88-99.
21. Flandry, F. and G. Hommel, Normal anatomy and biomechanics of the knee. Sports Med
Arthrosc, 2011. 19(2): p. 82-92.
22. Last, R.J., Some anatomical details of the knee joint. J Bone Joint Surg Br, 1948. 30b(4): p. 683-8.
23. Kurosawa, H., T. Fukubayashi, and H. Nakajima, Load-bearing mode of the knee joint: physical
behavior of the knee joint with or without menisci. Clin Orthop Relat Res, 1980(149): p. 283-90.
24. Englund, M., E.M. Roos, and L.S. Lohmander, Impact of type of meniscal tear on radiographic
and symptomatic knee osteoarthritis: a sixteen-year followup of meniscectomy with matched controls.
Arthritis Rheum, 2003. 48(8): p. 2178-87.
25. Fenn, S., A. Datir, and A. Saifuddin, Synovial recesses of the knee: MR imaging review of
anatomical and pathological features. Skeletal Radiol, 2009. 38(4): p. 317-28.
26. Makris, E.A., P. Hadidi, and K.A. Athanasiou, The knee meniscus: structure-function,
pathophysiology, current repair techniques, and prospects for regeneration. Biomaterials, 2011.
32(30): p. 7411-31.
27. Lane, L.B. and P.G. Bullough, Age-related changes in the thickness of the calcified zone and the
number of tidemarks in adult human articular cartilage. J Bone Joint Surg Br, 1980. 62(3): p. 372-5.
28. Shapiro, L.M., et al., Mechanisms of osteoarthritis in the knee: MR imaging appearance. J Magn
Reson Imaging, 2014. 39(6): p. 1346-56.
29. Martel-Pelletier, J., Pathophysiology of osteoarthritis. Osteoarthritis Cartilage, 2004. 12 Suppl A:

p. S31-3.
30. Martel-Pelletier, J., et al., New thoughts on the pathophysiology of osteoarthritis: one more step
toward new therapeutic targets. Curr Rheumatol Rep, 2006. 8(1): p. 30-6.
31. Hunter, D.J., et al., Change in joint space width: hyaline articular cartilage loss or alteration in
meniscus? Arthritis Rheum, 2006. 54(8): p. 2488-95.
32. Henderson, J.V., et al., Prevalence, causes, severity, impact, and management of chronic pain in
Australian general practice patients. Pain Med, 2013. 14(9): p. 1346-61.
33. AIHW analysis of unpublished ABS Australian Health Survey, 2011–12 (National Health Survey
Component). 2011-12.
34. Nguyen, U.S., et al., Increasing prevalence of knee pain and symptomatic knee osteoarthritis:
survey and cohort data. Ann Intern Med, 2011. 155(11): p. 725-32.
35. Jinks, C., K. Jordan, and P. Croft, Measuring the population impact of knee pain and disability
with the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC). Pain, 2002.
100(1-2): p. 55-64.
Page 155
36. Peat, G., R. McCarney, and P. Croft, Knee pain and osteoarthritis in older adults: a review of
community burden and current use of primary health care. Ann Rheum Dis, 2001. 60(2): p. 91-7.
37. Perrot, S., Osteoarthritis pain. Best Pract Res Clin Rheumatol, 2015. 29(1): p. 90-7.
38. O’Brien, T. and H. Breivik, The impact of chronic pain—European patients’ perspective over 12
months. Scandinavian Journal of Pain. 3(1): p. 23-29.
39. Parks, E.L., et al., Brain activity for chronic knee osteoarthritis: dissociating evoked pain from
spontaneous pain. Eur J Pain, 2011. 15(8): p. 843.e1-14.
40. Hannan, M.T., D.T. Felson, and T. Pincus, Analysis of the discordance between radiographic
changes and knee pain in osteoarthritis of the knee. J Rheumatol, 2000. 27(6): p. 1513-7.
41. Hunter, D.J., Osteoarthritis. Best Pract Res Clin Rheumatol, 2011. 25(6): p. 801-14.
42. Goldenberg, D.L., The interface of pain and mood disturbances in the rheumatic diseases. Semin
43. Bratus, A., et al., Candidate gene approach in genetic epidemiological studies of osteoarthritis-
related pain. Pain, 2014. 155(2): p. 217-21.
44. Zhang, Y. and J.M. Jordan, Epidemiology of osteoarthritis. Clin Geriatr Med, 2010. 26(3): p. 355-
69.
45. Silverwood, V., et al., Current evidence on risk factors for knee osteoarthritis in older adults: a
systematic review and meta-analysis. Osteoarthritis Cartilage, 2015. 23(4): p. 507-15.
46. Jarvholm, B., et al., Age, bodyweight, smoking habits and the risk of severe osteoarthritis in the
hip and knee in men. Eur J Epidemiol, 2005. 20(6): p. 537-42.
47. Oliveria, S.A., et al., Incidence of symptomatic hand, hip, and knee osteoarthritis among patients
in a health maintenance organization. Arthritis Rheum, 1995. 38(8): p. 1134-41.
48. Heidari, B., Knee osteoarthritis prevalence, risk factors, pathogenesis and features: Part I.
Caspian J Intern Med, 2011. 2(2): p. 205-12.
49. Blagojevic, M., et al., Risk factors for onset of osteoarthritis of the knee in older adults: a
systematic review and meta-analysis. Osteoarthritis Cartilage, 2010. 18(1): p. 24-33.
50. Yoshimura, N., et al., Association of knee osteoarthritis with the accumulation of metabolic risk
factors such as overweight, hypertension, dyslipidemia, and impaired glucose tolerance in Japanese
men and women: the ROAD study. J Rheumatol, 2011. 38(5): p. 921-30.
51. Grazio, S. and D. Balen, [Obesity: risk factor and predictor of osteoarthritis]. Lijec Vjesn, 2009.
131(1-2): p. 22-6.
52. Felson, D.T., et al., Weight loss reduces the risk for symptomatic knee osteoarthritis in women.
The Framingham Study. Ann Intern Med, 1992. 116(7): p. 535-9.
53. Srikanth, V.K., et al., A meta-analysis of sex differences prevalence, incidence and severity of
osteoarthritis. Osteoarthritis Cartilage, 2005. 13(9): p. 769-81.
Page 156
54. Muraki, S., et al., Prevalence of radiographic knee osteoarthritis and its association with knee
pain in the elderly of Japanese population-based cohorts: the ROAD study. Osteoarthritis Cartilage,
2009. 17(9): p. 1137-43.
55. Wilder, F.V., et al., History of acute knee injury and osteoarthritis of the knee: a prospective
epidemiological assessment. The Clearwater Osteoarthritis Study. Osteoarthritis Cartilage, 2002.
10(8): p. 611-6.
56. Rogers, J., L. Shepstone, and P. Dieppe, Is osteoarthritis a systemic disorder of bone? Arthritis
Rheum, 2004. 50(2): p. 452-7.
57. Kujala, U.M., et al., Knee osteoarthritis in former runners, soccer players, weight lifters, and
shooters. Arthritis Rheum, 1995. 38(4): p. 539-46.
58. Ingham, S.L., et al., Incident knee pain in the Nottingham community: a 12-year retrospective
cohort study. Osteoarthritis Cartilage, 2011. 19(7): p. 847-52.
59. Zhang, W., et al., Nottingham knee osteoarthritis risk prediction models. Ann Rheum Dis, 2011.
70(9): p. 1599-604.
60. D'Souza, J.C., et al., Analysis of the Third National Health and Nutrition Examination Survey
(NHANES III) using expert ratings of job categories. Am J Ind Med, 2008. 51(1): p. 37-46.
61. Vingard, E., et al., Occupation and osteoarthrosis of the hip and knee: a register-based cohort
study. Int J Epidemiol, 1991. 20(4): p. 1025-31.
62. Felson, D.T., et al., Risk factors for incident radiographic knee osteoarthritis in the elderly: the
Framingham Study. Arthritis Rheum, 1997. 40(4): p. 728-33.
63. Cooper, C., et al., Risk factors for the incidence and progression of radiographic knee
osteoarthritis. Arthritis Rheum, 2000. 43(5): p. 995-1000.
64. Verweij, L.M., et al., Physical activity and incident clinical knee osteoarthritis in older adults.
65. Dore, D.A., et al., The association between objectively measured physical activity and knee
structural change using MRI. Ann Rheum Dis, 2013. 72(7): p. 1170-5.
66. Felson, D.T., et al., Physical activity, alignment and knee osteoarthritis: data from MOST and the
OAI. Osteoarthritis Cartilage, 2013. 21(6): p. 789-95.
67. Oiestad, B.E., et al., No Association between Daily Walking and Knee Structural Changes in
People at Risk of or with Mild Knee Osteoarthritis. Prospective Data from the Multicenter
Osteoarthritis Study. J Rheumatol, 2015. 42(9): p. 1685-93.
68. Spector, T.D., et al., Genetic influences on osteoarthritis in women: a twin study. BMJ, 1996.
312(7036): p. 940-3.
69. Bijkerk, C., et al., Heritabilities of radiologic osteoarthritis in peripheral joints and of disc
degeneration of the spine. Arthritis Rheum, 1999. 42(8): p. 1729-35.
70. Zhai, G., et al., The genetic contribution to muscle strength, knee pain, cartilage volume, bone
size, and radiographic osteoarthritis: a sibpair study. Arthritis Rheum, 2004. 50(3): p. 805-10.
Page 157
71. Chitnavis, J., et al., Genetic influences in end-stage osteoarthritis. Sibling risks of hip and knee
replacement for idiopathic osteoarthritis. J Bone Joint Surg Br, 1997. 79(4): p. 660-4.
72. Hirsch, R., et al., Familial aggregation of osteoarthritis: data from the Baltimore Longitudinal
Study on Aging. Arthritis Rheum, 1998. 41(7): p. 1227-32.
73. Rodriguez-Fontenla, C., et al., Assessment of osteoarthritis candidate genes in a meta-analysis of

nine genome-wide association studies. Arthritis Rheumatol, 2014. 66(4): p. 940-9.
74. Altman, R., 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): p. 1039-49.
75. Englund, M., et al., Effect of meniscal damage on the development of frequent knee pain, aching,
or stiffness. Arthritis Rheum, 2007. 56(12): p. 4048-54.
76. Felson, D.T., et al., Correlation of the development of knee pain with enlarging bone marrow
lesions on magnetic resonance imaging. Arthritis Rheum, 2007. 56(9): p. 2986-92.
77. Roos, E.M., et al., Knee Injury and Osteoarthritis Outcome Score (KOOS)--development of a self-
administered outcome measure. J Orthop Sports Phys Ther, 1998. 28(2): p. 88-96.
78. Bellamy, N., et al., Validation study of WOMAC: a health status instrument for measuring
clinically important patient relevant outcomes to antirheumatic drug therapy in patients with
osteoarthritis of the hip or knee. J Rheumatol, 1988. 15(12): p. 1833-40.
79. Guermazi, A., et al., Osteoarthritis: a review of strengths and weaknesses of different imaging
options. Rheum Dis Clin North Am, 2013. 39(3): p. 567-91.
80. Conaghan, P.G., et al., Summary and recommendations of the OARSI FDA osteoarthritis
Assessment of Structural Change Working Group. Osteoarthritis Cartilage, 2011. 19(5): p. 606-10.
81. Kellgren, J.H. and J.S. Lawrence, Radiological assessment of osteo-arthrosis. Ann Rheum Dis,
1957. 16(4): p. 494-502.
82. Altman, R.D., et al., Atlas of individual radiographic features in osteoarthritis. Osteoarthritis
Cartilage, 1995. 3 Suppl A: p. 3-70.
83. Altman, R.D. and G.E. Gold, Atlas of individual radiographic features in osteoarthritis, revised.
Osteoarthritis Cartilage, 2007. 15 Suppl A: p. A1-56.
84. Jones, G., et al., Early radiographic osteoarthritis is associated with substantial changes in
cartilage volume and tibial bone surface area in both males and females. Osteoarthritis Cartilage,
2004. 12(2): p. 169-74.
85. Wildi, L.M., et al., Assessment of cartilage changes over time in knee osteoarthritis disease-
modifying osteoarthritis drug trials using semiquantitative and quantitative methods: pros and cons.
Arthritis Care Res (Hoboken), 2013. 65(5): p. 686-94.
86. Link, T.M., et al., Osteoarthritis: MR imaging findings in different stages of disease and
correlation with clinical findings. Radiology, 2003. 226(2): p. 373-81.
Page 158
87. Bedson, J. and P.R. Croft, The discordance between clinical and radiographic knee osteoarthritis:
a systematic search and summary of the literature. BMC Musculoskelet Disord, 2008. 9: p. 116.
88. Creamer, P., Osteoarthritis pain and its treatment. Curr Opin Rheumatol, 2000. 12(5): p. 450-5.
89. Ding, C., et al., Natural history of knee cartilage defects and factors affecting change. Arch Intern
Med, 2006. 166(6): p. 651-8.
90. Baum, T., et al., Association of magnetic resonance imaging-based knee cartilage T2
measurements and focal knee lesions with knee pain: data from the Osteoarthritis Initiative. Arthritis
Care Res (Hoboken), 2012. 64(2): p. 248-55.
91. Felson, D.T., et al., The association of bone marrow lesions with pain in knee osteoarthritis. Ann
Intern Med, 2001. 134(7): p. 541-9.
92. Hunter, D.J., et al., 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). Ann
Rheum Dis, 2008. 67(2): p. 206-11.
93. Dore, D., et al., Natural history and clinical significance of MRI-detected bone marrow lesions at
the knee: a prospective study in community dwelling older adults. Arthritis Res Ther, 2010. 12(6): p.
R223.
94. Ding, C., et al., Meniscal tear as an osteoarthritis risk factor in a largely non-osteoarthritic
cohort: a cross-sectional study. J Rheumatol, 2007. 34(4): p. 776-84.
95. Hill, C.L., et al., Synovitis detected on magnetic resonance imaging and its relation to pain and
cartilage loss in knee osteoarthritis. Ann Rheum Dis, 2007. 66(12): p. 1599-603.
96. Yusuf, E., et al., Do knee abnormalities visualised on MRI explain knee pain in knee
osteoarthritis? A systematic review. Ann Rheum Dis, 2011. 70(1): p. 60-7.
97. Peterfy, C.G., et al., Whole-Organ Magnetic Resonance Imaging Score (WORMS) of the knee in
98. Kornaat, P.R., et al., MRI assessment of knee osteoarthritis: Knee Osteoarthritis Scoring System
(KOSS)--inter-observer and intra-observer reproducibility of a compartment-based scoring system.
Skeletal Radiol, 2005. 34(2): p. 95-102.
99. Eckstein, F., et al., In vivo reproducibility of three-dimensional cartilage volume and thickness
measurements with MR imaging. AJR Am J Roentgenol, 1998. 170(3): p. 593-7.
100. Eckstein, F., et al., In vivo morphometry and functional analysis of human articular cartilage
with quantitative magnetic resonance imaging--from image to data, from data to theory. Anat
Embryol (Berl), 2001. 203(3): p. 147-73.
101. Hunter, D.J., et al., Definition of osteoarthritis on MRI: results of a Delphi exercise.
Osteoarthritis Cartilage, 2011. 19(8): p. 963-9.
102. Jones, G., et al., Genetic mechanisms of knee osteoarthritis: a population based case-control
study. Ann Rheum Dis, 2004. 63(10): p. 1255-9.
Page 159
103. Ding, C., et al., Sex differences in knee cartilage volume in adults: role of body and bone size,
age and physical activity. Rheumatology (Oxford), 2003. 42(11): p. 1317-23.
104. Jones, G., et al., Sex and site differences in cartilage development: a possible explanation for
variations in knee osteoarthritis in later life. Arthritis Rheum, 2000. 43(11): p. 2543-9.
105. Zhai, G., et al., Correlates of knee pain in younger subjects. Clin Rheumatol, 2007. 26(1): p. 75-
80.
106. Zhai, G., et al., Familial, structural, and environmental correlates of MRI-defined bone marrow
lesions: a sibpair study. Arthritis Res Ther, 2006. 8(4): p. R137.
107. Ding, C., et al., Association between age and knee structural change: a cross sectional MRI
based study. Ann Rheum Dis, 2005. 64(4): p. 549-55.
108. Ding, C., et al., Association of prevalent and incident knee cartilage defects with loss of tibial
and patellar cartilage: a longitudinal study. Arthritis Rheum, 2005. 52(12): p. 3918-27.
109. Ding, C., et al., Smoking interacts with family history with regard to change in knee cartilage
volume and cartilage defect development. Arthritis Rheum, 2007. 56(5): p. 1521-8.
110. Berthiaume, M.J., et al., Meniscal tear and extrusion are strongly associated with progression of
symptomatic knee osteoarthritis as assessed by quantitative magnetic resonance imaging. Ann Rheum
Dis. , 2005. 64(4): p. 556-563.
111. Raynauld, J.P., et al., Long term evaluation of disease progression through the quantitative
magnetic resonance imaging of symptomatic knee osteoarthritis patients: correlation with clinical
symptoms and radiographic changes. Arthritis Res Ther, 2006. 8(1): p. R21.
112. Raynauld, J.P., et al., Reliability of a quantification imaging system using magnetic resonance
images to measure cartilage thickness and volume in human normal and osteoarthritic knees.
113. Ding, C., et al., Knee cartilage defects: association with early radiographic osteoarthritis,
decreased cartilage volume, increased joint surface area and type II collagen breakdown.
114. Oda, H., et al., Bone bruise in magnetic resonance imaging strongly correlates with the
production of joint effusion and with knee osteoarthritis. J Orthop Sci, 2008. 13(1): p. 7-15.
115. Issa, S.N. and L. Sharma, Epidemiology of osteoarthritis: an update. Curr Rheumatol Rep, 2006.
8(1): p. 7-15.
116. Peterfy, C.G., Imaging of the disease process. Curr Opin Rheumatol, 2002. 14(5): p. 590-6.
117. Ratzlaff, C.R., et al., Influence of lifetime hip joint force on the risk of self-reported hip
osteoarthritis: a community-based cohort study. Osteoarthritis Cartilage, 2011. 19(4): p. 389-98.
118. Jones, G., H.M. Cooley, and J.M. Stankovich, A cross sectional study of the association between
sex, smoking, and other lifestyle factors and osteoarthritis of the hand. J Rheumatol, 2002. 29(8): p.
1719-24.
Page 160
119. Kellgren, J.H. and J.S. Lawrence, Osteo-arthrosis and disk degeneration in an urban population.
Ann Rheum Dis, 1958. 17(4): p. 388-97.
120. Davis, M.A., et al., The association of knee injury and obesity with unilateral and bilateral
osteoarthritis of the knee. Am J Epidemiol, 1989. 130(2): p. 278-88.
121. Lau, E.C., et al., Factors associated with osteoarthritis of the hip and knee in Hong Kong
Chinese: obesity, joint injury, and occupational activities. Am J Epidemiol, 2000. 152(9): p. 855-62.
122. Yoshimura, N., et al., Risk factors for knee osteoarthritis in Japanese women: heavy weight,
previous joint injuries, and occupational activities. J Rheumatol, 2004. 31(1): p. 157-62.
123. Toivanen, A.T., et al., Obesity, physically demanding work and traumatic knee injury are major
risk factors for knee osteoarthritis--a population-based study with a follow-up of 22 years.
Rheumatology (Oxford), 2010. 49(2): p. 308-14.
124. Hunter, D.J., Advanced imaging in osteoarthritis. Bull NYU Hosp Jt Dis, 2008. 66(3): p. 251-60.
125. Ding, C., F. Cicuttini, and G. Jones, Tibial subchondral bone size and knee cartilage defects:
relevance to knee osteoarthritis. Osteoarthritis Cartilage, 2007. 15(5): p. 479-86.
126. Cicuttini, F.M., et al., Rate of cartilage loss at two years predicts subsequent total knee
arthroplasty: a prospective study. Ann Rheum Dis, 2004. 63(9): p. 1124-7.
127. Roos, E.M., Joint injury causes knee osteoarthritis in young adults. Curr Opin Rheumatol, 2005.
17(2): p. 195-200.
128. Muthuri, S.G., et al., History of knee injuries and knee osteoarthritis: a meta-analysis of
observational studies. Osteoarthritis Cartilage, 2011. 19(11): p. 1286-93.
129. Dore, D., et al., Correlates of subchondral BMD: a cross-sectional study. J Bone Miner Res,
2009. 24(12): p. 2007-15.
130. Raynauld, J.P., et al., Quantitative magnetic resonance imaging evaluation of knee osteoarthritis
progression over two years and correlation with clinical symptoms and radiologic changes. Arthritis
Rheum, 2004. 50(2): p. 476-87.
131. Kauffmann, C., et al., Computer-aided method for quantification of cartilage thickness and
volume changes using MRI: validation study using a synthetic model. IEEE Trans Biomed Eng, 2003.
50(8): p. 978-88.
132. Cicuttini, F.M., A.E. Wluka, and S.L. Stuckey, Tibial and femoral cartilage changes in knee
osteoarthritis. Ann Rheum Dis, 2001. 60(10): p. 977-80.
133. Palmer, W.E., S.M. Levine, and D.E. Dupuy, Knee and shoulder fractures: association of
fracture detection and marrow edema on MR images with mechanism of injury. Radiology, 1997.
204(2): p. 395-401.
134. Frobell, R.B., et al., The acutely ACL injured knee assessed by MRI: are large volume traumatic
bone marrow lesions a sign of severe compression injury? Osteoarthritis Cartilage, 2008. 16(7): p.
829-36.
Page 161
135. Frobell, R.B., Change in cartilage thickness, posttraumatic bone marrow lesions, and joint fluid
volumes after acute ACL disruption: a two-year prospective MRI study of sixty-one subjects. J Bone
Joint Surg Am, 2011. 93(12): p. 1096-103.
136. Gensburger, D., et al., Radiologic assessment of age-related knee joint space changes in women:
a 4-year longitudinal study. Arthritis Rheum, 2009. 61(3): p. 336-43.
137. Wluka, A.E., et al., Tibial plateau size is related to grade of joint space narrowing and
osteophytes in healthy women and in women with osteoarthritis. Ann Rheum Dis, 2005. 64(7): p.
1033-7.
138. Wang, Y., et al., Patellofemoral and tibiofemoral articular cartilage and subchondral bone
health following arthroscopic partial medial meniscectomy. Knee Surg Sports Traumatol Arthrosc,
2012. 20(5): p. 970-8.
139. Elsaid, K.A., et al., Association of articular cartilage degradation and loss of boundary-
lubricating ability of synovial fluid following injury and inflammatory arthritis. Arthritis Rheum,
2005. 52(6): p. 1746-55.
140. Crema, M.D., et al., Relevant traumatic injury of the knee joint-MRI follow-up after 7-10 years.
Eur J Radiol, 2009. 72(3): p. 473-9.
141. Carnes, J., et al., Knee cartilage defects in a sample of older adults: natural history, clinical
significance and factors influencing change over 2.9 years. Osteoarthritis Cartilage, 2012. 20(12): p.
1541-7.
142. Englund, M., et al., Risk factors for medial meniscal pathology on knee MRI in older US adults:
a multicentre prospective cohort study. Ann Rheum Dis, 2011. 70(10): p. 1733-9.
143. Koster, I.M., et al., Predictive factors for new onset or progression of knee osteoarthritis one
year after trauma: MRI follow-up in general practice. Eur Radiol, 2011. 21(7): p. 1509-16.
144. Potter, H.G., et al., Cartilage injury after acute, isolated anterior cruciate ligament tear:
immediate and longitudinal effect with clinical/MRI follow-up. Am J Sports Med, 2012. 40(2): p. 276-
85.
145. Tandogan, R.N., et al., Analysis of meniscal and chondral lesions accompanying anterior
cruciate ligament tears: relationship with age, time from injury, and level of sport. Knee Surg Sports
Traumatol Arthrosc, 2004. 12(4): p. 262-70.
146. Rodriguez-Fontenla, C., et al., Assessment of osteoarthritis candidate genes in a meta-analysis of

9 genome-wide association studies. Arthritis Rheum, 2013.
147. Hunter, D.J., et al., A genome scan for joint-specific hand osteoarthritis susceptibility: The
Framingham Study. Arthritis Rheum, 2004. 50(8): p. 2489-96.
148. MacGregor, A.J., et al., The genetic influence on radiographic osteoarthritis is site specific at the
hand, hip and knee. Rheumatology (Oxford), 2009. 48(3): p. 277-80.
149. Hunter, D.J., et al., Genetic contribution to cartilage volume in women: a classical twin study.
Rheumatology (Oxford), 2003. 42(12): p. 1495-500.
Page 162
150. Zhai, G., et al., The genetic contribution to longitudinal changes in knee structure and muscle
strength: a sibpair study. Arthritis Rheum, 2005. 52(9): p. 2830-4.
151. Berthiaume, M.J., et al., Meniscal tear and extrusion are strongly associated with progression of
symptomatic knee osteoarthritis as assessed by quantitative magnetic resonance imaging. Ann Rheum
Dis, 2005. 64(4): p. 556-63.
152. Ding, C., et al., The genetic contribution and relevance of knee cartilage defects: case-control
and sib-pair studies. J Rheumatol, 2005. 32(10): p. 1937-42.
153. Andrew, T., et al., Are twins and singletons comparable? A study of disease-related and lifestyle
characteristics in adult women. Twin Res, 2001. 4(6): p. 464-77.
154. Zhai, G., et al., Genetic influence on the progression of radiographic knee osteoarthritis: a
longitudinal twin study. Osteoarthritis Cartilage, 2007. 15(2): p. 222-5.
155. Uitterlinden, A.G., et al., Adjacent genes, for COL2A1 and the vitamin D receptor, are
associated with separate features of radiographic osteoarthritis of the knee. Arthritis Rheum, 2000.
43(7): p. 1456-64.
156. Sandmark, H., et al., Osteoarthrosis of the knee in men and women in association with
overweight, smoking, and hormone therapy. Ann Rheum Dis, 1999. 58(3): p. 151-5.
157. Slemenda, C., et al., Reduced quadriceps strength relative to body weight: a risk factor for knee
osteoarthritis in women? Arthritis Rheum, 1998. 41(11): p. 1951-9.
158. Aitken, D., et al., Responsiveness of Magnetic Resonance Imaging-derived Measures Over 2.7
Years. J Rheumatol, 2014.
159. Sharma, L., D. Kapoor, and S. Issa, Epidemiology of osteoarthritis: an update. Curr Opin
Rheumatol, 2006. 18(2): p. 147-56.
160. Cushnaghan, J. and P. Dieppe, Study of 500 patients with limb joint osteoarthritis. I. Analysis by
age, sex, and distribution of symptomatic joint sites. Ann Rheum Dis, 1991. 50(1): p. 8-13.
161. Tanamas, S.K., et al., Bone marrow lesions in people with knee osteoarthritis predict
progression of disease and joint replacement: a longitudinal study. Rheumatology (Oxford), 2010.
49(12): p. 2413-9.
162. Dore, D., et al., Bone marrow lesions predict site-specific cartilage defect development and
volume loss: a prospective study in older adults. Arthritis Res Ther, 2010. 12(6): p. R222.
163. Davies-Tuck, M.L., et al., The natural history of bone marrow lesions in community-based adults
with no clinical knee osteoarthritis. Ann Rheum Dis, 2009. 68(6): p. 904-8.
164. Berry, P.A., et al., The natural history of bone marrow lesions in community-based middle-aged
women without clinical knee osteoarthritis. Semin Arthritis Rheum, 2009. 39(3): p. 213-7.
165. Hunter, D.J., et al., Increase in bone marrow lesions associated with cartilage loss: a
longitudinal magnetic resonance imaging study of knee osteoarthritis. Arthritis Rheum, 2006. 54(5):
p. 1529-35.
Page 163
166. Kornaat, P.R., et al., Bone marrow edema-like lesions change in volume in the majority of
patients with osteoarthritis; associations with clinical features. Eur Radiol, 2007. 17(12): p. 3073-8.
167. Sowers, M.F., et al., Magnetic resonance-detected subchondral bone marrow and cartilage
defect characteristics associated with pain and X-ray-defined knee osteoarthritis. Osteoarthritis
Cartilage, 2003. 11(6): p. 387-93.
168. Lo, G.H., et al., Bone marrow lesions and joint effusion are strongly and independently
associated with weight-bearing pain in knee osteoarthritis: data from the osteoarthritis initiative.
169. Torres, L., et al., The relationship between specific tissue lesions and pain severity in persons
with knee osteoarthritis. Osteoarthritis Cartilage, 2006. 14(10): p. 1033-40.
170. Hayes, C.W., et al., Osteoarthritis of the knee: comparison of MR imaging findings with
radiographic severity measurements and pain in middle-aged women. Radiology, 2005. 237(3): p.
998-1007.
171. Driban, J.B., et al., Evaluation of bone marrow lesion volume as a knee osteoarthritis biomarker-
-longitudinal relationships with pain and structural changes: data from the Osteoarthritis Initiative.
Arthritis Res Ther, 2013. 15(5): p. R112.
172. Zhang, Y., et al., Fluctuation of knee pain and changes in bone marrow lesions, effusions, and
synovitis on magnetic resonance imaging. Arthritis Rheum, 2011. 63(3): p. 691-9.
173. Phan, C.M., et al., MR imaging findings in the follow-up of patients with different stages of knee
osteoarthritis and the correlation with clinical symptoms. Eur Radiol, 2006. 16(3): p. 608-18.
174. Baranyay, F.J., et al., Association of bone marrow lesions with knee structures and risk factors
for bone marrow lesions in the knees of clinically healthy, community-based adults. Semin Arthritis
Rheum, 2007. 37(2): p. 112-8.
175. Brennan, S.L., et al., Does an increase in body mass index over 10 years affect knee structure in
a population-based cohort study of adult women? Arthritis Res Ther, 2010. 12(4): p. R139.
176. Conaghan, P.G., H. Vanharanta, and P.A. Dieppe, Is progressive osteoarthritis an atheromatous
vascular disease? Ann Rheum Dis, 2005. 64(11): p. 1539-41.
177. Davies-Tuck, M.L., et al., Smoking is associated with increased cartilage loss and persistence of
bone marrow lesions over 2 years in community-based individuals. Rheumatology (Oxford), 2009.
48(10): p. 1227-31.
178. Davies-Tuck, M.L., et al., Increased fasting serum glucose concentration is associated with
adverse knee structural changes in adults with no knee symptoms and diabetes. Maturitas, 2012.
72(4): p. 373-8.
179. Davies-Tuck, M.L., et al., Total cholesterol and triglycerides are associated with the
development of new bone marrow lesions in asymptomatic middle-aged women - a prospective cohort
study. Arthritis Res Ther, 2009. 11(6): p. R181.
180. Dore, D., et al., A longitudinal study of the association between dietary factors, serum lipids, and
bone marrow lesions of the knee. Arthritis Res Ther, 2012. 14(1): p. R13.
Page 164
181. Wang, Y., et al., Dietary fatty acid intake affects the risk of developing bone marrow lesions in
healthy middle-aged adults without clinical knee osteoarthritis: a prospective cohort study. Arthritis
Res Ther, 2009. 11(3): p. R63.
182. Foley, S., et al., Physical activity and knee structural change: a longitudinal study using MRI.
Med Sci Sports Exerc, 2007. 39(3): p. 426-34.
183. Valdes, A.M., et al., The Ile585Val TRPV1 variant is involved in risk of painful knee
184. Lowitz, T., et al., Bone marrow lesions identified by MRI in knee osteoarthritis are associated
with locally increased bone mineral density measured by QCT. Osteoarthritis Cartilage, 2013. 21(7):
p. 957-64.
185. Laberge, M.A., et al., Obesity increases the prevalence and severity of focal knee abnormalities
diagnosed using 3T MRI in middle-aged subjects--data from the Osteoarthritis Initiative. Skeletal
Radiol, 2012. 41(6): p. 633-41.
186. Vuori, I.M., Dose-response of physical activity and low back pain, osteoarthritis, and
osteoporosis. Med Sci Sports Exerc, 2001. 33(6 Suppl): p. S551-86; discussion 609-10.
187. McAlindon, T.E., et al., Level of physical activity and the risk of radiographic and symptomatic
knee osteoarthritis in the elderly: the Framingham study. Am J Med, 1999. 106(2): p. 151-7.
188. Fransen, M. and S. McConnell, Exercise for osteoarthritis of the knee. Cochrane Database Syst
Rev, 2008(4): p. Cd004376.
189. Laslett, L.L., et al., Zoledronic acid reduces knee pain and bone marrow lesions over 1 year: a
randomised controlled trial. Ann Rheum Dis, 2012. 71(8): p. 1322-8.
190. Laslett, L.L., et al., X-ray or MRI change for chondro-protective clinical drug trials[abstr], in
Controversies, Debates and Consensus in Bone, Muscle and Joint Diseases (BMJD). 2015: Montreal.
p. p.51.
191. Englund, M., et al., Meniscal pathology on MRI increases the risk for both incident and
enlarging subchondral bone marrow lesions of the knee: the MOST Study. Ann Rheum Dis, 2010.
69(10): p. 1796-802.
192. Englund, M., et al., Incidental meniscal findings on knee MRI in middle-aged and elderly
persons. N Engl J Med, 2008. 359(11): p. 1108-15.
193. Day, B., et al., The vascular and nerve supply of the human meniscus. Arthroscopy, 1985. 1(1):
p. 58-62.
194. Mine, T., et al., Innervation of nociceptors in the menisci of the knee joint: an
immunohistochemical study. Arch Orthop Trauma Surg, 2000. 120(3-4): p. 201-4.
195. Zanetti, M., et al., Clinical course of knees with asymptomatic meniscal abnormalities: findings
at 2-year follow-up after MR imaging-based diagnosis. Radiology, 2005. 237(3): p. 993-7.
196. Khan, H.I., et al., History of knee injury and MRI-assessed knee structures in middle- and older-
aged adults: a cross-sectional study. Clin Rheumatol, 2014.
Page 165
197. Beattie, K.A., et al., Abnormalities identified in the knees of asymptomatic volunteers using
peripheral magnetic resonance imaging. Osteoarthritis Cartilage, 2005. 13(3): p. 181-6.
198. Foong, Y.C., et al., The clinical significance, natural history and predictors of bone marrow
lesion change over eight years. Arthritis Res Ther, 2014. 16(4): p. R149.
199. Dillon, E.H., et al., Follow-up of grade 2 meniscal abnormalities in the stable knee. Radiology,
1991. 181(3): p. 849-52.
200. Boegard, T.L., et al., Magnetic resonance imaging of the knee in chronic knee pain. A 2-year
follow-up. Osteoarthritis Cartilage, 2001. 9(5): p. 473-80.
201. Baker, P., et al., Sports injury, occupational physical activity, joint laxity, and meniscal damage.
J Rheumatol, 2002. 29(3): p. 557-63.
202. Snoeker, B.A., et al., Risk factors for meniscal tears: a systematic review including meta-
analysis. J Orthop Sports Phys Ther, 2013. 43(6): p. 352-67.
203. Chan, W.P., et al., Osteoarthritis of the knee: comparison of radiography, CT, and MR imaging
to assess extent and severity. AJR Am J Roentgenol, 1991. 157(4): p. 799-806.
204. Zanetti, M., et al., Patients with suspected meniscal tears: prevalence of abnormalities seen on
MRI of 100 symptomatic and 100 contralateral asymptomatic knees. AJR Am J Roentgenol, 2003.
181(3): p. 635-41.
205. Walker, P.S. and M.J. Erkman, The role of the menisci in force transmission across the knee.
Clin Orthop Relat Res, 1975(109): p. 184-92.
206. Crema, M.D., et al., Factors associated with meniscal extrusion in knees with or at risk for
osteoarthritis: the Multicenter Osteoarthritis study. Radiology, 2012. 264(2): p. 494-503.
207. Lo, G.H., et al., Strong association of MRI meniscal derangement and bone marrow lesions in
knee osteoarthritis: data from the osteoarthritis initiative. Osteoarthritis Cartilage, 2009. 17(6): p.
743-7.
208. Chang, A., et al., Subregional effects of meniscal tears on cartilage loss over 2 years in knee
209. Felson, D.T., et al., Bone marrow edema and its relation to progression of knee osteoarthritis.
Ann Intern Med, 2003. 139(5 Pt 1): p. 330-6.
210. Cibere, J., et al., Natural history of cartilage damage and osteoarthritis progression on magnetic
resonance imaging in a population-based cohort with knee pain. Osteoarthritis Cartilage, 2011. 19(6):
p. 683-8.
211. Khan, H.I., et al., Natural history of meniscal tears over 8 years in a largely non-osteoarthritic
cohort[abstract]. Internal Medicine Journal, 2014. 44(S2): p. 1-7.
212. Crema, M.D., et al., Progression of cartilage damage and meniscal pathology over 30 months is
associated with an increase in radiographic tibiofemoral joint space narrowing in persons with knee
OA--the MOST study. Osteoarthritis Cartilage, 2014. 22(10): p. 1743-7.
Page 166
213. Wluka, A.E., et al., The clinical correlates of articular cartilage defects in symptomatic knee
osteoarthritis: a prospective study. Rheumatology (Oxford), 2005. 44(10): p. 1311-6.
214. Wang, Y., et al., Meniscal extrusion predicts increases in subchondral bone marrow lesions and
bone cysts and expansion of subchondral bone in osteoarthritic knees. Rheumatology (Oxford), 2010.
49(5): p. 997-1004.
215. Davies-Tuck, M.L., et al., Development of bone marrow lesions is associated with adverse
effects on knee cartilage while resolution is associated with improvement--a potential target for
prevention of knee osteoarthritis: a longitudinal study. Arthritis Res Ther, 2010. 12(1): p. R10.
216. Khan, H.I., et al., A family history of knee joint replacement increases the progression of knee
radiographic osteoarthritis and medial tibial cartilage volume loss over 10 years. Osteoarthritis
Cartilage, 2015. 23(2): p. 203-9.
217. Wang, X., et al., Association between MRI-detected knee joint regional effusion-synovitis and
structural changes in older adults: a cohort study. Ann Rheum Dis, 2014.
218. Amin, S., et al., The relationship between cartilage loss on magnetic resonance imaging and
radiographic progression in men and women with knee osteoarthritis. Arthritis Rheum, 2005. 52(10):
p. 3152-9.
219. Davies-Tuck, M.L., et al., The natural history of cartilage defects in people with knee
220. Biswal, S., et al., Risk factors for progressive cartilage loss in the knee: a longitudinal magnetic
resonance imaging study in forty-three patients. Arthritis Rheum, 2002. 46(11): p. 2884-92.
221. Laasanen, M.S., et al., Quantitative ultrasound imaging of spontaneous repair of porcine
cartilage. Osteoarthritis Cartilage, 2006. 14(3): p. 258-63.
222. Shapiro, F., S. Koide, and M.J. Glimcher, Cell origin and differentiation in the repair of full-
thickness defects of articular cartilage. J Bone Joint Surg Am, 1993. 75(4): p. 532-53.
223. Bhosale, A.M. and J.B. Richardson, Articular cartilage: structure, injuries and review of
management. Br Med Bull, 2008. 87: p. 77-95.
224. Wang, Y., et al., Factors affecting progression of knee cartilage defects in normal subjects over
2 years. Rheumatology (Oxford), 2006. 45(1): p. 79-84.
225. Cicuttini, F., et al., Association of cartilage defects with loss of knee cartilage in healthy, middle-
age adults: a prospective study. Arthritis Rheum, 2005. 52(7): p. 2033-9.
226. Khan, H.I., et al., Natural history and clinical significance of meniscal tears over 8 years in a
midlife cohort. BMC Musculoskelet Disord, 2016. 17: p. 4.
227. Javaid, M.K., et al., Pre-radiographic MRI findings are associated with onset of knee symptoms:
the most study. Osteoarthritis Cartilage. , 2010. 18(3).
228. Ding, C., et al., Genetic mechanisms of knee osteoarthritis: a population-based longitudinal
study. Arthritis Res Ther, 2006. 8(1): p. R8.
Page 167
229. Adams, J.G., et al., Contribution of meniscal extrusion and cartilage loss to joint space
narrowing in osteoarthritis. Clin Radiol, 1999. 54(8): p. 502-6.
230. Shelbourne, K.D. and J.F. Dickens, Joint space narrowing after partial medial meniscectomy in
the anterior cruciate ligament-intact knee. J Am Acad Orthop Surg, 2007. 15(9): p. 519-24.
231. Messner, K., et al., Radiographic joint space narrowing and histologic changes in a rabbit
meniscectomy model of early knee osteoarthrosis. Am J Sports Med, 2001. 29(2): p. 151-60.
232. Bruyere, O., et al., Longitudinal study of magnetic resonance imaging and standard X-rays to
assess disease progression in osteoarthritis. Osteoarthritis Cartilage, 2007. 15(1): p. 98-103.
233. Cicuttini, F., et al., Comparison of conventional standing knee radiographs and magnetic
resonance imaging in assessing progression of tibiofemoral joint osteoarthritis. Osteoarthritis
Cartilage, 2005. 13(8): p. 722-7.
234. Pelletier, J.P., et al., What is the predictive value of MRI for the occurrence of knee replacement
surgery in knee osteoarthritis? Ann Rheum Dis, 2013. 72(10): p. 1594-604.
235. Hall, J., et al., Change in knee cartilage volume and incident meniscal extrusion as predictors of
change in joint space width of the tibiofemoral joint: 5 year longitudinal study[abstract]. Arthritis
Rheum 2011. 63(suppl 1625): p. S635.
236. Rice, J.A., Ch.14. Linear Least Sqares. In: Correlation and Regression. Vol. 2nd ed. Belmont,
CA: Mathematical Statistics and Data Analysis 1995:526-529.
237. Kenney, J.F., Keeping, E. S, Ch. 15. Linear Regression and Correlation. In: Van Nostrand, Pt. 1,
3rd ed. Princeton, NJ: Mathematics of Statistics 1962:252-285.
238. Arden, N. and M.C. Nevitt, Osteoarthritis: epidemiology. Best Pract Res Clin Rheumatol, 2006.
20(1): p. 3-25.
239. Aitken, D., et al., Magnetic resonance imaging (MRI)-assessed knee abnormalities in Australian
rules football players-AFL tas knee study. Osteoarthritis and Cartilage. 23: p. A240.
Appendix A
Page 168
This article has been removed for
copyright or proprietary reasons.
The appendices contain a letter to the editor and reply by the thesis
author. The citations are:
Kuijer, P. P. F. M. et al., 2015. Knee joint replacement and individual
susceptibility for progression of knee osteoarthritis and tibial cartilage
volume loss: not only genes run in the family, Osteoarthritis and
cartilage, 23(10), 1817-1818
Khan, H. I. et al., 2015. Reply letter to the editor: Knee joint

replacement and individual susceptibility for progression of knee
osteoarthritis and tibial cartilage volume loss: not only genes run in
the family, Osteoarthritis and cartilage, 23(10), 1819-1820
This article has been removed for
copyright or proprietary reasons.
Khan, H. I., et al., 2013. Association between

hip and knee cartilage measured using
radiographs and magnetic resonance
imaging: the Tasmanian older adult cohort
study, Rheumatology, 52(11), 2009-2015
Appendix C
Appendix C. Does cartilage volume measurement or radiographic osteoarthritis at baseline
independently predict ten-year cartilage volume loss?
Page 178
Page 179
Page 180
Page 181
Page 182
Page 183
Page 184
Page 185

B7de PDF

Uploaded by

Copyright:

Available Formats

B7de PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

B7de PDF

Uploaded by

Copyright:

Available Formats

DETERMINANTS OF STRUCTURAL CHANGES IN

THE KNEE JOINT IN A MIDDLE-AGED COHORT

Hussain Ijaz Khan, MBBS

Menzies Institute for Medical Research

Submitted in fulfilment of the requirements for the degree of

Signed: . Date: . ?. �.J _1:J o I b

The Contribution of each author:

Signed: ..... .................... . z�/o1-/;;,ot/:,

Hussain Ijaz Khan, MBBS

Signed: ...... .. . . Prof. Graeme Jones

Articles under-going peer-review:

2014 ¥ Natural history and clinical significance of meniscal tears in a mid-life

2015 ¥ Natural history and clinical significance of cartilage defects in a mid-life

2012 International Postgraduate Research Scholarship (IPRS)

2014 Finalist for Young Investigator award at Australian Rheumatology

2015 Travel Award, Graduate Research Travel Fund, University of Tasmania,

2016 World Osteoarthritis Research Society Young Researcher Scholarship

2016 European Rheumatology Association Travel Grant- EULAR congress 2016

I would like to start by thanking my primary supervisor, Professor Graeme Jones. I am

I am also extremely grateful to my co-supervisors, Dr Dawn Aitken and Associate Professor

Associate Professor Leigh Blizzard was my statistical supervisor. Early on in my candidature,

ACR American College of Rheumatology

AQoL Assessment of quality of life

BME Bone marrow edema

BMI Body mass index

BML Bone marrow lesion

CTX-II Type II collagen C-terminal telopeptide

dGEMRIC Delayed gadolinium-enhanced magnetic resonance imaging of cartilage

DMOAD Disease-modifying osteoarthritis drug

FDA Food and Drug Administration

FFQ Food frequency questionnaire

FSE Fast spin echo

GRE Gradient-recalled echo

HDL High-density lipoprotein

ICC Intraclass correlation coefficient

JSN Joint space narrowing

JSW Joint space width

K/L Kellgren and Lawrence

KOOS Knee injury and Osteoarthritis Outcome Score

MOST Multicentre Osteoarthritis Study

MRI Magnetic resonance imaging

NHMRC The National Health and Medical Research Council

NSAIDS non-steroidal anti-inflammatory drugs

RCT Randomised controlled trial

ROA Radiographic osteoarthritis

ROI Region of interest

sBMD Subchondral bone mineral density

STIR Short tau inversion recovery

TASOAC Tasmanian Older Adult Cohort study

TKR Total knee replacement

THR Total hip replacement

VAS Visual analog score

WOMAC Western Ontario and McMasters Universities Osteoarthritis Index

WORMS Whole-Organ magnetic resonance imaging score

List of Abbreviations ............................................................................................................... xv

Chapter One-Literature Review.................................................................................... 4

1.2 Knee OA............................................................................................................................... 5

1.2.1 Burden of disease....................................................................................................................... 6

1.2.2 Applied anatomy of the knee ..................................................................................................... 7