Title: THE NATURAL HISTORY OF RADIOGRAPHIC KNEE OSTEOARTHRITIS: A FOURTEEN YEAR POPULATION-BASED COHORT STUDY

Transcription

1 Full Length DOI /art Title: THE NATURAL HISTORY OF RADIOGRAPHIC KNEE OSTEOARTHRITIS: A FOURTEEN YEAR POPULATION-BASED COHORT STUDY Running Head: Natural history of radiographic knee OA Authors: K.M.Leyland 1, D.Hart 2, M.K.Javaid 1, A.Judge 1, A.Kiran 1, A.Soni 1, L.M.Goulston 3, C.Cooper 1,3, T.D.Spector 2, N.K.Arden 1,3,4 1 NIHR Musculoskeletal Biomedical Research Unit, University of Oxford, Oxford, UK 2 Department of Twin Research and Genetic Epidemiology, King s College London, London, UK 3 MRC Lifecourse Epidemiology Unit, University of Southampton, Southampton, UK 4 Department of Rheumatology, Southampton University Hospitals, Southampton, UK Corresponding Author: Prof. Nigel Arden Botnar Research Centre Institute of Musculoskeletal Sciences University of Oxford Oxford, OX3 7LD Tel: +44 (0) Fax: +44 (0) nka@mrc.soton.ac.uk Disclosures: Authors have no conflicts of interest to declare. Acknowledgements: Support was received from the Oxford NIHR Musculoskeletal Biomedical Research Unit, University of Oxford. We would like to thank all the participants of the Chingford Women Study and Maxine Daniels and Dr Alan Hakim for their time and dedication and Arthritis Research UK for their funding support to the study. This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process which may lead to differences between this version and the Version of Record. Please cite this article as an Accepted Article, doi: /art American College of Rheumatology Received: Feb 04, 2011; Revised: Dec 20, 2011; Accepted: Jan 26, 2012

2 Page 2 of 23 ABSTRACT Objective: The purpose of this study was to establish the natural history of radiographic knee osteoarthritis (ROA) over fourteen years in a community-based cohort. Methods: We examined women from the Chingford study, a community based cohort followed for over 14 years. We selected those women with bilateral, fully-extended radiographs taken at approximately five year intervals. Radiographs were scored for OA using Kellgren and Lawrence (K&L) grades blinded to order and patient identity. Descriptive statistics and odds ratios were used to compare incidence (K&L 2 from grade 0/1), worsening ( 1 increase in grade) and progression ( 1 increase in grade from a 2) of ROA. Results: A complete radiographic series was available on 56% (561) of the original 1003 subjects enrolled in the study, who had a median baseline age of 53 (Interquartile Range 48, 58). 13.7% of subjects had prevalent ROA (K&L 2) in at least one knee at baseline, which increased to 47.8% by year 15. Annual cumulative incidence, progression and worsening were 2.3%, 2.8%, and 3.0%, respectively. Baseline grade 1 s were more likely to worsen than any other baseline grade (Odds Ratio 6.6, 95% Confidence Interval (3.6, 12.7)) by year 15. Conclusion: We report the longest natural history study of radiographic knee OA to date. Our data identified relatively low rates of incidence and progression; almost half of all subjects remained free from any radiographic evidence of ROA for fifteen years; and a baseline grade 1was more likely to worsen than any other baseline grade. 2

3 Page 3 of 23 INTRODUCTION Knee osteoarthritis (OA) is one of the leading health burdens, costing 14.6 billion dollars in the US for knee replacements in 2004 alone (1). This figure does not address the additional expense of pain management, loss of work due to disability, and various treatment options such as physiotherapy and revision surgery. The economic burden of OA is increasing, with 54% more knee replacements performed in 2004 than four years earlier with this number estimated to rise to 1.4 million by the year 2015 (1). The trend has been further substantiated in a long term study based in the UK, where the rate of knee replacements tripled between 1991 and 2006 (2). Because of the increasing health burden due to the ageing population and a projected lifetime risk of developing symptomatic knee OA of 45%, there is an urgent need to understand the natural course of knee osteoarthritis in order to target preventative therapies and reduce known risk factors for both incidence and progression (3). Plain film radiographs are the most common diagnostic imaging technique used to evaluate knee OA; although other imaging modalities such as MRI, are being assessed within the research community, their advantage over radiographic assessment in clinical practice remains uncertain (4). Research into the natural history of radiographic knee osteoarthritis (ROA) has primarily focused on incidence and progression in symptomatic patients (5-8) and the progression of disease in older cohorts (9-11). Community-based studies have been limited by follow-up times that vary between three and twelve years (9-14), with little data extending beyond this period (15). Previous studies have also focused on radiographic data from only baseline and follow-up. They reported annual incidence rates ranging between 2% and 4% (9, 11, 12); rates of ROA that were significantly higher in women than men; and that rates of incident ROA were twice as much as incident symptomatic OA (9). 3

4 Page 4 of 23 The aim of this study is to assess the long-term prevalence, incidence and progression of mild and moderate radiographic disease in a well described population based cohort of middle aged women; to compare incident unilateral and bilateral disease and progressive unilateral ROA using data at five year intervals; and to assess the movement of individual Kellgren and Lawrence (K&L) grades over fourteen years. This research will establish a long-term natural history of radiographic knee osteoarthritis. SUBJECTS and METHODS Subjects Subjects were selected from the Chingford Study, a prospective population-based longitudinal study of osteoporosis and osteoarthritis. All women between the ages of 45 and 64 from a single GP practice in Chingford, North London, were contacted to take part in women out of a possible 1,353 (78% response rate) attended the baseline visit with an actual age range of 44 to 67 due to the two year recruitment period. Clinic visits included detailed musculoskeletal questionnaires, physical evaluation and knee x-rays, with pertinent risk factors for osteoarthritis such as physical activity, smoking and age evaluated by nurse-administered questionnaire. Height and weight were collected by staff (from which body mass index (BMI) has been calculated). BMI groups were calculated using World Health Organisation categories, with normal BMI as less than 25, overweight as 25 to less than 35 and obese as 35 and above. Pain was evaluated using data from the baseline clinic visit assessing the presence/absence of current pain in each knee. Much of this information was evaluated repeatedly over the course of the prospective study. Information about comorbidities, such as rheumatoid arthritis and fractures were collected, and knee replacements (TKRs) were confirmed by GP records in addition to self-report. Subjects identified as 4

5 Page 5 of 23 having rheumatoid arthritis (RA) (n=6), including those with TKRs due to RA (as determined from self-report) were excluded from the final analysis. Subjects of the Chingford study are a well-described predominantly Caucasian cohort who have been found to be representative of women of the general UK population in terms of height, weight, and rates of hysterectomy, but a lower percentage of current smokers (16). By the year 15 clinic visit, 98 women had died, 76 moved away, 22 were unable to be contacted, and 149 declined to attend. The study was approved by the Outer North East London Research Ethics Committee and written consent was obtained from each woman. Radiographic Protocols Anterio-posterior fully-extended weight bearing knee radiographs were taken at baseline (year 1), year 5, year 10 and year 15. Both knees for every subject present for the visit were radiographed by experienced radiographers using the same equipment each year. Standardised protocols were established at baseline and used for all subsequent visits, with the back of the knee kept in contact with the cassette and the patella centred over the lower portion of the femur. A tube-to-film distance of 100 cm was used, with the beam centred 2.5 cm below the apex of the patella (12). Radiographic Grading Radiographs were scored using the Kellgren and Lawrence global score (0= normal; 1=possible osteophyte, no joint space narrowing; 2= definite osteophyte, possible joint space narrowing (JSN); 3= multiple osteophytes, definite JSN, sclerosis and possible deformity of bone ends; 4= large osteophytes, marked JSN, severe sclerosis and definite deformity of bone ends) (15, 17). Total knee replacements and partial knee replacements were identified by a combination of self-report and GP records, and were further confirmed by reviewing the original x-rays. Subjects with replacements were included in the final analysis and coded 5

6 Page 6 of 23 separately. Radiographs were read individually by visit year and were blinded to order, patient identity and symptoms. Radiographs were read by the same two observers (T.S and D.H) for baseline and year 5, with a single observer (D.H) reading year 10 and year 15. As previously reported, inter/intra-observer reproducibility was calculated by reading a subset of 100 knees with an interval of three weeks. Kappas for intra-observer reproducibility were %CI (0.87, 0.89) and % CI (0.78, 0.80). Inter-observer reproducibility was high with a kappa of %CI (0.79, 0.81) (18). Statistical Methods The present analysis was conducted using a sub-set of the Chingford cohort, which included only subjects that had complete knee x-rays at baseline, year 5, year 10 and year 15 and had all pertinent baseline characteristics available. Due to the inherent limitations of complete case analysis, a post hoc available case analysis was done where possible to check for drop out bias. Subjects and knees with replacements were included in the analysis and placed in the groups K&L grade 2 or above and ROA unless explicitly listed separately. Baseline characteristics of subjects lost to follow-up were compared with those selected for this study, by age, BMI and K&L grade. As none of the continuous variables had normal distributions, Mann-Whitney U tests were used. For categorical data, Pearson s chi-square test was used except where expected cell counts were 5 or less in which case the Fishers exact test was used. Prevalence was calculated at both the subject level (using the worst knee of each subject) and the knee level (with each subject supplying two knees to the analysis) and was defined using a K&L grade 2 or above as indication of disease presence and K&L grades 0 and 1 as lack of disease. The worst-knee of each subject was determined by the knee with the highest K&L grade acting as an index knee to be used in the analysis. Incidence was calculated at both the subject-level (worst-knee) and the knee-level and was defined by 6

7 Page 7 of 23 having a K&L grade 0 or 1 at the first period of observation and a grade 2 or above at the second period of observation. Annual cumulative incidence was calculated by dividing the incidence by the number of years under observation. Incident unilateral and bilateral disease was defined as having a grade 0 or 1 in both knees at the first observation and having a 2 or above in one or both knees at the next observation point, respectively. Progression was calculated at the knee level and was defined by having a K&L grade 2 or above at the first period of observation and showing an increase of at least one K&L grade by the second period of observation. At the subject-level, progression was defined as unilateral disease at the first period of observation and bilateral disease at the second. Worsening calculated at the knee level was defined as an increase of one K&L grade from any other grade (including grades 0 and 1). This group essentially includes incident cases, progressive cases as well as mild progressors who moved between grade 0 and 1. The development of incident ROA at each time point (year 5, 10, and 15) from a baseline grade 1 was compared with baseline grade 0 s by calculating percentages for each with 95% confidence intervals. This was stratified by quartiles of age (under 50, 50 to 54, 55 to 59 and over 60 s). Differences between groups were assessed by chi-square tests. Odds ratios were used to compare incident ROA at year 15 between baseline grade 0 s and 1 s and to assess the odds of each baseline K&L grade going on to have a TKR by year 15. These were calculated by generalized estimating equation logistic regression models in order to account for clustering due to each subject contributing two knees to the analysis. Baseline characteristics (age, BMI, pain and smoking status) of subjects who developed unilateral or bilateral disease were compared using logistic regression models. Finally, a cross-tabulation was used to assess individual K&L grades at baseline and year 15. Statistical analysis was carried out using STATA version 10 (19) and SPSS version 17.0 (20). 7

8 Page 8 of 23 RESULTS The baseline median age of subjects was 53 (IQR 48, 58). Out of the original 1003 Chingford women seen at baseline, 970 women attended x-rays at baseline, 831 at year 5, 819 at year 10 and 613 at year women were present for all four x-rays had complete demographic data recorded at baseline. 442 women had incomplete follow-up data and were excluded from the complete case analysis. Subjects lost to follow-up were slightly older (+6 years, p<0.05), were more likely to be current smokers and had a slightly higher percentage of knee pain (16% vs. 11%) (table 1). The prevalence of subjects with ROA (worst-knee with K&L 2) was 13.7% at baseline, 23.9% at year 5, 36.4% at year 10 and 47.8% at year 15. Out of all knees (n=1122), ROA prevalence was 9.5% at baseline, 17.5% at year 5, 27.5% at year 10 and 38.6% at year 15. Interval rates of annual cumulative incidence at the subject-level (worst-knee) were 3.0% between baseline and 5, 3.4% between 5 and 10 and 3.9% between 10 and 15, with 39.5% of subjects developing incident ROA in at least one knee between baseline and year 15. At the knee level, annual cumulative incidence also increased steadily between each five-year period, from 2.3% between baseline and year 5, 2.6% between year 5 and 10 and 3.3% between year 10 and 15, with 32.5% of knees developing incident ROA between baseline and year 15. The annual cumulative incidence between baseline and year 15 was 2.3% at the knee-level and 2.8% per year for worst-knee subject-level analysis. Out of knees with a grade 2 or above at baseline (n=106), 38.7% progressed between baseline and year % of knees (n=1122) showed worsening (increase of at least one grade) between baseline and year 5, 23.4% between year 5 and 10, and 23.8% between 10 and 15. Analysing only baseline and year 15 data, 41.5% of knees showed worsening by at least one grade. A sensitivity analysis using all available knees at each time point did not show any significant difference from the complete case analysis (data not shown). 8

9 Page 9 of 23 When 5-year cumulative incidence of ROA examined at each time point and age is stratified into quartiles (<50, 50-54, 55-60, >60), a linear trend (p<0.002) is evident with the oldest age group having the highest percentage of incident ROA (figure 1). By year 15, 26.0% of subjects with a baseline age less than 50 had developed incident ROA, 34.1% of subjects aged 50 to 54, 31.7% of 55 to 60 year olds and 42.2% of subjects who had a baseline age greater than 60. This difference was significant between youngest and oldest age groups (p<0.01), although not between the two middle age-groups (p=0.584). When age is stratified into two age bands (<55 and 55), there is a statistically significant difference (p= 0.017) between incident ROA in the two groups. When 5-year cumulative incidence is analysed by BMI category, the percentage of incident ROA knees at year 5 is roughly similar between groups. By year 10 and 15, however, the cumulative incidence for obese subjects was almost 20% higher than both the normal and overweight categories. No difference was found between pre and post-menopausal (baseline) subjects in terms of ROA incidence (p=0.193) A cross-tabulation of individual K&L grades (0-4) and TKR s between baseline and year 15 (table 2) demonstrates that 51.3% of all knees (n=1122) remained a grade 0 throughout the study period, while 41.5% of knees worsened by at least one grade. Out of the subjects with a K&L of 1 or above at baseline (n=167), 37.1% remained at the same grade and 51.5% worsened (including to TKR) by year 15. Knees with a baseline grade 1 (n=61) had a higher percentage of progression than any other baseline grade (73.8%, 45/61). Grade 2 s (n=76) were the next most likely to progress, with 47.7% increasing by at least one grade over 15 years. 1.7% of knees were scored as having regressed to a lower grade by year % (10/955) of baseline grade 0 s went on to have a total knee replacement by year 15, 4.9% (3/61) of grade 1 s, 5.3% (4/76) of grade 2 s and 6.7% (2/30) of grade 3 s. 1.2% (11/957) of baseline grade 0 s went on to have a total knee replacement by year 15, 4.9% (3/61) of grade 1 s, 5.3% (4/76) of grade 2 s and 6.7% (2/30) of grade 3 s. The percentage of 9

10 Page 10 of 23 knees with baseline pain when broken down by baseline K&L grade was 21.5 for grade 0, 19.7 for grade 1, 39.5 for grade 2 and 26.7 for grade 3. When rates are analysed for all knees (n=1122), the group with the highest number of total knee replacements by year 15 had a baseline grade of 0. The odds of having a TKR by year 15 were similar for baseline grades 1 through 3 when compared with a baseline grade of 0 (grade 1: OR 4.7 (95% CI 1.0, 22.2); grade 2: OR 5.9 (95% CI 1.9, 18.2); grade 3: OR 4.6 (95% CI 0.3, 65.3)) percent of knees with a TKR by year 15 were reported to have pain at the baseline visit. While a K&L grade 1 is not considered diagnostic of ROA, when stratifying by an initial baseline grade of 0 or 1 (figure 2), there is a significant difference in cumulative incidence between the groups at each visit (p<0.001). The odds of a baseline grade 1 developing incident ROA by year 15 is 4.5 (95% CI 2.7, 7.4) times the odds for a baseline grade 0. When stratified by age groups, these differences remained significant (p<0.05) at all visits and in all age groups except at year 5 in the 60+ group (OR 2.8, 95% CI (0.9, 9.0), p=0.085). The prevalence of unilateral and bilateral ROA (table 3) at baseline was 8.6 and 5.2% respectively, while by the year 15 it was 18.4% for unilateral disease and 29.4% for bilateral disease. Of the subjects without ROA at baseline, 60.5% (293/484) remained disease free, 29.2% (14/48) of the subjects with unilateral disease and 96.6% (28/29) of the bilateral subjects remained stable by year 15. Out of the subjects who developed either incident unilateral or bilateral disease (n=200) within any five-year follow-up period, 32.0% developed bilateral disease while 68.0% developed unilateral disease. Out of subjects with unilateral disease between baseline and year 10 (n=143), 56.6% had progressed to bilateral disease by year 15. When cumulative incidence was stratified by age quartiles for baseline and year 15 data, a significant difference was found between the youngest (<55) and oldest 10

11 Page 11 of 23 (>60) age groups (p=0.003), although adjacent age groups were not significantly different from one another (p=0.642). Baseline characteristics were compared between subjects who remained disease free over the 15 year study period and subjects who progressed to unilateral or bilateral disease. All subjects who progressed were more likely to have a higher BMI; those developing bilateral disease (from no disease or unilateral) were more likely to be older and subjects developing unilateral disease and those progressing from unilateral to bilateral were more likely to have pain at baseline than the subjects who remained free from radiographic disease. DISCUSSION The novel findings of this research were that annual cumulative incidence, progression and worsening of knee ROA were 2.3, 2.8 and 3.0% respectively over 14 years of follow-up; three potential phenotypes exist for knee ROA based on symmetry: incident unilateral, incident bilateral and progressive unilateral to bilateral disease; and that although risk of TKR was associated with increasing baseline K&L grade, the majority of knees with a TKR by year 15 had a baseline grade 0. Over half the subjects (52.2%) remained free from radiographic knee osteoarthritis over the course of the study. 38.6% of knees had prevalent ROA at visit 15, which increased from 9.5% at baseline. Annual rates of knee progression (2.8%) and worsening (3.0%) between baseline and year 15 were found to be slightly lower than those observed in other community-based cohorts which varied between 3.5 and 8.0% for progression (8, 9, 11) and 4.4% for worsening (11). Rates in established symptomatic cohorts were similar, varying between 3.3 and 7.7% for worsening (6, 21) and 4.0 to 8.8% for progression (7, 8). The slightly lower rates found in Chingford are likely a consequence of both the relatively young age of the cohort at the start of the study and the length of the study. 11

12 Page 12 of 23 This study found an annual cumulative incidence of 2.3% between baseline and year 15. When broken down by the five-year interval data, the annual rate between baseline and year 5 is in the lower range of 2.3% with a high of 3.3% between year 10 and 15. This is likely a result of the increasing age of the sample, with a median (IQR) age of 53 (48, 58) at baseline and 68 (63.5, 72.5) by year 15. When cumulative incidence was stratified by age quartiles, a significant difference was found between the youngest (< 55) and oldest age groups (> 60), although adjacent age groups were not significantly different from one another. Analyses conducted in primarily symptomatic cohorts have found much higher percentages of annual cumulative incidence, with rates up to 4.0% (8). Cohort studies and case-control studies which included both symptomatic and non-symptomatic subjects were more comparable to these results ranging between 2.0 and 2.5% (9-11). As would be expected, obese subjects had higher percentages of incidence than other BMI groups, with almost 20% more by clinic visit 10. Assessing a cross-tabulation of individual K&L grades over 15 years, rather than using K&L as a binary variable, demonstrated that specific grades are more likely to progress to a higher grade (grade 1), or to remain stable (grade 3) even over a long period of time. Less than half of all knees (41.5%) worsened by at least one K&L grade over the 14 years of the study. The majority of subjects (68.4%) who went on to have a knee replacement by year 15 did not have evidence of conventional ROA (K&L 2) at baseline, and over half had current knee pain. This suggests that radiographs are not necessarily the optimal tool for predicting long-term TKR outcomes in younger patients (median age 53.0 at baseline). The comparison of incident ROA between subjects with baseline grades 0 and 1 extends the timeline of an earlier nested case-control study within this cohort (22) and other recent research (23) which emphasise the importance of grade 1 s being treated distinctly from grade 0 s. The higher risk of a grade 1 going on to develop incident ROA (6.6 times the 12

13 Page 13 of 23 chance of a grade 0) suggests that they are an important indicator of longitudinal radiographic incidence. Lachance et al. looked at incidence and progression of mild K&L grades over 3 years and found that a grade 1 was 6.4 times more likely to progress to a grade 2 than a grade 0 (13). Of the subjects who developed incident ROA, approximately one-third developed bilateral ROA while the remaining two-thirds developed unilateral ROA between each clinic visit. Over one-third of the unilateral knees progressed to bilateral disease between each clinic visit, while the rest remained stable. These groups showed differences in baseline characteristics (age, BMI and pain) from subjects who remained free from radiographic OA over the study period. This data is possibly describing three distinct subsets of ROA, where some have slow progression from no disease through unilateral then bilateral ROA, while others have more rapid progression to bilateral disease within a five-year period. This could reflect a difference between environmental factors (i.e. functional effect of having contralateral knee OA) and genetic factors (i.e. genetic predisposition to ROA) (24), although further work is required to validate these findings. Limitations of this study include the effect of radiographic views, scoring methods, radiographic blinding, inclusion criteria and the most common limitation of studies of this length, loss to follow-up. The standard radiographic view for knees at the start of the Chingford study was AP, fully extended and weight-bearing. Although the prevailing opinion is for standard semi-flexed views (25, 26), and the more rigorous fluoroscopy-assisted positioning (27) due to underestimation of joint space narrowing in fully extended views (28), long-term studies often continue using the same radiographic protocol as the baseline visit in order to more accurately evaluate change. The patellofemoral (PF) compartment is known to be an important component of whole organ knee OA, and the presence of PF ROA is highly associated with pain and disability (29, 30). The lack of additional radiographic views able to 13

14 Page 14 of 23 image the PF compartment such as skyline and/or lateral in this study is a limitation which should be addressed in future radiographic natural history studies. The Kellgren and Lawrence scoring system has been the primary method for evaluating radiographic knee OA in similar studies (5, 7-11, 13, 14) but is commonly criticised for several known limitations. K&L assumes a non-validated natural disease progression which is extremely osteophyte-centric, as well as having several different official and modified versions which are all in use (15, 17, 31). Despite these negatives, K&L scores have a high level of reproducibility (8, 11, 32), a strong correlation with pain and increasing K&L grades (33, 34), and is present in the majority of subjects who present with knee pain (35). While this relationship with pain is by no means perfect, other imaging modalities, such as MRI, have not yet demonstrated a better specificity than plain film x-ray (36). A potential limitation of reading radiographs blind to order is that grades may decrease over time as they are not being read in a method that allows for the evaluation of change. The percentage of regressive grades in this study (1.7%) is much lower than studies using similar blinding methods ( %) (13, 22). Interval censoring was utilised due to the lack of an exact date of radiographic incidence/progression/worsening for each knee. Scores were only evaluated on the date of each radiograph, so that knees that progressed immediately after a visit would be recorded as not having worsened until the next clinic visit 4-5 years later. This may have contributed to an overestimation of the amount of time it took subjects to have incident or progressive ROA. Due to the original study design, the results of this study are also restricted to the natural history of ROA in Caucasian women. While the results should be able to be loosely applied to men, and women of other ancestry, there are known differences in prevalence, incidence and progression between these groups (23, 29, 34, 37). The high projected lifetime risk for symptomatic knee OA (45%) emphasizes the importance of using symptoms when 14

15 Page 15 of 23 defining OA (3). The lack of pain information in this analysis is a limitation for clinical application and should be addressed in future work. Subjects lost to follow-up are a major limitation of all long-term cohort studies, such as Chingford. There is a potential for study bias due to deaths, subjects dropping out because of disability and illness and generally having a healthier cohort attending the follow-up visits. While the comparison of baseline characteristics of those included in analysis versus those lost to follow-up are similar enough not to imply a severe bias (i.e. slightly older, more likely to smoke and slightly more knee pain), the unknown possibility remains that subjects lost to follow-up, for whatever reason, would have had significantly worse ROA than those included in this analysis. There is no way to know the potential effect of this type of bias on any study of this design. The use of complete-case analysis can add further bias due to the need to drop subjects not present for intermediate visits, however only a small percentage of subjects present at baseline and year 15 were missing additional visits. Out of subjects with both baseline and year 15 data, 29 were excluded due to missing data at year 5, eight for missing year 5 data, and three for missing both visits. This analysis is the longest natural history study of radiographic knee OA to date, and is intended to provide novel data regarding the trend of individual Kellgren and Lawrence grades during 5 year intervals over 14 years. No other natural history study of this type has included intermediate radiographic scores between the baseline and follow-up data. The inclusion criteria were deliberately non-restrictive in order to gain a more accurate picture of ROA in a normal population, including the analysis of the progression of mild radiographic OA in a relatively young cohort. Chingford has an extremely high response rate of subjects for a study of this length, with over 50% of the original 1003 women attending all clinic visits involving knee radiographs over the 14 years of follow-up 15

16 Page 16 of 23 In conclusion, this research has identified that annual cumulative incidence, progression and worsening of knee ROA were 2.3, 2.8 and 3.0%, respectively; over half of the subjects remained free from ROA over 14 years; three potential phenotypes exist for knee ROA based on symmetry; the odds of baseline grade 1 s to develop incident ROA were 4.5 times greater than a baseline 0; and that the majority of knees that had a TKR by follow-up, did not have ROA at baseline. 16

17 Page 17 of 23 REFERENCES 1. Kim S: Changes in surgical loads and economic burden of hip and knee replacements in the US: Arthritis Rheum 2008, 59(4): Culliford DJ, Maskell J, Beard DJ, Murray DW, Price AJ, Arden NK: Temporal trends in hip and knee replacement in the United Kingdom: 1991 to J Bone Joint Surg Br 2010, 92(1): Murphy L, Schwartz TA, Helmick CG, Renner JB, Tudor G, Koch G, Dragomir A, Kalsbeek WD, Luta G, Jordan JM: Lifetime risk of symptomatic knee osteoarthritis. Arthritis Rheum 2008, 59(9): Hannan MT, Felson DT, Pincus T: Analysis of the discordance between radiographic changes and knee pain in osteoarthritis of the knee. J Rheumatol 2000, 27(6): Spector TD, Dacre JE, Harris PA, Huskisson EC: Radiological progression of osteoarthritis: an 11 year follow up study of the knee. Ann Rheum Dis 1992, 51(10): Massardo L, Watt I, Cushnaghan J, Dieppe P: Osteoarthritis of the knee joint: an eight year prospective study. Ann Rheum Dis 1989, 48(11): Pavelka K, Gatterova J, Altman RD: Radiographic progression of knee osteoarthritis in a Czech cohort. Clin Exp Rheumatol 2000, 18(4): Thorstensson CA, Andersson ML, Jonsson H, Saxne T, Petersson IF: The natural course of knee osteoarthritis in middle-aged individuals with knee pain - A 12 year follow-up using clinical and radiographic criteria. Ann Rheum Dis Felson DT, Zhang Y, Hannan MT, Naimark A, Weissman BN, Aliabadi P, Levy D: The incidence and natural history of knee osteoarthritis in the elderly. The Framingham Osteoarthritis Study. Arthritis Rheum 1995, 38(10): Felson DT, Zhang Y, Hannan MT, Naimark A, Weissman B, Aliabadi P, Levy D: Risk factors for incident radiographic knee osteoarthritis in the elderly: the Framingham Study. Arthritis Rheum 1997, 40(4): Cooper C, Snow S, McAlindon TE, Kellingray S, Stuart B, Coggon D, Dieppe PA: Risk factors for the incidence and progression of radiographic knee osteoarthritis. Arthritis Rheum 2000, 43(5): Hart DJ, Doyle DV, Spector TD: Incidence and risk factors for radiographic knee osteoarthritis in middle-aged women: the Chingford Study. Arthritis Rheum 1999, 42(1): Lachance L, Sowers MF, Jamadar D, Hochberg M: The natural history of emergent osteoarthritis of the knee in women. Osteoarthritis Cartilage 2002, 10(11): Schouten JS, van den Ouweland FA, Valkenburg HA: A 12 year follow up study in the general population on prognostic factors of cartilage loss in osteoarthritis of the knee. Ann Rheum Dis 1992, 51(8): Kellgren JH, Lawrence JS: Radiological assessment of osteo-arthrosis. Ann Rheum Dis 1957, 16(4): Hart DJ, Mootoosamy I, Doyle DV, Spector TD: The relationship between osteoarthritis and osteoporosis in the general population: the Chingford Study. Annals of the Rheumatic Diseases 1994, 53(3): Kellgren JH, Lawrence JS: Atlas of standard radiographs: the epidemiology of chronic rheumatism. Vol. 2. In.: Oxford: Blackwell;

18 Page 18 of Spector TD, Hart DJ, Byrne J, Harris PA, Dacre JE, Doyle DV: Definition of osteoarthritis of the knee for epidemiological studies. Ann Rheum Dis 1993, 52(11): StataCorp: Stata Statistical Software: Release 10. In. College Station, TX: StataCorp LP; SPSS: Release In. Chicago: SPSS Inc.; Spector TD, Cooper C, Cushnaghan J, Hart DJ, Dieppe PA: A radiographic atlas of knee osteoarthritis. In.: London: Springer Verlag; Hart DJ, Spector TD: Kellgren & Lawrence grade 1 osteophytes in the knee-- doubtful or definite? Osteoarthritis Cartilage 2003, 11(2): Lachance L, Sowers M, Jamadar D, Jannausch M, Hochberg M, Crutchfield M: The experience of pain and emergent osteoarthritis of the knee. Osteoarthritis Cartilage 2001, 9(6): Driban JB, Sitler MR, Barbe MF, Balasubramanian E: Is osteoarthritis a heterogeneous disease that can be stratified into subsets? Clin Rheumatol 2010, 29(2): Brandt KD, Mazzuca SA, Conrozier T, Dacre JE, Peterfy CG, Provvedini D, Ravaud P, Taccoen A, Vignon E: Which is the best radiographic protocol for a clinical trial of a structure modifying drug in patients with knee osteoarthritis? J Rheumatol 2002, 29(6): Buckland-Wright C: Radiographic assessment of osteoarthritis: comparison between existing methodologies. Osteoarthritis Cartilage 1999, 7(4): Mazzuca SA, Brandt KD, Katz BP: Is conventional radiography suitable for evaluation of a disease-modifying drug in patients with knee osteoarthritis? Osteoarthritis Cartilage 1997, 5(4): Buckland-Wright C: Protocols for precise radio-anatomical positioning of the tibiofemoral and patellofemoral compartments of the knee. Osteoarthritis Cartilage 1995, 3 Suppl A: McAlindon TE, Snow S, Cooper C, Dieppe PA: Radiographic patterns of osteoarthritis of the knee joint in the community: the importance of the patellofemoral joint. Ann Rheum Dis 1992, 51(7): Duncan RC, Hay EM, Saklatvala J, Croft PR: Prevalence of radiographic osteoarthritis--it all depends on your point of view. Rheumatology (Oxford) 2006, 45(6): Lawrence JS: Rheumatism in Populations. In. London, UK: William Heinemann Medical Books Ltd; Neame RL, Muir K, Doherty S, Doherty M: Genetic risk of knee osteoarthritis: a sibling study. Ann Rheum Dis 2004, 63(9): Neogi T, Felson D, Niu J, Nevitt M, Lewis CE, Aliabadi P, Sack B, Torner J, Bradley L, Zhang Y: Association between radiographic features of knee osteoarthritis and pain: results from two cohort studies. BMJ 2009, 339:b Cho HJ, Chang CB, Yoo JH, Kim SJ, Kim TK: Gender differences in the correlation between symptom and radiographic severity in patients with knee osteoarthritis. Clin Orthop Relat Res 2010, 468(7): Duncan R, Peat G, Thomas E, Hay E, McCall I, Croft P: Symptoms and radiographic osteoarthritis: not as discordant as they are made out to be? Ann Rheum Dis 2007, 66(1): Kornaat PR, Bloem JL, Ceulemans RY, Riyazi N, Rosendaal FR, Nelissen RG, Carter WO, Hellio Le Graverand MP, Kloppenburg M: Osteoarthritis of the knee: 18

19 Page 19 of 23 association between clinical features and MR imaging findings. Radiology 2006, 239(3): Odding E, Valkenburg HA, Algra D, Vandenouweland FA, Grobbee DE, Hofman A: Associations of radiological osteoarthritis of the hip and knee with locomotor disability in the Rotterdam Study. Ann Rheum Dis 1998, 57(4): TABLES Table 1. Baseline Characteristics of the entire cohort, subjects with and without complete follow-up Entire Cohort Complete Followup Lost to follow-up p-value N=1003 N=561 N=442 Age (years), median (IQR) 54.0 (49.0,60.0) 53.0 (48.0, 58.0) 56.0 (50.0, 61.0) < BMI (kg/m 2 ), median (IQR) 24.8 (22.6,27.6) 24.7 (22.7, 27.3) 25.1 (22.6, 28.2) BMI-Normal (<25 ), % Current Smokers, % Worst Knee-KL 0, % Worst Knee-KL 1, % Worst Knee-KL 2, % Worst Knee-KL 3, % Worst Knee-KL 4, % Worst Knee-TKR, % Pain, % p-values are comparing subjects with complete follow-up versus subjects lost to follow-up with significant p-values in bold Table 2. Cross Tabulation of baseline and year 15 K&L grades* (n=1122knees) Baseline N Year TKR (60.2) 95 (10.0) 157(16.4) 116 (12.2) 2 (0.2) 10(1.1) (19.7) 4 (6.6) 24 (39.3) 18 (29.5) 0 (0.0) 3 (4.9) (0.0) 1 (1.3) 39 (51.3) 32 (42.1) 0 (0.0) 4 (5.3) (3.3) 1 (3.3) 4 (13.3) 19 (63.3) 3 (10.0) 2 (6.7) *Results are displayed as N (%), with percentages calculated by row 19

20 Page 20 of 23 Table 3. Cross Tabulations of None, Unilateral and Bilateral knee ROA (K&L 2+) for baseline by year 5, year 5 by year 10, year 10 by year 15 Baseline Year 5 Year 10 Year 5 N None Uni Bi None (88.0) 42 (8.7) 16 (3.3) Unilateral 48 1 (2.1) 30(62.5) 17 (35.4) Bilateral 29 0 (0.0) 0 (0.0) 29(100.0) Year 10 None (83.1) 51(11.9) 21 (4.9) Unilateral 72 2 (2.8) 44(6.1.1) 26 (36.1 Bilateral 62 0 (0.0) 5 (8.1) 57 (91.9) Year 15 None (80.4) 43 (12.0) 27 (7.7) Unilateral (6.0) 56 (56.0) 38 (38.0) Bilateral (0.0) 4 (3.8) 100(96.2) Results are displayed as N(%), with percentages calculated by row FIGURE CAPTIONS Figure 1. The point estimate and 95% CI for the cumulative percentage of women with incident radiographic OA is shown for each visit stratified by baseline age group. (N = 1016 knees) Figure 2. The point estimate and 95% CI for the cumulative percentage of women with incident radiographic OA is shown for each visit stratified by subjects with a baseline K&L grade 0 or 1. (N= 1016 knees) Figure 3. The point estimate and 95% CI for the cumulative percentage of women with incident radiographic OA is shown for each visit stratified by baseline BMI category. (N= 1016 knees) 20

21 Page 21 of 23 Figure 1. The point estimate and 95% CI for the cumulative percentage of women with incident radiographic OA is shown for each visit stratified by baseline age group. (N = 1016 knees)

22 Page 22 of 23 Figure 2. The point estimate and 95% CI for the cumulative percentage of women with incident radiographic OA is shown for each visit stratified by subjects with a baseline K&L grade 0 or 1. (N= 1016 knees)

23 Page 23 of 23 Figure 3. The point estimate and 95% CI for the cumulative percentage of women with incident radiographic OA is shown for each visit stratified by baseline BMI category. (N= 1016 knees)