Some recent clinical approaches to osteoarthritis research

Some recent clinical approaches to osteoarthritis research

Some Recent Clinical Approaches to Osteoarthritis Research By Paul Dieppe Clinical investigation of osteoarthritis (OA) is the starting point of t...

936KB Sizes 0 Downloads 24 Views

Some Recent Clinical


to Osteoarthritis


By Paul Dieppe Clinical investigation of osteoarthritis (OA) is the starting point of the “research triangle,” leading to appropriate animal and in vitro studies. This article discusses the historical perspective of clinical investigation, definition of OA, case studies taking place in Bristol, assessment and measurement of OA, including imaging of various types and possible biochemical “markers” of the OA process. The different aspects of the process of OA need to be related to outcome, and the future for clinical research in OA is postulated. 0 1990 by W.B. Saunders Company. INDEX WORDS:

Osteoarthritis; pathogenesis; cartilage:

history of medicine.


HE “RESEARCH TRIANGLE” of disease investigation includes clinical observation, animal modelling, and in vitro experimentation (Fig 1). Clinical observation comes first, and points the way towards the appropriate animal and in vitro work. This article argues the case for continued clinical investigation of osteoarthritis (OA), and suggests that much of the current animal and in vitro work is inappropriate because clinicians have not been able to formulate the right questions. OA Research in Bristol The Rheumatology Unit at Bristol University has devoted much of its research effort towards the clinical investigation of OA. Over the last 10

From the Department ofRheumatology, University of Bristol, Bristol, UK. Supported by the Arthritis and Rheumatism Council (UK). Presented at the ILAR Conference in Rio de Janeiro in 1989. Received the Roussel Prize for Clinical Research in Osteoarthritis. Paul Dieppe, BSc, MD, FRCP: ARC Professor of Rheumatology, University of Bristol, Bristol, UK. Address reprint requests to Paul Dieppe, MD, Department ofRheumatology, Bristol Royal Infirmary, Malborough Street, Bristol BS2 Hi W, UK. 0 1990 by W.B. Saunders Company. 0049-0172/90/2003-1002$S.O0/0 2


years numerous collaborators have been involved, some of whom are listed.* HISTORICAL PERSPECTIVE

Clinical research on OA is not new. Many superb observers and clinical scientists have made great contributions in the past. They include many of the well known fathers of modern rheumatology such as Heberden, Charcot, and the Garrods, father and son. Many of our recent observations have only underlined how much our predecessors knew, in spite of inferior technology. Three clinicians have had a special influence on the development of OA research in the Rheumatology Unit in Bristol: Haygarth, Adams, and Kellgren . John Haygarth was a great believer in the power of clinical observation. He wrote, in 1805: The term rheumatism has been applied without sufficient discrimination to a great variety of disorders, which, except pain, have but few symptoms that connect them together.’

He continues to explain to the reader that, inspired by William Heberden, he decided, from 1767, to record notes on all his patients. One of the conditions to emerge from his careful record keeping was “nodosity of the joints” which he had observed in 34 out of 10,549 patients seen. There was only one man in Haygarth’s series, and he records that the condition generally started around the time of the menopause. The findings of painful nodules at the distal interphalangeal joints of the hands, and the association with knee disease are well described. He recorded gradual increase in the number of joints affected and followed many of his patients for years. He was not sure of the cause or the treatment, writing as follows: “as so little knowledge has been practically ascertained upon this subject, I should advise an attentive and comparative trial of Bathing and * Sara Browning, Giles Campion, Margaret Clarke, Cyrus Cooper, Janet Cushnaghan, Michael Doherty. Charles Hutton, John Kirwan, Tim McAlindon, Fiona McCrae, Maria Palmer, Iain Watt, Jennie Shouls, Angela Swan, Juliet Rogers.

in Aflhritis and Rheumatism, Vol20,

No 3, Suppl 1 (December), 1990: pp 2-l 1












The “research


of disease


Douching.” A good deal of our work has confirmed Haygarth’s findings, and the cynic might remark that little has changed in the 285 years since he published his monograph. Robert Adams’s landmark rheumatology book of 18517~combined clinical observation with pathology, and illustrated the text with a wonderful atlas. He clearly distinguished between many different forms of arthritis, recognizing the importance of joint site in the evolution of “chronic rheumatic arthritis” (OA). He distinguished a localized and a generalized form of hip disease, and stressed the role of bone hypertrophy. As described elsewhere,3 he also documented a case of chondrocalcinosis with a chronic arthropathy. Adams’s work paved the way for subsequent clinicians, pathologists and radiologists to make distinctions between the “atrophic” (RA) and “hypertrophic” (OA) responses of synovial joints.4 His work is another beautiful example of the value of careful clinical observation. J.H. Kellgren has been one of the greatest contributors to OA research amongst contemporary rheumatologists. His work on pain, published in the 1930s and 1940s5.’ remains unsurpassed, and is arguably the most important and relevant contribution to our understanding of the symptoms of OA. Among his many other contributions, his papers on “generalized osteoarthritis” (GOA) are still widely quoted.7.s We now believe that there needs to be some adjustment to the Kellgren concept of GOA, but his work and encouragement have been an inspiration and a fine example, and his papers, like those of Haygarth and Adams, are standing the test of time.

as it relates

to current



on osteoarthritis.


Haygarth, Adams, and Kellgren avoided the terminological issues and problems of definition that have bedevilled OA of late. Obsessions with disease definition and diagnostic criteria are relatively recent, and perhaps unhelpful developments.“.” However, although the name is unimportant, it is impossible to avoid the issue of definition completely. Haygarth recognized a clinical disease pattern, without the use of complex statistics. Adams went further, dividing arthritis according to the joint sites involved and the pattern of pathological reaction. Both realized that chronic destructive arthritis was a heterogeneous form of joint disorder, involving a mixture of damage, inflammation, and a repair reaction. Arthritis was subsequently divided into atrophic and hypertrophic forms by pathology and radiology.4 Kellgren and Lawrence then formulated the classical radiological criteria and grading system used to define OA.‘* The problem with OA is the lack of a good relation between the clinical features and the pathology, as shown by a roentgenogram. ‘x’~ As the roentgenogram has become the “gold standard” for epidemiological and other studies of OA, this presents problems. Our own approach has been to adhere to the concepts of our predecessors. OA is defined radiologically, with focal areas of joint space narrowing regarded as the sine-qua-non, and bone hypertrophy the second most important feature (ie, we have followed the concept of the hypertrophic form of joint reaction). The presence of


any clinical features, including possible associations or aetiological factors, is noted. The chief joint site(s) of involvement are also used as a major way of categorizing patients. Certain other rheumatic conditions may result in exclusion from a diagnosis of OA, although the coexistence of the OA response with other forms of arthritis is well recognized. Many of our case studies, referred to below, of necessity, have involved only those patients who have significant symptoms and disability as well as radiological evidence of OA in their index joints. However, we are now doing parallel community-based studies, recording radiological evidence of OA in the knee, and relating this to symptoms and risk factors separately. The problems and advantages of these approaches have been fully discussed in the recent literature.‘0.“,‘5-‘7 In summary we define/describe OA as a heterogeneous group of conditions of synoviai joints, characterized by focal loss of articular cartilage and hypertrophy of the underlying/surrounding bone. The roentgenogram is used as a convenient surrogate for pathology to recognize this state of the joint. OA is an age-related, dynamic, changing process, combining features of joint damage and a repair response. This state of the joint is sometimes accompanied by symptoms (especially userelated pain and/or disability). As it has different associations and outcome at different joint sites, the main site(s) of involvement must be used as a major descriptive feature. CASE STUDIES OF LIMB JOINT OA

In the belief that further descriptive work on the nature of OA and its natural history is needed, a series of case studies have been established in Bristol by Janet Cushnaghan, Fiona McCrae, and others. The most important of these is a prospective study of 500 people with limb joint OA-The OA 500 study.9,‘8 OA 500 Five hundred consecutive patients referred to a rheumatology clinic because of joint problems, and subsequently described with limb joint OA as the major cause of their problems, have entered this study. The investigation began in 1985, and the majority of those entered have already had a major review 3 years after entry. The study group comprises 342 women with a mean age of 65 years at entry and 158 men

with a mean age of 60 years. The age range was from 25 to 87, and the majority reported a long history of joint pain in the index sites of involvement when first seen. Full details of the entry criteria and methods of study, and some of the cross-sectional data have been reported elsewhere.’ Further analysis of both cross-sectional and longitudinal data is currently under way. One of the first questions we have asked of this data base concerns the relation between the sites of involvement, age, sex, and risk factors (Table 1). As expected, the main joints involved were knees (4 1%) hands (30%), and hips ( 19%). The majority had one, two, or three sites of involvement. Hips, ankles, and elbows had a male preponderance; hands, knees, and shoulders had a strong female preponderance. Hip involvement occurred in the youngest group; shoulder and wrist disease was associated with increased age. There was also a strong association between the total number ofjoint sites involved and increasing age (Table 1). There would seem to be two possible reasons for this-either people with an older onset tend to get more joints affected, or, as Haygarth suggested, the disease is characterized by slow addition of new joint sites. Analysis of both this and other patient series” strongly suggest that Haygarth was right. If that is the case, GOA may not be a distinct entity, as patients will often first have monoarticular disease, subsequently developing oligo and polyarticular involvement as they age. The existence of GOA and of OA subsets has also been investigated by analyzing the association between OA at different joint sites when more than one type of joint is affected. If OA strikes randomly at the vulnerable joints, we would expect no relation between sites, even though polyarticular disease is common. If, however, there are definable subsets, we might expect to see a positive correlation between joint sites in certain age or sex groups. As shown in Table 2, such associations were sought in the 182 of 500 patients with only two joint sites involved. A strong positive correlation was found between hand and knee disease in women, but there was no linkage of hip involvement to other sites. This is in accord with the anecdotal observations of Haygarth, Kellgren and Moore, *,7and others. The knee-hand group were also the most likely to be overweight and hypertensive, and often had a family history, suggesting a strong systemic pre-





disposition to OA in these patients. To the best of our knowledge, this is the first good numerical support for the concept of GOA, even though it changes the definition somewhat. Our data also emphasizes the striking difference in age, sex, and associations between hip and knee disease noted by others2 These two disorders must be regarded as quite different conditions, and not amalgamated for clinical investigations and trials, as so often happens.


1) Study


in 500




and Women





Men Hip



+ Knee








Hand 15 3 (NV

K:;(Ns,H 1 ::;,Ns,:::” ‘;;!

Reprinted with permission

from Cushnaghan and Dieppe.

OA, showing that the presence of hand OA relates to the development of OA after a menisectomy, would tend to support the hypothesis.2’ These studies also indicate that the oft-quoted, simple classification of OA into primary and secondary forms is both inappropriate and inadequate. OA: PROCESS AND OUTCOME

Our definition of OA states that it represents a dynamic process, with a mixture of features, and a variable outcome. The important work of Byers et al has further emphasized that both proPeripheral





Mean age 65 yrs

158 Men, Mean age 60 yrs 2) No. of Joints


1 joint





of Symptomatic




1 joint site



2 joint sites


M :z F

3 joint sites



4 joint 3) Commonest


M .- F










M :> F






F :> M



M :> F


Men Age (years) One or Two Joint

Sites Involved

Three or More Sites Involved * Older versus younger patlents, ReprInted with permisslon


Age and Number Joint



all groups P <


from Cushnaghan and Dleppe




(P < 0.001)

evolutionary explanation, the vulnerable joints suffering from “under-design” problems. Alexander2’ suggests that “underutilization” is to blame. These and other authors stress mechanical factors to explain the localization of OA. How is this equated with the strong evidence of an agerelated systemic disease emerging from the OA 500 and many other studies? Our own hypothesis is shown in Figure 2. It suggests that a systemic predisposition is present in a proportion of the population, and that biomechanical factors then dictate the site and severity of OA that subsequently develops. We can take this further by suggesting that the systemic factors are more important in knee and hand disease than they are in hip OA. Doherty’s work on postmenisectomy 1:


in Men


Why should OA huve such a singular predilection .ftir [email protected] joints at diflerent ages in men and women? Hutton2’ suggests that there is an































An hypothesis on the interaction of biochemical and biological factors in the aetiopathogenesis

of osteoarthritis.

gressive and nonprogressive forms of OA exist 24,25and the factors that initiate OA may be dikerent from those that control progression.26 These concepts of OA as a dynamic set of processes leading to various outcomes, or health states, is represented diagrammatically in Figure 3. The remainder of this article is largely concerned with our efforts to assess both process and outcome in OA. Assessing and Measuring the Disease Process

In theory, the OA process must involve a variable degree of degradation, inflammation, and repair, and should involve all the major tissues of the joint, including cartilage, bone, synovium, and capsule. The articular cartilage is generally regarded as the main target tissue of the response, although the changes in subchondral bone are just as striking, and the synovium is generally abnorma1.27,28 In the majority of cases, the clinician cannot examine the pathology of the joint. He/she can only examine it clinically, take blood, and obtain a roentgenogram. The radiograph is

but a historical record of past anatomical changes and conventional blood tests are all normal. Synovial fluid can sometimes be aspirated, but arthroscopy or surgery is only undertaken in a tiny minority of advanced cases. We need better methods of assessing intraarticular pathophysiology. Two types of approach are currently being investigated-different ways of imaging joints, and new biochemical assays of products of the OA process. New images of OA. Imaging the OA joint can provide two types of information: anatomical (roentgenogram, computed tomography [CT], arthrography, ultrasound magnetic resonance imaging [MRI]) or physiological (MRI, scintigraphy, thermography). We have been investigating the role of the plain radiograph, scintigraphy and, MRI. The plane radiograph remains the single most important investigation in the diagnosis and assessment of OA. However, the use of composite scores, such as the Kellgren and Lawrence 0 to 4 scale,i2 presents several problems, including


THE OUTCOME Discordance Between:-

A Variable Balance of:Inflammation IN

Degradation Repair




Cartilage L

Fig 3:



Process and outcome in osteoarthritis.

Anatomical Changes Symptoms Functional Changes








Between of the

24 Paleopathological




Bony Changes

Scintigraphy in OA has been investigated extensively. For example, some years ago Daniellson and colleagues investigated bone scans in hip OA3’ and Thomas et al documented the sensitivity of the technique at the knee joint.32 Charles Hutton and colleagues investigated the meaning of bone scans in OA of the hand in patients from our case series in Bristol. They found that the scan provided different information from roentgenograms, and that scintigraphy predicted radiographic change in a longitudinal study (Table we and others have been 4). 33.34Subsequently, assessing the role of scintigraphy in knee joint OA. Several different patterns of change have been detected, and it would appear that different types of synovial and bony change can be imaged (Table 4).35 The predictive value of these findings is currently being explored. MRI is the one imaging technique that provides both anatomic and physiological information. Its potential in arthritis is well known.36 McAlindon et al have recently completed an interesting comparative study of MRI, scintigraphy, and roentgenograms in OA of the knee joint.37 MRI was the most sensitive technique for detecting effusions and soft tissue changes. It also showed two abnormalities of bone not seen by radiograph that correlated with scintigraphic ab-


of OA in

Knee Specimens Visual






Osteophytosis Present

~eig28~~~ ; ; Reprinted with permission from Rogers, Watt, and Dieppe.

having to lump several features together. We prefer to assess the major features individually, on the presumption that the different bony and soft tissue changes and the interbone distance might not go together. We have investigated our own reproducibility in scoring these features at the knee joint.29 Juliet Rogers et al have also examined the sensitivity of the radiograph to the bony changes of OA by comparing visual and roentgenographic evidence of changes in paleopathological samples. 3oNot surprisingly, the radiograph emerges as a very insensitive assessment of focal osteophytosis or bone sclerosis (Table 3).





in OA of the Hand

and Knee % Showing Radiographic Changes

Total No.




No Change
























X-Ray Positive Scan Negative X-Ray and Scan Posmve X-Ray Negatwe Scan Positive Scan Posmve

(8) Early “Blood Pool” Phase Late Bone Phase

1. Generalized 2.





Joint line




Hot Patella

(A) Relatron between radiographic and scintrgraphrc frndings at entry, and radiographrc progression over 5 years In 14 patients with hand OA (Data from Hutton et al.) (6) Drfferent patterns of scrntigraphrc uptake seen in patients with knee OA. (Data from McCrae et al.)



normalities: a “bright” osteophyte correlating with isotope uptake at the edge of the joint, and changes in the water content of subchondral bone correlating with extended isotope uptake in the same region. It would appear from these studies that physiological changes in the joint, indicative of various aspects of the OA process, can now be imaged. As the anatomical, symptomatic, and functional consequences of the process are relatively slow to change, it will take many years to assess the true meaning and prognostic value of these findings. However, it seems likely that they provide information of predictive value. Biochemical “markers” for OA. The list of biochemical assays of joint products measurable in the synovial fluid, blood or urine is growing rapidly. Table 5, which documents some of them, is bound to be out of date by the time this article is published. The idea is simple enough-a product, ideally recognized by a distinctive epitope, is shed into the synovial fluid as a direct result of one aspect of the OA process, be it degradation or repair. Measurement of this product might then be a “marker” of the process, and provide a test for either the presence or activity of the condition. The tendency for OA to be in more than one joint reduces the power of a blood or urine assay, as do potential problems with passage and changes in epitopes moving from the joint to the circulation. There are also bound to be problems with concentrations in synovial fluid assays. However, the idea has attracted considerable attention in many places.38a The Bristol group is working in collaboration with many of these laboratories. One of the first products to be investigated was with Thonar keratan sulphate. 4’ In conjunction et al and Campion et al, we have explored the relationship of serum and synovial fluid levels of keratan sulfate (KS) to clinical, radiographic, and scintigraphic findings in 125 subjects with knee OA (35 men and 90 women; mean age, 65 yrs). Serum KS levels were higher in men than women (456 & 135 v 368 + 74) and showed a relation with the number of involved joints in women.42 Synovial fluid levels were much higher than, and unrelated to, serum levels. There was a significant negative correlation between synovial fluid KS and joint space narrowing.







of the OA Process 1. Hyaluronan 2.



Total glycosaminoglycans, keratan sulphate, proteoglycan

3. Collagen Products


Collagen pro-peptides. cross-linking compounds,


fragments 4.

eg, “COMP”

Noncollagenous cartilage proteins proteins

eg, osteocalcin




Synovial Products

eg, cytokines,



eg, calcium





These sort of data are intriguing, although they do not suggest that measurement of a product like KS alone is going to provide very useful data for day-to-day clinical use. We and others are looking at many other molecules, some of which show positive, rather than negative associations with radiographic damage. We have also been relating levels of cytokines in the synovial fluid43 of enzyme-inhibitor with the presence complexes44 and potential marker molecules. The aim of our investigations is to see whether it is possible to detect a variety of different activities of joints within the heterogeneous OA spectrum. As it becomes possible to measure both synthetic and degradative products, we may be able to differentiate between joints with excess degradative activity and those with excess synthesis. A combination of imaging and biochemical studies may provide a comprehensive picture of the current state of activity of the OA process.

Assessing and Measuring the Outcome The true value of new imaging and biochemical techniques in the assessment of OA will only become apparent if correlated with longitudinal clinical studies with effective and relevant measures of changing health status or outcome. As outlined above, one of the major problems with OA is the discordance between anatomical changes (as shown by radiographic evidence), symptoms and functional consequences. This is good news for the victim, who may be free of




pain and disability in spite of severe joint damage, but it means that the investigator must regard each of these aspects of outcome as potentially unrelated variables. Satisfactory outcome measures for OA have yet to be developed, although, with the possible exception of death, the “5 Ds” of Fries et al [death, disability, discomfort, dollar costs, drug side-reactions)45 would seem to be appropriate. Very little is known about the natural history of 0A.4” Most data are anecdotal, perhaps because of the relatively slow rate of change that occurs in most patients, making formal prospective studies a daunting undertaking. However, it is apparent that the textbook descriptions of OA as a slowly progressive disorder are inappropriate. Cases of rapid progression and spontaneous healing are well described and not rare.46-49 We recently reported a small, prospective, g-year study of outcome in OA of the knee.” Most patients showed a deterioration in function during that time, and pain worsened in many. However, 2 of 32 patients showed a marked improvement in function and disability. Many radiographs did not change over 8 years, and clinical changes showed no apparent relation to radiographic changes (Table 6). We have also analyzed the changes seen in a small group of patients with erosive interphalangeal joint OA, studied over 3 years5’ The erosive changes often healed, and there were no apparent demographic differences of this group compared with other “control” patients with hand OA. This suggests that “erosive OA” merely reflects a stage of severe disease activity, and is not a different disorder. The findings from these prospective studies run contrary to some of our current concepts of OA. As most scientific investigations are based on clinical perceptions of the disease problem (progressive cartilage damage for example), this presents problems. More clinical studies, particularly addressing the natural history of the condition and the determinants of outcome, are urgently needed. RELATING






Table 6: Relation Between Changes in Symptoms and Radiographs in 26 Patients With Knee OA Studied Over 8 Years

11 Data from Massardo

et al.

poor correlation between symptoms and radiographs, and the relatively slow rate of change seen in many patients. In my opinion, these problems make it difficult to study the condition in its earliest, most interesting phase, and make investigations of therapy for anything other than symptoms almost impossible. The clinical studies mentioned in this article have led us to postulate two “new” approaches to some of these problems. Haygarth would have found nothing novel in the two suggestions.

1. How to Study Early OA Haygarth knew about the slow addition of new joint sites that we have also observed. In particular, an obese woman presenting with OA of one knee, or of the hands, will almost certainly develop OA of both knees within a few years.” This immediately provides a strategy for the investigation of early OA: study the asymptomatic, “uninvolved” knee joint(s) in this high risk group. 2. How to Detect Drug Effects in OA The slow addition of new sites is again the clue. If a drug was effective in slowing the development or progression of OA, then it should reduce the rate at which new sites of radiographic change appear in someone with the condition. The strategy is simply to count the number of new sites with definite radiographic changes in a treated and a control group of patients.



Four specific problems have contributed to the paucity of clinical investigations in OA: heterogeneity, the inability to detect early disease, the


Since the all important differentiation of arthropathies into atrophic and hypertrophic types, 4,5’ the less common atrophic group has attracted the most clinical and laboratory atten-



tion.52 We are now witnessing a rapid increase of interest and investigation into hypertrophic arthritis, or OA. This has centered on the belief that progressive cartilage degeneration is the key feature of the condition. Clinical studies do not confirm this belief.


I accepted the Roussel Prize for Clinical Research in Osteoarthritis on behalf of the Bristol University Rheumatology Unit; all those listed,* and many others, have contributed to this work, and I would like to take this opportunity to thank them.

REFERENCES 1. Haygarth J: A Clinical History of the Joints, 1805

of Diseases: 2. Nodosity

2. Adams R: A Treatise on Rheumatic Gout or Chronic Rheumatic Arthritis of all the Joints. London, Churchill, 1857 3. Dieppe drocalcinosis.

P, Doherty M: The first descriptions of chonArthritis Rheum 32: 1339- 1340, 1989

4. McCrae T: The pathology formans. JAMA 43:1027-1032,

and aetiology 1904

of arthritis


5. Kellgren JH: Observations on referred pain arising from muscle. Clin Sci 3: l75- 190, I938 6. Kellgren 949, 1948

JH: Deep pain and sensibility.


7. Kellgren JH, Moore R: Generalised osteoarthritis Heberden’s nodes. Br Med J 1:181-187, 1952 8. Kellgren JH: Primary 1954


Bull Rheum

1:943and Dis 4:

9. Cushnaghan J, Dieppe P: A Study of 500 patients with limb joint osteoarthritis: 1. Analysis by age, sex and distribution of symptomatic joint sites. Ann Rheum Dis (in press), 1990 10. Altman RD, Bloch DA, Bole GG, et al: Development of clinical criteria for osteoarthritis. J Rheumatol 14:3-6, 1987 (SUPPI 14) 11. McAlindon T, Dieppe P: Osteoarthritis: criteria. Ann Rheum Dis 48:531-532, 1989



12. Kellgren JH, Lawrence JS: Radiological assessment of osteoarthritis. Ann Rheum Dis 16:494-50 I, 1957 13. Lawrence JS, Bremner JM, Bier F: Osteoarthritis: prevalence in the population and relationship between symptoms and X-ray changes. Ann Rheum Dis 25:1-24, 1966 14. Cobb S, Merchant WR, Ruben T: The relation of symptoms to osteoarthritis. J Chron Dis 5: 197-204, 1957 15. Wright V: Osteoarthritis: Diagnosis is not as simple as it seems, Br Med J 299: 1476-1477, 1989 16. Nassonova VA, Podchalimova VV, Brzhezovsky MM, et al: Osteoarthritis diagnostic criteria for epidemiological studies. “R” XVI No83-2:85-88, 17. Altman RD, Block DA, Brandt KD, et al: Osteoarthritis: Definitions and criteria (letter). Ann Rheum Dis 49: 201, 1990 18. Dieppc P, Cushnaghan J: Prospective study of outcome in Osteoarthritis (OA). Arthritis Rheum 31:574, 1988 19. Massardo L, Watt 1, Cushnaghan J, et al: An eight year prospective study of osteoarthritis of the knee joint. Ann Rheum Dis 48:893-897, 1989 20. Felson PT: Epidemiology of hip and knee osteoarthritis. Epidemiol Rev lO:l-28, 1988 2 1. Hutton CW: Generalised osteoarthritis: An evolutionary problem? Lancet 1:1463-1465, 1987

22. Alexander CJ: Relationship between the utilisation profile of individual joints and their susceptibility to primary osteoarthritis. Skeletal Radio] 18: 199-205, 1989 23. Doherty M, Watt I, Dieppe P: Intluence of primary generalised osteoarthritis on development of secondary osteoarthritis. Lancet 2:8- 11, 1983 24. Byers PD, Contepomi CA, Farkas TA: A postmortem study of the hip joint. Ann Rheum Dis 29:31, 1970 25. Youngman HR, Cooper F, Byers P: Prevalence of cartilage lesions in foot joints: a test of the concept of limited and progressive lesions. Ann Rheum Dis 46:5 15-5 19, 1957 26. Raden E: Osteoarthritis: What is known about prevention. Clin Orthop Rel Res 222:60-62, 1987 27. Meacham G, Brooke G: The pathology of osteoarthritis, in Moskowitz R, Howell D, Goldberg V, et al (eds): Osteoarthritis: Diagnosis and Management. Philadelphia, PA, Saunders, 1984, pp 29-42 28. Dieppe P, Harkness AL, H&s ER: Osteoarthritis, in Wall PD, Melzack R (eds): Textbook of Pair (ed 2). London, Churchill Livingstone, 1989, pp 306-3 16 29. Cooper C, Cushnaghan J, Kirwan J, et al: Radiographic assessment of the knee joint in osteoarthritis. Br J Rheumatol 29:Al9, 1990 (suppl I) 30. Rogers J, Watt I, Dieppe P: Comparison of visual and radiographic detection of bony changes at the knee joint. Br Med J 300:367-368, 1990 3 I. Danielsson LG, Dymling JF, Heripret G: Coxarthrosis in man studied with external counting of 85Sr and 47Ca. Chn Orthop 31:184-199, 1963 32. Thomas RH, Resnick D, Alazraki NP, et al: Compartmental evolution of osteoarthritis of the knee. A comparative study of available diagnostic modalities. Radiology 116:858-894, 1975 33. Hutton CW, H&s ER, Jackson PC, et al: 99mTc HMDP bone scanning in general&d osteoarthritis. 1. Comparison of the standard radiograph and four hour bone scan image of the hand. Ann Rheum Dis 45:6 17-62 1, 1986 34. Hutton CW, Higgs ER, Jackson PC, et al: 99mTc HMDP bone scanning in generalised osteoarthritis. 2. The four hour bone scan image predicts radiographic change. Ann Rheum Dis 45:622-626, 1986 35. McCrae F, Shouls J, Dieppe P, et al: The heterogeneity of osteoarthritis of the knee. Br J Rheumatol 26:A45, 1987 (SUPPI I) 36. Ehman RL, Bergust TH, McLeod RA: MR Imaging of the musculoskeletal system: A five year appraisal. Radiology 166:313-320, 1988 37. McAlindon TE, McCrae F, Watt I, et al: Magnetic resonance imaging in osteoarthritis of the knee: Correlation with Clinical, radiographic and scintigraphic appearances. Br J Rheumatol (in press), 1990





38. Hascall V, Clam T: Proteoglycan epitopes as potential markers

of normal

Arthritis Rheum

and pathological

39. Thompson

pyridinium. involvcmcnt


30:568-589 P: Laboratory

MJ. Duncan

markers ofjoint




A. Robins SP: IJrinary


diseases. Br J Rheumatol


070. IYSY 4



Thonar E. Schnitzer TJ. Kuettner K: Kerdtan Sulphate 14:

IYXY (suppI 14)

42. Campion

Ci. McCrae

teoarthritis? Br J Rheumatol

F. Schnitzer T. et al: Do serum of os-

27:A IS, 1988 (suppl 2)

43. Westacott Cl. Whither JT. Thompson D, ct al: Synovial fluid concentration

of five drtTcrcnt cytokines

discascs. Rr J Rheumatol

in rheumatic

of polymorphonuclear

elastasc - 0,

proteinase inhibitor (elastase - U, antitrypsin) in patients with rheumatoid

arthritis: interference hy rheumatoid

Rheum IXs 4Y:IX-21.


R. Howell D. Goldberg V. et al (eds):

Diagnosis and Management

(cd 2). Philadel-

phia. PA. 1990 47.

Perry GH,

Smith MJG.


recovery of the joint space in degenerative Dis 31:440-44X.

CC;: Spontaneous hip disease. Ann


48. Storey GO, Landello JW: Restoration of femoral head after collapse in osteoarthritis,



Ann Rheum Dis 30:406-412.

Doherty M. Holt M. MacMillan

of analgesic hip. Ann Rheum 50. Cushnaghan

factor. Ann

J. Cobby

osteoarthritis: a longitudinal Rr J Rheumatol

27 (suppl 2) A67

Banks RE. Evans SW. Taylor KF. ct al: Measurement

of plasma concentration

of pattent outcome

Rheum 26:697-704,


Ic\cls in keratan sulphate help with the heterogeneity


arthritis. in Moskowitz


rn blood as a marker of cartilage catabolism. J Rheumatol


Dieppe P. Cushnaghan J: The natural history of osteo-


Crosslinks provide indices of cartilage and bone in arthritic

45. Fnes JF: Towards an undemanding measurement. 46.

and damage. Rr J Rheumatol 40. Sietxl


M. Creamer comparative


f’. et al: firosive study of 24 cases.

28: I02 (suppl 2). 1989 (ahstr)

5 I. Goldthwaitc

JE: The treatment

sulting from the so-called rheumatic Surg J 136:79-X4. 52. Dieppe

P. et al: A reappraisal

Dis 45:272-276.

of disabled joints rcdiseases. Boston Med


P: Osteoarthritis:

problem. in Russell RGG.

The size and scope of the

Dicppc P (eds): Osteoarthritis