Annals of the Rheumatic Diseases 1997;56:85–90 85 Ann Rheum Dis: first published as 10.1136/ard.56.2.85 on 1 February 1997. Downloaded from EXTENDED REPORTS

Bone formers: osteophyte and enthesophyte formation are positively associated

Juliet Rogers, Lee Shepstone, Paul Dieppe

Abstract phenomenon, unrelated to any joint disease.4 Objective—To test the hypothesis that New bone can form at individual entheses in enthesophyte formation and osteophyte response to a seronegative spondarthritis.5 growth are positively associated and to More commonly, they are seen in several sites look for associations between bone forma- as part of the condition first described in the tion at diVerent sites on the skeleton so spine by Forrestier and Rotes-Querol6 and now that a simple measure of bone formation known as diVuse idiopathic skeletal hyperost- could be derived. osis (DISH).7 Methods—Visual examination of 337 adult The presence of periarticular osteophytes skeletons. All common sites of either has been noted by Resnick and Niwayama in enthesophyte or osteophyte formation DISH1 but the relation of enthesophyte and were inspected by a single observer who marginal osteophytosis in this condition has graded bone formation at these sites on a not been specifically investigated. This study 0-3 scale. The total score for each feature tests the hypothesis that some individuals have was divided by the number of sites exam- a greater tendency to form bone at both joint ined to derive an enthesophyte and an margins and entheses than others. The osteophyte score. Cronbach’s á and hypothesis has been derived from the observa- principal components analysis were used tion in skeletal studies of striking osteophyte to identify groupings. formation in a subgroup of skeletons, including Results—Enthesophyte formation was those with DISH, and from the reports quoted associated with gender (M>F) and age. above. The latter suggests that a variable

There was a positive correlation between amount of bone formation occurs at these two http://ard.bmj.com/ enthesophytes and osteophytes (r = 0.65, sites as a result of a mixture of age, a systemic 95% confidence interval, 0.58 to 0.71) predisposition, and local biomechanical which remained after correction for age factors. and gender. Principal components analy- Radiographs provide an insensitive and sis indicated four diVerent groupings of inadequate way of assessing osteophyte forma- enthesophyte formation. By choosing one tion and enthesophyte changes.8 In contrast, site from each group a simple index of the visual examination of skeletons allows all total skeletal bone formation could be aspects of the joint margin and several different on September 26, 2021 by guest. Protected copyright. derived. ligament and tendon insertion sites to be Conclusions—Osteophytes and entheso- examined in detail, and graded for the amount phytes are associated, such that a propor- of bony change. We have therefore system- tion of the population can be classified as atically examined a number of skeletons for “bone formers”. Enthesophyte groupings evidence of bony changes at several joint mar- provide some clues to aetiopathogenesis. gins and enthesophyte sites, and examined the Bone formation should be investigated as data for associations. a possible determinant of the heterogene- The unique opportunity of visual examina- ity of outcome and of treatment responses tion of whole skeletons also allowed us to in common musculoskeletal disorders. document enough enthesis sites to investigate the data for possible groupings, which might Rheumatology Unit, (Ann Rheum Dis 1997;56:85–90) Department of provide insights into the aetiopathogenesis of Medicine, University Musculoskeletal disorders are associated with enthesopathy as well as developing a method of of Bristol, Bristol the formation of new bone at two main sites: quantifying the phenomenon and thus Royal Infirmary, the joint margin (osteophytosis) and ligament Bristol BS2 8HW, allowing us to derive a simple measure of United Kingdom and tendon insertions (enthesophyte forma- “bone formation”. The development of a J Rogers tion).12 Osteophytes are strongly associated working definition is necessary before any L Shepstone with osteoarthritis, probably forming in clinical implications can be investigated. P Dieppe response to abnormal stresses on the joint 3 Correspondence to: margin. However, the degree of osteophytosis Methods Dr Juliet Rogers. in osteoarthritis varies considerably, and there Three hundred and thirty seven adult skeletons Accepted for publication is some evidence that small marginal from routine studies of the skeletons from 24 October 1996 osteophytes can also develop as an age related three separate archaeological sites were 86 Rogers, Shepstone, Dieppe

Table 1 Osteophyte and enthesophyte locations all skeletal elements surviving, including the Ann Rheum Dis: first published as 10.1136/ard.56.2.85 on 1 February 1997. Downloaded from spine and the main long bones. Missing bones Osteophytes: spinal and peripheral joint sites Enthesophytes: ligament insertion sites were assumed to be missing at random. There Odontoid Distal interphalangeal joints Rotator cuV were 202 male skeletons, 129 female, and six Cervical facet joints Hip Olecranon—ulna were unsexed. The age ranges were young Thoracic facet joints Knee: medial compartment Deltoid tubercle—humerus Lumbar facet joints Knee: lateral compartment Bicipital tuberosity—radius adult (20-25 years) to old adult (more than 60 Temporomandibular joint Knee: patellofemoral joint Greater trochanter—femur years), as aged by standard anthropological Acromioclavicular joint Ankle Lesser trochanter—femur 910 Sternoclavicular joint Tarsal joints Linear aspera—femur techniques. The skeletons were dated from Glenohumeral joint Metatarsophalangeal joints Tibial tubercle between the 9th century and the 16th century; Elbow Interphalangeal joints Soleal line—tibia 91 were from excavations at Wells Cathedral, Wrist Quadriceps insertion—patella Thumb base Iliac crest—pelvis 165 from St Oswald’s priory in Gloucester, and Other carpal metacarpal Ischial tuberosity—pelvis 81 from St Peter’s Church, Barton-on- joints Posterior spur—calcaneum Humber. All available skeletons from the sites Metacarpophalangeal joints Inferior spur—calcaneum Proximal interphalangeal at Wells and St Oswalds were used, and a fur- joints ther group from Barton-on-Humber was added to increase the power of the study. The selected for entry into this project. All the data skeletons were observed for osteophyte around were collected according to our standard the margins of vertebral bodies and at 23 protocols and were acquired before deciding to peripheral joint sites, and enthesophyte at 14 use them to examine the hypothesis addressed ligament insertion sites (table 1). The joint in this paper. Each skeleton had at least half of sites were selected to include most synovial http://ard.bmj.com/ on September 26, 2021 by guest. Protected copyright.

Figure 1 (A) Knees showing marked osteophytes grade III and enthesophytes at the Figure 1 (A) Knees showing marked tibial tubercle grade II. (B) Grade III osteophyte osteophytes grade III and enthesophytes at the around the carpo-metacarpal joint of the thumb. tibial tubercle grade II. (B) Grade III (C) Calcaneum with grade II enthesophyte and osteophyte around the carpo-metacarpal joint of patellae with grade III enthesophyte. (D) Radius the thumb. (C) Calcaneum with grade II with grade II enthesophyte at occipital enthesophyte and patellae with grade III protuberance and grade I osteophyte at enthesophyte. (D) Radius with grade II radioulnar articulation. Ulna humeral joint has enthesophyte at occipital protuberance and grade grade II osteophyte. (E) Hip joint showing I osteophyte at radioulnar articulation. Ulna greater trochanter of femur with grade II humeral joint has grade II osteophyte. (E) Hip enthesophyte and lesser trochanter with grade I joint showing greater trochanter of femur with enthesophyte. The iliac crest also has grade II enthesophyte and lesser trochanter with enthesophytes, grade I, and the ischial tuberosity grade I enthesophyte. The iliac crest also has enthesophytes grade II. enthesophytes, grade I, and the ischial tuberosity enthesophytes grade II. Correlation of osteophyte and enthesophyte formation 87

articulations around the skeleton. The between sexes). A general linear model was Ann Rheum Dis: first published as 10.1136/ard.56.2.85 on 1 February 1997. Downloaded from odontoid peg was observed separately from the used to assess the strength of the relations rest of the cervical facet joints. In the spine the between enthesophyte scores and osteophyte presence of osteophyte at any cervical, scores while allowing for other potential thoracic, or lumbar facet joint was deemed to confounding variables. A square root transfor- be positive for that segment. The three mation was taken of the enthesophyte score compartments of the knee joint were observed (the dependent variable) in order to achieve as individual joints and the thumb base was normally distributed residuals. Statistical treated separately from the rest of the significance was set at the 5% level. carpal-metacarpal joints. The presence of An attempt was made to find a small number osteophyte in any metacarpal-phalangeal of enthesophyte sites that could be used to (MCP) joint, a proximal interphalangeal joint- construct an enthesophyte score with little loss (PIP), or a distal interphalangeal joint counted of information (referred to as the reduced as positive for that site. A similar regime was score). Cronbach’s á (a form of intraclass cor- adopted for the foot. We selected a wide distri- relation coeYcient) was used to identify sites bution of ligament insertion sites around the that correlated poorly with others (that is, skeleton, mainly those noted by Resnick to those sites which resulted in an increase in produce enthesopathy in DISH. They were Cronbach’s á when omitted) in order to make also sites where new bone formation at the an initial reduction in the number of sites used enthesis was unequivocal. The osteophytes and for scoring. Principal components analysis was enthesophytes were graded on a scale of 0-3 used to identify possible “groupings” of (none, mild, moderate, severe) (fig 1). The four enthesophyte sites with the intention of select- point grading scales were adopted as being a ing sites to represent each grouping. usual way of grading osteophytes, for example, We have also considered diVerent definitions on x rays11 and have been adopted in of a “bone former” based on a dichotomy of palaeopathological studies so as to be as the reduced score (that is, defining individuals compatible with clinical work as possible. The as Bone formers or not, depending on whether presence or absence of DISH was also their reduced score is above or below a recorded according to criteria defined by particular cut oV point). We have attempted to Resnick.12 validate this by comparing our bone former definition with DISH as defined by Resnick 12 STATISTICAL METHODS and Niwayama. An osteophyte score for each skeleton was obtained by adding all scores together and Results dividing by the number of joint sites for which Osteophyte and enthesophyte scores were observations could be made. Likewise entheso- calculable in all 337 skeletons. Both scores phyte scores were obtained by adding them showed a positively skewed distribution (fig 2 together and dividing by the number of and 3). The median osteophyte score was

ligament insertion sites that were present for 0.087, (interquartile range, 0 to 0.318). A total http://ard.bmj.com/ that individual. Thus the maximum osteophyte of 130 skeletons had a score of zero. The and enthesophyte scores for an individual was median enthesophyte score was 0.071, 3 for osteophyte and 3 for enthesophyte. (interquartile range, 0 to 0.214). A total of 168 Both osteophyte and enthesophyte scores skeletons had a score of zero. have been summarised with median scores and There was a strong positive correlation interquartile ranges. Spearman’s rank correla- between osteophyte score and enthesophyte tion was used to assess the strength of relations score (Spearman’s rank correlation, r = 0.647, between continuous measures. Mann-Whitney 95% confidence limits = 0.579 and 0.706). on September 26, 2021 by guest. Protected copyright. tests were used to test for a diVerence in The median enthesophyte score for males median scores between groups (for example, (0.080, interquartile range, 0 to 0.357) was

250

200

150

100 Frequency

50

0 < 0.2 0.4–0.6 0.8–1.0 1.4–1.6 1.8–2.0 2.2–2.4 2.6–2.8 0.2–0.4 0.6–0.8 1.2–1.4 1.6–1.8 2.0–2.2 2.4–2.6 2.8–3.0 Score Figure 2 Histogram of the frequencies of skeletons in each category of osteophyte score. Each category has a range of 0.2 units. A positive skew is clearly evident 88 Rogers, Shepstone, Dieppe

250 Ann Rheum Dis: first published as 10.1136/ard.56.2.85 on 1 February 1997. Downloaded from

200

150

100 Frequency

50

0 < 0.2 0.4–0.6 0.8–1.0 1.4–1.6 1.8–2.0 2.2–2.4 2.6–2.8 0.2–0.4 0.6–0.8 1.2–1.4 1.6–1.8 2.0–2.2 2.4–2.6 2.8–3.0 Score Figure 3 Histogram of the frequencies of skeletons in each category of enthesophyte score. Each category has a range of 0.2 units. A positive skew is clearly evident.

significantly diVerent (P < 0.001, Mann- 0.051, partial correlation = 0.534, 95% Whitney test) from the median score for confidence limits, 0.450 and 0.609). females (0, interquartile range, 0 to 0.100). The presence of DISH was identified in 28 Those in the older age group (> 45 years) had individuals. The median enthesophyte score a higher median enthesophyte score (0.214, for these individuals was 0.667 (interquartile interquartile range, 0.071 to 0.538) than those range, 0.432 to 0.893) compared to a median in the younger age group (0, interquartile score of zero (interquartile range 0 to 0.167) in range, 0 to 0.083). This was also statistically those without DISH. This diVerence was significant (P < 0.001, Mann-Whitney test). statistically significant (P < 0.001, Mann- A general linear model using enthesophyte Whitney test). The median osteophyte score score as the dependent variable was for those with DISH was 0.711 (interquartile constructed to test for a significant relation range 0.457 to 0.929) compared to a median with osteophyte score whilst allowing for both score of 0.067 (interquartile range 0 to 0.222) sex and age eVects. The parameter estimates for those without. This diVerence was also for this model indicated a significantly higher statistically significant (P < 0.001, Mann- enthesophyte score for males than for females Whitney test). (parameter estimate, males-females, = 0.079, The overall value of Cronbach’s á (0.923) SE = 0.026, P = 0.003) and also higher for the indicated a high degree of correlation between http://ard.bmj.com/ older age group (parameter estimate = 0.132, the enthesophyte scores at diVerent ligament SE = 0.029, P < 0.001). In the presence of insertion sites. The omission of any single site these eVects there was also a significant osteo- resulted in very little change in Cronbach’s á. phyte eVect (parameter estimate = 0.679, SE = The lowest value (0.910) resulted from the

0.6 Rotator cuff on September 26, 2021 by guest. Protected copyright. 0.5

0.4 Olecranon Deltoid tubercle Ischial tuberosity0.3

0.2

0.1 Linear aspera Third PC 0 Posterior spur –0.4 –0.3 –0.2 –0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Greater trochanter –0.1 Quadriceps insertion Biceptial tuberosity –0.2 Iliac crest Inferior spur Lesser trochanter –0.3 Soleal line Linear tubercle –0.4

Second PC Figure 4 The loadings from each of the ligament insertion sites for the second and third principal components. This plot has been used to informally identify four groups of sites. Correlation of osteophyte and enthesophyte formation 89

Table 2 Four groups of ligament insertion sites diVerences examined. The mean diVerence Ann Rheum Dis: first published as 10.1136/ard.56.2.85 on 1 February 1997. Downloaded from was 0.011 (standard error = 0.0090); the Group 1 Group 3 standard deviation was 0.160. Therefore, Posterior spur, calcaneum Greater trochanter (femur) Inferior spur of the calcaneum Lesser trochanter (femur) assuming these errors follow a normal Tubercle (tibia) distribution, approximately 90% of the Group 2 Group 4 reduced scores would lie within the range Soleal line (tibia) Rotator cuV (-0.252, 0.274) of the full score. Biceptial tuberosity (radius) Olecronon Possible definitions of a “bone former”were Iliac crest Deltoid tubercle (ulna) Quadriceps insertion (patella) Ischial tuberosity (pelvis) investigated using diVerent dichotomies of this four-site reduced score. The sample prevalence of Bone formers was calculated for each defini- omission of the iliac crest; the highest value tion together with percentage agreement (both (0.928) resulted from the omission of the rota- sensitivity and specificity) with the presence or tor cuV. Cronbach’s á was therefore of little use absence of DISH. These are shown graphically in the selection of a reduced number of sites. in fig 5. A definition based on a reduced score The principal components analysis resulted of greater than 0.30 agrees to a reasonable in three components which, combined, extent with the presence or absence of DISH accounted for more than 68% of total (sensitivity = 78.6%, specificity = 83.2%) and variation. The first component (accounting for would indicate a prevalence of Bone formers of 53% of the total variation) contained positive approximately 22%. coeYcients from all sites and can be interpreted as a “size” component. The second Discussion (accounting for 8% of total variation) and the The visual examination of skeletons provides a third (accounting for 7% of total variation) unique opportunity to describe bony changes represent diVerent modes of variation and were related to age and disease. They allow all used to select four sites for consideration. aspects of bony surfaces to be assessed, free of These two components plotted against each soft tissues. The commonest types of bony other are shown in fig 4. Four groupings were change seen in adult skeletons include identified: (1) the posterior and inferior spurs osteophytes and enthesophyte formation.13 of the calcaneum; (2) the soleal line of the tibia, Osteophytes can be defined as lateral the bicipital tuberosity of the radius, the iliac outgrowths of bone at the margin of the articu- crest, and the quadriceps insertion of the lar surface of a synovial joint. An enthesophyte patella; (3) greater and lesser trochanters of the is a bony spur forming at a ligament or tendon femur and the tubercle of the tibia; (4) the insertion into bone, growing in the direction of rotator cuV, olecranon process of the ulna, del- the natural pull of the ligament or tendon toid tubercule of the humerus, and ischial involved. tuberosity of the pelvis. These groups are sum- Both osteophyte and enthesophyte can be marised in table 2. One ligament insertion site regarded as skeletal responses to stress. Osteo-

was arbitrarily chosen from each group. These phytes can occur as a part of the aging process http://ard.bmj.com/ four sites were the posterior spur of the but are more commonly associated with calcaneum, the bicipital tuberosity of the osteoarthritis. There is experimental evidence radius, the greater trochanter of the femur, and that osteophyte formation is related to instabil- the ischial tuberosity (see fig 1C-1E). ity of joints and their growth has been Based on these four sites, a reduced score described as part of the attempt of a synovial (the average score of these four sites) was joint to adapt to injury, limiting excess calculated. In one individual all four sites were movement and helping to recreate a viable missing and hence a score could not be calcu- joint surface.14 Enthesophytes can form in on September 26, 2021 by guest. Protected copyright. lated. In the remaining 336 the reduced score response to the inflammation of the enthesis was compared with the full score and the occurring in seronegative spondarthropathies and in response to repetitive strain, as in the 15 100 spiking of tibial spines seen in footballers. Enthesophyte formation also occurs in the 90 absence of any clear cause. Multiple idiopathic 80 enthesophytes are characteristic of diVuse Sensitivity idiopathic skeletal hyperostosis (DISH).7 70 Specificity The hypothesis that there may be a relation Prevalence 60 between marginal osteophytes and entheso- 1 50 phytes has been alluded to by Resnick and was suggested to us by the observation of skeletons Percent 40 with excessive bone formation of both types, 30 with or without DISH. However, this is the first systematic study to examine this 20 hypothesis. A biological rationale might be that 10 the response of the skeleton to stress depends on common cellular mechanisms, which site is 0 aVected, and the genetic control of the

0.01 0.06 0.11 0.16 0.21 0.26 0.31 0.36 0.41 0.46 0.51 0.56 0.61 0.66 0.71 0.76 0.81 0.86 0.91 0.96 1.01 1.06 1.11 1.16 1.21 1.26 1.31 1.36 1.41 1.46 reaction. Cutt off point To address this hypothesis, multiple sites Figure 5 Characteristics of diVerent “bone former” definitions—agreement with prone to either osteophyte or enthesophyte occurrence of DISH and prevalence. formation were examined in a large number of 90 Rogers, Shepstone, Dieppe

skeletons. The data were collected by a single Bone formation is one of the major Ann Rheum Dis: first published as 10.1136/ard.56.2.85 on 1 February 1997. Downloaded from experienced observer, able to grade the severity components of the response of the musculo- of bone formation at diVerent sites, as well as skeletal system to stress and injury, as the presence or absence of change. It is evidenced by WolV’s law of bone remodelling. possible that due to a lack of blinding the This study suggests that the observed variation observation of one change may have influenced in bone formation could be due to diVerences the recording of the other. We think this is in individual ability to form bone in response unlikely to have been of importance, however, to stress rather than diVerences in stress. This because the data were collected before the suggests a heterogeneity in one of the hypothesis developed and because the changes fundamental aspects of the pathogenesis of are so clear on skeletons. The data shows a musculoskeletal disorders which may be under similar distribution of each type of bony genetic control. Individuals who are good bone growth: these “Bone formers” forming a formers may have diVerent disease outcomes substantial minority of the skeletal population. to those of poor bone formers. Our findings As expected,2 enthesophyte formation was indicate that simple indices of bone formation associated with age and was more common in can be derived from skeletons and these may males than females. be generalisable to radiographic examinations. The major finding of the study was the The concept of bone formers should now be strong positive relation between enthesophytes examined in relation to the heterogeneity of and osteophytes. The data were analysed care- outcome and treatment responses in disorders fully to exclude the possibility of this being due such as osteoarthritis and osteoporosis. to age and sex associations alone. The validity of the association was reinforced by the finding 1 Resnick D, Niwayama G. Diagnosis of bone and joint disor- that skeletons with the most extensive entheso- ders. Philadelphia :WB Saunders, 1994. 2 Resnick D, Niwayama G. Enthesis and enthesopathy: phyte formation (that is, those with DISH) also anatomical, pathological and radiological correlation. had very high osteophyte scores. Radiology 1983;146:1–9. Enthesophytes, when present, were fre- 3 Moskowitz RW, Goldberg VM. Studies of osteophyte pathogenesis in experimentally induced osteoarthritis. J quently seen in many sites. The data were Rheumatol 1987;14:311–20. examined, therefore, to look for patterns or 4 Hernborg J, Nilsson N. The relationship between osteoar- thritis in the knee joint, osteophyte and aging. Acta Orthop groupings. Using Cronbach’s á, a strong asso- Scand 1973;44:69–74. ciation between all sites was suggested, but no 5 Niepal GA, Sitaj S. Enthesopathy. Clin Rheum Dis 1979;5:857–72. obvious groups emerged. However, a principal 6 Forrestier J, Rotes-Querol J. Senile ankylosing hyperostosis components analysis indicated a possible of the spine. Ann Rheum Dis 1950;9:321–30. 7 Resnick D, Shaul SR, Robins JM. DiVuse idiopathic skeletal division into four groups of sites which were hyperostosis (DISH). Forrestiers disease with extra spinal more strongly associated with each other. The manifestations. Radiology 1975;115:513–24. grouping, together of the two calcaneal sites in 8 Rogers J, Watt I, Dieppe P. Comparison of visual and radio- graphic detection of bony changes at the knee joint. BMJ one group and the two femoral trochanters 1990;300:367–8. with the tibial tubercle in another, suggests that 9 Brothwell D. Digging up bones. London: British Museum of Natural History, 1981.

local biomechanical factors may lie behind 10 Brooks S, Suchey J. Skeletal age determination based on the http://ard.bmj.com/ these groupings. However, the two other os pubis. J Hum Evolut 1990;53:227–38. 11 Kellgren JH. The epidemiology of chronic rheumatism. Vol groupings link upper and lower limb sites, per- 2. Atlas of standard radiographs of arthritis. Oxford: haps indicating that systemic factors such as Blackwell Scientific, 1963. 12 Resnick D, Niwayama G. Radiographic and pathologic fea- body habitus may also be important. In this tures of spinal involvement in diVuse idiopathic skeletal context DISH is known to be associated with hyperostosis (DISH). Radiology 1976;115:559–68. both diabetes and obesity.16 17 13 Rogers J, Waldron T. A field guide to joint disease in archae- ology. New York: John Wiley, 1995. By using one site from each grouping, a sim- 14 Bullough P, Vigorta V. Atlas of orthopaedic pathology. Lon- ple assessment of overall skeletal bone don: Gower, 1984. on September 26, 2021 by guest. Protected copyright. 15 Smillie IS. Injuries of the knee joint. Edinburgh: Churchill formation was derived, which was a reasonable Livingstone, 1970. approximation of the total enthesophyte score. 16 Julkunnen H, Heinonen OP, Pyörälä K. Hyperostosis of the spine in an adult population. Ann Rheum Dis 1971; This could allow skeletons to be assessed 30:605–12. quickly and easily. It also implies that a few 17 Daragon A, Mejjad O, Czernichow P, Louvel JP, Vittecoq O, Durr A, et al. Vertebral hyperostosis and diabetes simple radiographs could be used to provide an mellitus:a case-control study. Ann Rheum Dis 1971; assessment of bone formation in vivo. 30:605–12.