MICROSCOPY RESEARCH AND TECHNIQUE 37:358–371 (1997) Bone Density and Local Growth Factors in Generalized Osteoarthritis JAN DEQUEKER,* LUC MOKASSA, JEROEN AERSSENS, AND STEVEN BOONEN Arthritis and Metabolic Bone Disease Research Unit, K.U. Leuven, U.Z. Pellenberg, B-3212 Pellenberg, Belgium KEY WORDS osteoarthritis; bone density; growth factors; subchondral bone; osteoporosis ABSTRACT Osteoarthritis is usually considered to be a primary disorder of chondrocyte function with secondary changes in bones. However, a defect in the subchondral bone resulting in loss of its shock absorbing capacity could transfer the stress of loading directly to the articular cartilage with secondary changes in the cartilage. Review of histomorphometric and bone densitometric studies at sites of osteoarthritis at the hip or knee revealed that cartilage fibrillation could not be dissociated from bony changes even in the earliest stages of osteoarthritis and that subchondral trabeculae are thickened and more spaced in osteoarthritis. Microfractures of subchondral trabecular bone were less frequently seen in osteoarthritis compared to controls. Changes of the tidemark were found to be multiform and metabolically active in the osteoarthritic process. Endochondral ossification depletes the calcified cartilage at the cartilage/bone interface and the tidemark has been thought of as a calcification front advancing in the direction of non-calcified cartilage. Duplication of the tidemark is cited as evidence of this advancement. In the few experimental animal studies of subchondral bone in osteoarthritis, thicker trabeculae which were closer together were found in guinea pigs already when only mild cartilage changes were present. In the dog, with cruciate ligament transection, changes in bone were later than in the cartilage, but the changes in bone could still contribute to the progression of osteoarthritis. To study if bone changes may precede injury to the cartilage and if metabolic and systemic influences can also alter the subchondral bone, rendering it less able to withstand normal mechanical stresses, bone at different sites in the body has been studied extensively by the authors. Epidemiological and case control studies have revealed that osteoarthritis cases have more bone at all sites than expected and that bone in cases with generalized osteoarthritis shows both quantitative and qualitative differences, including increased contents of growth factors and hypermineralization. These findings suggest that a more generalized bone alteration may be the basis of the pathogenesis of osteoarthritis. Microsc. Res. Tech. 37:358–371, 1997. r 1997 Wiley-Liss, Inc. INTRODUCTION The damage to articular cartilage that is characteristic of osteoarthritis is accompanied by changes in the subchondral bone. It is not certain whether the bony changes precede or follow injury to the cartilage. Osteoarthritis is usually considered to be a primary disorder of chondrocyte function with secondary changes in bone. However, a defect in the subchondral bone resulting in loss of its shock absorbing capacity could transfer the stress of loading directly to the articular cartilage with secondary changes in the cartilage. Changes in the subchondral bone may occur as a result of local factors, such as altered weight bearing, or as part of a generalized disease of bone, such as Paget’s disease. Simon et al. (1972) induced cartilage damage in rabbits by repetitive excessive impulse loading. The damage was preceded by changes in the subchondral bone. Radin and Paul (1970) postulated that excessive impulse loading of the joint causes thickening of the subchondral bone as a result of repair of numerous microfractures. This bone is stiffer and less pliable so that the impact of loading is borne principally by the overlying cartilage, which, as a result, degenerates. r 1997 WILEY-LISS, INC. The morphological appearances in osteoarthritis are well known: the subchondral bone undergoes a number of changes. Sclerosis occurs in areas of thinning and loss of the cartilage, particularly in the central area lateral to the fovea of the femoral head. Osteophyte production occurs around the circumference of the joint margin, and the so-called cysts, which contain fibrovascular tissue or metaplastic cartilage, are often surrounded by bone showing active new bone formation (Amir et al., 1992). What initiates this subchondral remodeling? Wolff’s law states that the structure of any bone is a function of the mechanical demands made on that bone. The load applied to any joint can be modified by many factors, such as weight, amount and type of exercise, occupation, integrity of neuromuscular control. Metabolic and systemic influences can also alter the subchondral bone, rendering it less able to withstand normal mechanical demands. The latter hypothesis will be dis*Correspondence to: Prof. Dr. J. Dequeker, Arthritis and Metabolic Bone Disease, Research Unit, U.Z. Pellenberg, Weligerveld 1, B-3212 Pellenberg, Belgium. Received 22 January 1995; Accepted 18 March 1995 BONE DENSITY AND LOCAL GROWTH FACTORS IN OA cussed here in more detail and illustrated by means of fundamental studies on bone. OSTEOARTHRITIS OF THE HIP Changes in Bone Mechanics In 1969, Sokoloff studied the pathology of osteoarthritis in human hip joints and found that cartilage fibrillation could not be dissociated from bony changes, even in the earliest stage of osteoarthritis. Radin and Paul (1970) were the first investigators to propose that subchondral bone stiffening, due to healing microfractures, might be the primary pathogenic event in osteoarthritis. They performed biomechanical tests on subchondral bone plugs, removed at autopsy, from the weight-bearing portion of the medial femoral condyle in 43 male patients with varying degrees of osteoarthritis, as well as normal controls. There was a significant decrease in the energy-absorbing capacity of the subchondral bone in mild osteoarthritis vs. normal energy absorption in moderate and severe cases as well as in the controls. It was suggested that in mild osteoarthritis, the articular cartilage would be exposed to increased forces and eventually wear, due to the decreased energy attenuating properties of the underlying subchondral bone. Further support for the above mechanism was provided by Ewald et al. (1982) when they replaced the subchondral bone in living dogs with methacrylate cement which was 2.6 times stiffer than normal cancellous bone. After 2 years of normal abulation, there was uniform loss of the articular cartilage in the weight-bearing regions of all the femoral heads. Other studies of osteoarthritic human hips noted that the bony changes (including trabecular bone volume, ash mineral, alkaline and acid phosphatase activity) were greater in the weight-bearing region (Cameron and Fornasier 1979; Reinmann and Christensen 1979; Reinmann et al., 1977). In 67% of osteoarthritic femoral heads studied histologically by Batra and Charnley (1969), osteoid was found exclusively in the pressure zones. Histomorphometric Alterations Amir et al. (1992), studying quantitative histomorphometry after double tetracycline labeling in human osteoarthritic femoral heads, found that remodeling of subchondral bone in osteoarthritis is fairly constant and is not related to weight-bearing. They also stated that endochondral ossification is an unlikely mechanism for bone sclerosis in osteoarthritis. In order to determine whether subchondral bone in osteoarthritis differs from that seen in normal human aging osteoarthritic femoral heads removed for total hip arthroplasty compared with normal age-matched and young autopsy controls, Grynpas et al. (1991) studied extensively the femoral head. Standardized, 1-cm deep, weight-bearing and nonweight-bearing subchondral bone blocks, as well as cancellous core bone, 2–4 cm deep to the articular surface, were examined in each femoral head. Mineralization was assessed using density fractionation and chemical analyses, and compared to histomorphometry. In osteoarthritis, both weight-bearing and nonweight-bearing surface subchondral bone showed a lower degree of mineralization than age-matched and young controls. Histomorphometric analysis showed that subchondral bone thickness, as well as all osteoid 359 parameters and eroded surfaces, were increased in osteoarthritic samples vs. controls. Mineralization in the deep cancellous core bone increased with normal aging but underwent less change with osteoarthritis. Histomorphometry of the cancellous core showed that osteoid parameters, but not bone volume, were increased in osteoarthritis vs. controls. They concluded that advanced osteoarthritis of the femoral head is associated with a thickening of the subchondral bone with an abnormally low mineralization pattern. The observations of Grynpas are in line with those of Kusakabe (1977), who found that bone remodeling is increased in the subchondral proximal femoral head bone of osteoarthritis cases and that the osteoid surface and osteoblastic activity are so great that trabeculae and larger relative bone area are produced. Bone volume in the femoral head of osteoarthritis cases and controls has also been studied by Moore et al. (1994). They found that the cartilage overlying the subchondral compressive and tensile regions was significantly thinner and that bone volume was increased at the principal compressive region and diminished at both of the tensile regions. Figure 1 illustrates the structural differences between osteoarthritis and controls at the compression and tensile regions of the femoral head subchondral bone. In an earlier study of the same group, Fazzalari et al. (1985) studied histomorphometric changes in the trabecular structures of the principal compressive stress regions in the femoral head of patients with osteoarthritis and patients with femoral neck fracture. They found that the trabeculae in osteoarthritis were significantly thicker than in femoral neck fracture and that spacing in osteoarthritis was significantly greater than in femoral neck fracture. The different relationships indicate, therefore, that the structural arrangement of the cancellous bone in each group is unique and reflects the different pathophysiology and biomechanics of the femoral head in osteoarthritis and femoral neck fracture. Taking the derived parameters of trabecular thickness and trabecular spacing, it was possible to draw the average trabecular dimensions for each group that reflect the differences of architecture in the principal compressive stress region (Fig. 2). Trabecular Microfractures In Radin’s concept on the cause of stiffening of the subchondral bone is the healing of trabecular microfractures. When the joint is subjected to repetitive impulsive loading, the bone underlying the joint, especially the well-vascularized subchondral area, responds by microfracturing and remodeling of its internal architecture to better resist these stresses. This stiffer bone is no longer as effective a shock-absorber as it was before. The articular cartilage is now relatively unprotected and the peak dynamic forces it is subjected to are increased. Fazzalari and co-workers (1987) examined the head of the femur from 67 patients who had joint replacement for advanced osteoarthritis and 66 autopsy controls without discernible joint disease. They used a microscopic approach to ascertain fractures in coronal slabs and to quantify the mineralization and thickness of trabecular bone in a principal compressive zone (Fig. 360 J. DEQUEKER ET AL. A. B. Fig. 1. A: Sections from the subchondral principal compressive region of control (a) and osteoarthritic (b) cases, which show the differences in the structural variables. From Fazzalari et al. (1992). B: Sections taken from the subchondral principal tensile region of control (a) and osteoarthritic (b) cases, which show more complex structural changes than occur in the subchondral principal compressive region. From Fazzalari et al. (1992). 3). There was a reduction in the number of trabecular microfractures in osteoarthritis patients, compared with the controls, with a lack of correlation between numbers of microfractures and age in osteoarthritis patients (Fig. 4). There was no evidence for the hypothesis that increased numbers of microfractures led to the increase of bone to support the view that microfractures play a role in maintaining osteoarthritic joint structure. This study confirms earlier studies of trabecular stress fractures in early and advanced osteoarthritis of the hip (Cameron and Fornasier, 1975). Koszyca et al. (1989) evaluated the role of healing trabecular microfractures by studying their nature and distribution in femoral heads from patients with osteoarthritis of the hip and subcapital fracture of the femoral neck. The nature and distribution of healing trabecular fractures of femoral heads were studied in 50 autopsy specimens, 30 patients with osteoarthritis of the hip, and 45 patients with subcapital fractures of the femoral neck. Macerated coronal slices were examined with a dissecting microscope, and the trabecular microfractures were identified where there was callus formation, either nodular or smooth. The number of trabecular microfractures did not differ significantly in the subcapital fractures of the femoral neck and age-matched control groups. The number of trabecular microfractures in the osteoarthritis group was lower than in subcapital fractures of the femoral neck and age-matched control groups. The numbers of nodular trabecular microfractures in the control and subcapital fractures of the femoral neck groups did not differ significantly but were fewer in the osteoarthritis group. Smooth trabecular microfractures were present in similar numbers in osteoarthritis, control, and subcapital fractures of the femoral neck groups. The low number of trabecular microfractures in osteoarthritic bone may indicate that the proposed pathologic remodeling of bone following repeated impact loading has already occurred and that the trabecular network has become stronger as a result of trabecu- BONE DENSITY AND LOCAL GROWTH FACTORS IN OA Fig. 2. Schematic drawings of histologic sections showing the average trabecular dimensions in the selected stress region of the femoral neck for control, osteoarthritis, and fractured neck of femur samples. The drawings show the correct relative relationships among the three groups. From Fazzalari et al. (1985). lar thickening and increased connectivity, decreasing the likelihood of fracture. Another possibility is that the thicker trabeculae were already present as part of a constitutional stronger bone built up during adolescence. OSTEOARTHRITIS OF THE KNEE The tibial subchondral bone in osteoarthritis of the tibia was extensively studied histologically and microradiographically by Havdrup et al. (1976). They found that the tidemark was double or multiple as has also been described by Green et al. (1970). The borderline between the mineralized cartilage and the subchondral bone was very irregular and each tissue showed extensions which fitted in between one another as has been described by Stougård (1974). The subchondral bone had a strong tendency to undergo sclerosis in the form of thickening and confluence of the trabeculae to form massive blocks of bone containing single vessel canals. 361 When the bone was completely denuded the surface layer exhibited necrotic osteocytes. The severity of the sclerosis clearly varied with damage to the overlying cartilage. When the cartilage was relatively intact the underlying trabeculae were of normal appearance. The thinner the cartilage the more massive the sclerosis. The sclerosis was clearly focal. The sclerotic changes varied in depth from 500 to 3,000 µ but were often abruptly bordered by spongy bone of normal appearance. No microscopic fractures or nodular aggregates (‘‘birds nests’’) (Todd et al., 1972) were found in this spongy bone. Mostly the trabeculae were highly mineralized without appreciable signs of remodeling of the bone. Where the osteoarthritis was severe, it had locally broken through the mineralized cartilage, which otherwise appeared to be very resistant. In such cases resorption cavities extending downwards from the surface and filled with granulation tissue were visible. The spaces between the bone trabeculae contained fatty marrow, except where the trabeculae were sclerotic; here the tissue was fibromyxotic and poor in cells. The resorption cysts contained highly cellular tissue. Osteoblasts were observed on the sclerotic trabeculae and to a lesser extent on the normal trabeculae. The osteoblasts were invariably flattened, suggesting low activity (so called resting osteoblasts). Only in the resorption cavities (cysts) the osteoblasts were more active and protrudent. The amount of osteoid tissue found in osteoarthritic bone stained ad modum Goldner was small, but increased with the severity of the osteoarthritis and was found particularly on trabeculae surrounding granulation tissue and in resorption cavities. These cavities also contained osteoclasts which were hardly ever seen elsewhere. New bone formation occurred inside the cysts. The specimens contained very few osteophytes, but where they were found, they were seen to form on the remains of the mineralized cartilage, frequently with a persisting tidemark. As shown for the femoral head, Pugh et al. (1974) have found at the femoral condyle evidence of changes in geometric pattern of trabeculae in the early stages of degenerative arthritis in the subchondral bone; changes in the spatial arrangement of the trabeculae (rather than the gross density of bone) were shown to correlate with the observed increase in subchondral bone stiffening associated with loss of cartilage mucopolysaccharides. Data on bone mineral density (BMD) of the proximal tibia in osteoarthritic knees assessed in vivo by photon absorption were recently reported by Madsen et al. (1994). Compared to the controls, BMD of the subchondral plate, BMD of the medial condyle, the medial to lateral distribution ratio of subchondral BMD, and the ratio between BMD of the subchondral plate and of the immediately underlying region were significantly increased in patients with predominantly medial osteoarthritis of the knee. The BMD of the lateral condyle was lower in patients with osteoarthritis of the knee than in controls. After adjustment for weight differences between patients with osteoarthritis of the knee and controls, the differences in bone density were still present. Mokassa et al. (1993) studied the effect of body weight and osteoarthritis grade on the mineralization pattern of subchondral trabecular bone of cadaveric 362 J. DEQUEKER ET AL. Fig. 3. Trabecular microfractures on trabecular bone in macerated coronal slabs of the femoral head. a, nodular lesion; b, fusiform lesion. Original magnification 325. From Fazzalari et al. (1987). opposite to the effect of body weight. As shown in Figure 5, the mineralization profile is shifted to higher density in the high osteoarthritis grade while in the obese a shift to a lower density mineralization profile was found, indicating that despite a positive relation between obesity and osteoarthritis they have opposite effects on bone quality. Fig. 4. Pooled male and female data for trabecular microfractures vs. age for both normal controls and osteoarthritis patients. No significant relationship was found for osteoarthritis patients. Their scatter with respect to the 95% confidence interval for controls is shown. Controls, trabecular microfractures 0.47 3 Exp (0.064 3 Age). Osteoarthritis, NS. From Fazzalari et al. (1987). proximal tibial epiphyses. The effect of osteoarthritis grade on mineralization degree as defined by gradient density separation of trabecular bone powder was THE SUBCHONDRAL PLATE AND TIDEMARK IN OSTEOARTHRITIS Subchondral Plate and Tidemark in Tibial Bone The subchondral plate consists of two mineralized layers which together act as a single unit, separating the articular cartilage from the bone marrow cavity. On the articular side at the line of contact between subchondral plate and articular cartilage is a discrete band of mineralized cartilage which is more radiodense than adjacent bone, has affinity for numerous histological stains, and is recognized as the ‘‘tidemark’’ in hematoxylin stained sections (Green et al., 1970). From the tidemark, the calcified cartilage extends for a varying distance toward the bone marrow cavity, where it is remodeled and replaced with woven, or lamellar bone similar to the supporting trabeculae. The thickness of the subchondral plate varies in different places and in different joints. In most joints where the bones have a concave and convex component, the dome-shaped convex subarticular bone structures are thinner and more uniform in shape than the complementary part of the joint where the center of the concavity is associated with a much thicker subchondral plate than at the periphery (Simkin et al., 1980). In the case of the tibial plateau, some modification of these general observations is noted since the bony plateau is almost flat; the presence of the wedge-shaped meniscal cartilage (measuring thickness from 0 mm at the center of the plateau to 5 to 7 mm at the periphery) produces a functional BONE DENSITY AND LOCAL GROWTH FACTORS IN OA 363 Fig. 5. Opposite effect of body weight and osteoarthritis grade of the knee on distribution of trabecular bone density of the proximal tibia. From Mokassa (1993). concavity in which the condylar end of the femur articulates. At the center—contact area—of each plateau the subchondral bone is thicker than at the more peripheral regions. The hyaline articular cartilage is closely attached to the irregular surface of the subchondral bone at the tidemark. This cartilage is traversed from the tidemark to its surface by ‘‘arcades’’ of collagen fibers which emanate at approximately 90° from the calcified cartilage in bundles 55 µ in diameter (Minns and Stevens, 1977). Deep to the surface and anchoring these collagen bundles at the tidemark level is a ramification or ‘‘root system’’ as they penetrate deeper into the calcified cartilage (Broom and Poole, 1982; Inove, 1981). Whilst not all collagen fibers are directed in this fashion, some are known to be aligned at more acute or obtuse angles. The chondrocytes buried between the collagen fibers maintain control over a region or domain in their proximity and appear to control the turnover of the noncollagenous matrix. The mechanisms of repair, if any, of the collagen fibers themselves is uncertain. Subchondral Plate Including Tidemark in Osteoarthritis The role of subchondral plate in the development of osteoarthritis is of particular interest and has not received much attention in the search for the pathogenesis of osteoarthritis. Its role in the development of osteoarthritis may depend not only upon its own mechanical and metabolic properties, but also upon its influence on the nutrition of the adjacent cartilage. While it is stated (Schubert and Hamerman, 1968) that the nutrition of cartilage is facilitated by the compression and relaxation during normal joint use, and that the influx and expulsion of synovial fluid from the interstices of spongelike hyaline cartilage provides the essential nutrition, other authors have suggested that some of the deeper layers of cartilage cells are metabolically influenced by the subchondral capillaries (Ekholm and Norbäck, 1951; Ingemark, 1950; Sokoloff, 1969; Trueta and Harrison, 1953; Woods et al., 1970). If one presumes that both the superficial diffusion and also the permeation of nutrients from beneath the subchondral plate may control the metabolism of the cartilage, then the development of osteoarthritis might follow nutritional interference by 1) primary damage and disintegration of the surface of the articular cartilage due to compression or excessive use independently of the subchondral plate structure, 2) remodeling of the subchondral plate may interfere with the nutrition of the deeper cartilage layer, or 3) combinations of both. According to Duncan et al. (1985), injury to the subchondral plate may generate a response which alters the access of nutrients to the deeper cartilage layers either by 1) fracture of the subchondral plate and its immediate local metabolic effect on the adjacent cartilage; 2) the occlusion of the vascular channel by the repair process, giving rise to relative ischemia; 3) perforations and identations of the nonarthritic tibial plateau may be the sites of increased damage penetration and cyst formation seen in the osteoarthritic joint; or 4) a stiffened joint architecture predisposes to direct cartilage damage. Oettmeier et al. (1989) further analysed the tidemark changes in osteoarthritic and control femoral heads. Histological and histomorphometric examinations were carried out on 88 human femoral heads, 63 from patients with osteoarthritis and 25 controls. The changes of the tidemark were found to be multiform in the osteoarthritic process and could be classified into three degrees of severity. Low-grade tidemark changes were characterized by reduplications of the tidemark and discontinuities of the tidemark line, occurring also in about one third of the controls. Vascular invasion into the tidemark as well as incipient calcification of basal hyaline cartilage were observed in middle-grade tidemark alterations. High-grade changes were distinguished by the disappearance of the tidemark, advanced mineralization and ossification of basal hyaline and calcified cartilage and finally by the ‘‘tidemark and bald bone.’’ Increased Metabolic Activity at the Tidemark in Osteoarthritis Differences of metabolic activity in the calcified zone of cartilage in particular at the tidemark might be of importance in the development of osteoarthritis. Active though slow remodeling takes place at the articular cartilage/subchondral bone interface in adult humans (Green et al., 1970) and in rabbits (Lemperg, 1971). Lane et al. (1977) showed that this endochondral ossification increases in late middle age, especially in more loaded areas. Lemperg (1971) and Bullough and Jagannath (1983) showed that the tidemark is a metabolically active area involved in mineralization of the overlying noncalcified hyaline cartilage. Endochondral 364 J. DEQUEKER ET AL. ossification depletes the calcified cartilage at the cartilage/bone interface and the tidemark has been thought of as a calcification front advancing in the direction of the noncalcified cartilage. Duplication of the tidemark is cited as evidence for this advancement and indeed duplication is increased in normal late middle age and in osteoarthritic femoral heads (Lane and Bullough, 1980). Apart from this activity at the calcified-hyaline cartilage interface, the calcified zone of cartilage has been considered inactive. Revell et al. (1990) studied tetracycline labelled articular cartilage of 50 osteoarthritic femoral heads in an attempt to understand the mechanisms of endochondral ossification and tidemark advancement. They confirmed that the tidemark is a metabolically active part of calcified cartilage with a much higher level of activity than previously suspected. That there was tetracycline incorporation in 20 of 43 tidemarks during the 18-day period before joint replacement is testimony to the extent to which the tidemark and calcified cartilage represent a metabolically active zone. The metabolic role of the tidemark zone in the pathogenesis is further emphasized by Gannon et al. (1991), who detected type X collagen in the pericellular matrix of the chondrocytes at or just above the tidemark, and this expression was immediately before mineralization. They suggest that the expression of growth platelike properties would allow the deep chondrocytes of mature articular cartilage to play a role in the remodeling of the joint with age and in the pathogenesis of osteoarthritis. Rees and Ali (1988) examined the distribution of alkaline phosphatase activity in human articular cartilage from normal and osteoarthritic joints by an electron microscope technique. Chondrocytes and matrix vesicles close to the tidemark were positive for alkaline phosphatase activity in osteoarthritic cartilage within the extracellular matrix of osteoarthritic cartilage. Large numbers of matrix vesicles were found. Because there is a specific relation between alkaline phosphatase activity, number of matrix vesicles, and initial mineral formation in the tidemark region of articular cartilage, these results give evidence for increased metabolic activity in articular damage in osteoarthritic joints. SUBCHONDRAL BONE DENSITY AT THE SITE OF OSTEOARTHRITIC JOINTS Remodeling of subchondral bone has been considered to have a key role in the pathogenesis and perpetuation of the changes in osteoarthritis. Bone sclerosis can result from an increased rate of appositional bone formation or decreased resorptive activity, an increased amount of bone forming surface (increased osteoid), and increased endochondral ossification of the calcified layer of the articular cartilage. Ossification also occurs in plugs of metaplastic cartilage within the eburnated bone at the joint surface, but this mechanism probably has a minor role, if any at all, in the alteration of the bony contour. SUBCHONDRAL BONE DENSITY IN EXPERIMENTAL MODELS OF OSTEOARTHRITIS Subchondral bone density has been studied in a few animal models for osteoarthritis. Layton et al. (1988) used microscopic computed axial tomography to evaluate the subchondral bone structure in femoral heads from a guinea pig model of spontaneous osteoarthritis. Their findings showed that trabecular remodeling occurs deep within the femoral heads in this animal model of early osteoarthritis. These changes occurred when only mild cartilage changes were present. The trabecular bone changes consisted of thicker trabeculae that were closer together. These findings are consistent with the notion that bone changes are an early event in this model of osteoarthritis. Dedrick et al. (1993) evaluated the sequence of changes in articular cartilage, trabecular bone, and subchondral plate in dogs with osteoarthritis, 3 months, 18 months, and 54 months after anterior cruciate ligament transection. Specimens of the medial tibial plateau were analyzed with microscopic computed tomography (micro-CT) at a resolution of 60 µm, and biochemical and morphologic changes in the femoral articular cartilage were assessed. At 3 months and 18 months after anterior cruciate ligament transection, the articular cartilage in the unstable knee showed histologic changes typical of early osteoarthritis and increased water content and uronic acid concentration; by 54 months, full-thickness ulceration had developed. Micro-CT analysis showed a loss of trabecular bone in the unstable knee, compared with the contralateral knee, at all time points. At both 18 and 54 months, the differences in trabecular thickness and surface-to-volume ratio were greater than at 3 months. Although the mean subchondral plate thickness, especially in the medial aspect of the medial tibial plateau, was greater in the osteoarthritic knee than in the contralateral knee 18 months and 54 months after anterior cruciate ligament transection, these differences were not statistically significant; however, the difference was significantly greater at 54 months than at 3 months. They concluded that thickening of the subchondral bone is not required for the development of cartilage changes of osteoarthritis in this model. The bony changes that develop after anterior cruciate ligament transection, however, could result in abnormal transmission of stress to the overlying cartilage and thereby contribute to the progression of cartilage degeneration. OVERALL BONE DENSITY AND OSTEOARTHRITIS In line with Radin’s aetiological ‘‘impulse loading’’ concept of osteoarthritis and the role of subchondral bone, we have studied the hypothesis that the subchondral bone stiffness is part of a more general bone alteration instead of local microfractures in patients sufferering from generalized osteoarthritis. Osteoporosis Protects Against Osteoarthritis It is known that osteoporosis spares osteoarthritis (Smith and Rizek, 1966; Urist, 1960), and that osteopetrosis, on the contrary, is highly associated with osteoarthritis in the few patients with this condition who reach adulthood (Milgram and Jasty, 1982). Osteoporotic bone would be relatively soft and would act as an excellent shock-absorber easily sustaining compression fracture of its relatively weak structure. The osteoporotic bone simply lacks the mass, even in the presence BONE DENSITY AND LOCAL GROWTH FACTORS IN OA of numerous micro- and macrofractures, to stiffen sufficiently and lose its protective shock-absorbing role vis-à-vis the overlaying cartilage. Fracture is a very effective alternative way of absorbing energy. Although much has been studied about the possible causes of osteoarthritis, we are still unable to readily explain why some individuals remain free from the disease throughout life while others are affected. The observation that not everyone will develop osteoarthritis or osteoporotic fractures with ageing, and that primary osteoarthritis and primary osteoporosis rarely coexist (Verstraeten et al., 1991), stimulated us to study these two conditions in order to elucidate pathogenetic pathways. Osteoporosis and osteoarthritis are two common agerelated musculoskeletal disorders associated with considerable morbidity and mortality. Patients with postmenopausal osteoporosis and those with osteoarthritis appear to represent anthropometrically different populations (Dequeker et al., 1983). The osteoporotic women were shorter, more slender, and had less fat, muscle girth, and strength, while the women with osteoarthritis, although of comparable age and skeletal size, were more obese and had more fat, muscle mass, and strength. General clinical experience is that, although both common in elderly patients, osteoporosis and osteoarthritis (Christiansen et al., 1981; Cooper et al., 1991; Dequeker et al., 1975; de Sèze et al., 1962; Foss and Byers, 1972; Roh et al., 1974; Urist, 1960; Verstraeten et al., 1991), and, in particular, fractures and osteoarthritis of the hip, seldom occur together (Aström and Beertema, 1992; Pogrund et al., 1982). Osteoarthritis of the hip is in addition a negative risk factor for hip fracture (Dequeker et al., 1993a) and for compression fractures of the spine (Healy et al., 1985). Despite strong evidence that osteoarthritis protects against or retards the development of osteoporosis, the biological explanation for this link is not yet clear. Osteoarthritis is usually considered as a disease of cartilage failure with secondary bone changes such as osteophytes and subchondral sclerosis. Osteoporosis, on the other hand, is considered to be an age-related condition characterized by a reduced amount of bone, leading to diminished physical strength of the skeleton and an increased susceptibility to fracture. Primary Osteoarthritis: Initially a Bone Disease? At present, evidence is accumulating that primary osteoarthritis might initially be a bone disease rather than a cartilage disease. Osteoarthritis cases have a better preserved bone mass (Carlsson et al., 1979; Foss and Byers, 1972; Hannan et al., 1993; Moore et al., 1994; Roh et al., 1974), even independently of body weight (Belmonte-Serrano et al., 1993; Gotfriedsen et al., 1990; Hart et al., 1994; Hordon et al., 1993; Lane and Nevitt, 1994; Mokassa et al., 1993; Nevitt et al., 1992; Vandermeersch et al., 1990). No significant differences in calcitropic hormones, PTH, 25(OH) and 1.25(OH)2 vitamin D3 metabolite levels could be established (Geusens et al., 1983), but fasting growth hormone levels and growth hormone levels after stimulation with hypoglycaemia were above normal in osteoarthritis (Dequeker et al., 1982). Calcium excretion, both over 24 hours and when fasting, was significantly higher in osteoporotic cases. 365 Dynamic tests for osteoblast function stimulation of osteocalcin production by 1.25 dihydroxyvitamin D3 treatment revealed no difference between osteoarthritis and osteoporosis cases and between osteoarthritisosteoporosis and young controls, indicating that osteoblasts remain responsive to stimulation in both conditions (Geusens et al., 1991). Serum levels of IGF-I are elevated in patients with acromegaly, a disease associated with the development of secondary osteoarthritis (Bluestone et al., 1972). New bone formation has also been related to increased serum levels of IGF-I in patients with diffuse idiopathic skeletal hyperostosis (Littlejohn et al., 1986), and, in one study, both the size and the growth of osteophytes were directly correlated with serum levels of IGF-I in subjects with osteoarthritis of the knee (Schouten et al., 1993). Other studies, however, have yielded conflicting results (Denko et al., 1990; Hochberg et al., 1994; McAlindon et al., 1992). Hochberg et al. (1994) examined the relationship between serum levels of IGF-I and the presence and severity of radiographic changes of knee osteoarthritis in subjects participating in the Baltimore Longitudinal Study of Aging. Mean serum IGF-I levels were significantly lower in subjects with knee osteoarthritis; however, after adjustment for age-related changes in IGF-I levels, these differences were no longer significant. We examined serum levels of IGF-I in osteoarthritis and osteoporosis cases and did not find significant differences and also not after stimulation with growth hormone releasing factor (Dequeker et al., 1994). We also studied serum osteocalcin levels as marker of osteoblast function and bone turnover and bone mineral content in the spine and radius in osteoarthritis and osteoporosis cases. Bone mineral content in the spine (corrected for excessive osteophyte formation) and in the radius was significantly higher in osteoarthritis cases. No difference in serum osteocalcin levels was found between osteoarthritis and osteoporosis but osteocalcin levels were significantly negatively correlated with bone mineral content in the spine in osteoporosis, but no correlation was found in osteoarthritis (Gevers et al., 1988). In a study of 19 cases in whom there was a coexistence of spinal osteoarthritis and osteoporosis, we found that these cases compared to cases with osteoarthritis alone, had the following characteristics: older age, more menopausal years, smaller stature and lower body weight, a higher serum PTH level, and more nulliparity than both other groups. They have less osteoarthritis of the hip than osteoarthritis alone cases, and less forearm and other fragility fractures than osteoporosis alone cases, though the incidence of femoral neck fractures was comparable. These results suggest that osteoarthritis might have a protective or retarding effect on the development of osteoporosis (Verstraeten et al., 1991). The observed differences in bone mass and bone metabolism parameters between osteoarthritis and osteoporosis raises the possibility that the physiopathology of osteoarthritis might be linked rather to bone changes with secondary cartilage suffering. The increase in bone mass and change in bone quality might alter the mechanical properties of subchondral bone resulting in a reduced shock absorbing efficiency and 366 J. DEQUEKER ET AL. Fig. 6. Osteoarthritis grades of the hand. From Kellgren (1963). leading to subchondral fracture and cartilage degeneration. This hypothesis might explain why in osteoporotic cases osteoarthritic changes at the joints and the spine are rarely seen (Dequeker et al., 1975; Pogrund et al., 1982; Weintroub et al., 1982). These data are in support of the hypothesis of Radin and Paul (1970), who postulated that the primary defect in osteoarthritis is not in the articular cartilage, but rather in the subchondral bone. According to Radin et al. (1973), the initial process of bone thickening in osteoarthritis is due to callus formation at the sites of multiple fatigue fractures. Our hypothesis is that in osteoarthritis there is a generalised bone alteration which might be accentuated in weight bearing areas. BONE QUALITY OF THE ILIAC CREST IN OSTEOARTHRITIS In order to study the hypothesis that osteoarthritis and osteoporosis are part of a generalised bone disease, we decided to study bone at the iliac crest, because this is a part of the skeleton where no effect of local joint destruction can be expected. Data on the analysis of iliac crest bone, consecutively removed in the autopsy room of a large university hospital, will now be reported. Iliac crest specimens, obtained from 38 women, aged 60–75 years, were analysed for biochemical, quantitative histomorphological, and biomechanical characteristics. For each subject, an X-ray of the right hand was obtained in the necropsy room for the grading of osteoarthritis in the small hand joints. The osteoarthritic changes were graded 0 to 4 according to the criteria of the Atlas of Standard Radiographs of Arthritis (Kellgren, 1963) by two independent readers (Fig. 6). Eleven patients had grade 0, 10 had grade I, and 17 had grade II to IV. Grouping of patients was done before TABLE 1. Physical characteristics of the iliac crest in women graded according to the degree of osteoarthritis Grade 0 (n 5 8) Grade I (n 5 10) Grade II–IV (n 5 14) Apparent density (trabecular) Fat-free dry weight/crude bone (g/cm3) 16.9 6 4.8 23.8 6 3.6*** 26.27 6 8.97** Pure bone density (trabecular) Fat-free dry weight/crude bone 2.1 6 0.1 2.2 6 0.1 2.1 6 0.3 % Pure/crude bone Trabecular 8.03 6 2.80 10.9 6 2.2* 16.76 6 9.0** Trabecular 1 cortical 31.8 6 9.4 36.3 6 3.9 42.6 6 8.5** Mean 6 SD. Significance compared to osteoarthritis grade 0: *P , 0.05; **P , 0.02; ***P , 0.005, Student’s t test. results of the different tests were available. One patient was excluded because of local metastasis in the iliac crest. Bone Density Pure crude bone volume and iliac crest density of bone cylinders were determined, using the principle of Archimedes, with carbon tetrachloride as a medium (Dequeker et al., 1971). Apparent density (fat-free dry weight/pure bone volume) was significantly higher in the osteoarthritis group (Table 1). Since analysis on whole bone samples represents a weighted average of microscopic portion of tissue of different physiological maturity, metabolic changes occurring in a small portion of the total bone mass may 367 BONE DENSITY AND LOCAL GROWTH FACTORS IN OA Fig. 7. Distribution of cortical bone of various densities in young adult control (A) 60- to 75-year-old control (B) and osteoarthritis (C) patients. Values are expressed as weight of the fraction per total bone volume (%). Bars indicate SEM. Significant differences: *vs. control 60–75 years; Q vs. control young adults; *Q P , 0.05; **QQ P , 0.01; QQQ P , 0.001. From Raymaekers et al. (1992). TABLE 2. Bone histomorphometry of the iliac crest in women graded according to the degree of osteoarthritis Osteoarthritis grade 0 I n58 n 5 10 II–IV n 5 14 Iliac bone thickness (mm) Cortical thickness (µm) Cortical thickness/iliac bone thickness (%) Cortical bone volume (%) Trabecular bone volume (%) Trabecular width (µm) 6.09 6 1.18 6.99 6 2.43 NS 7.53 6 0.92 P , 0.01 521.68 6 188.96 731.12 6 494.7 NS 893.41 6 210.06 P , 0.001 8.68 6 3.15 9.68 6 3.67 NS 11.67 6 2.02 P , 0.02 89.71 6 1.67 89.77 6 2.51 NS 88.19 6 3.38 NS 12.67 6 3.28 13.2 6 5.4 NS 17.87 6 5.98 P , 0.05 116.41 6 18.4 129.31 6 36.67 NS 158.82 6 12.33 P , 0.015 Mean 6 SD. Significance compared to osteoarthritis grade 0. not be detected. Pathogenetic mechanisms behind diseases can thus be overlooked, especially in the elderly, when phases of low and high turnover bone may be mixed. The density fractionation technique of bone powder gives us an unique procedure to study ageing in situ and to disclose early pathogenetic events. The distribution pattern of cortical bone powder from the iliac crest (Raymaekers et al., 1992) shows a shift to higher density fractions in the osteoarthritis cases, which implies more mineralized bone with lower proportions of young osteons (Fig. 7). histomorphometric changes in the trabecular structure of a selected stress region in the femur in patients with osteoarthritis: thicker trabeculae were found in osteoarthritis compared to normals and femoral neck fracture cases (Fazzalari et al., 1985). Thus, in osteoarthritis not only is cortical bone mass increased, but so is trabecular bone as shown by our studies using two different techniques for trabecular bone measurement. Figure 8 illustrates the striking difference in iliac crest bone density in a case with osteoarthritis at the hand (grade II–IV) (B) and a case without hand osteoarthritis (A). Histomorphometry Using quantitative histomorphological methods, it was found that in the osteoarthritis patients of grades II–IV, the cortical thickness, percentage of trabecular bone, and trabecular thickness of iliac bone were significantly higher compared to the group which had no osteoarthritic changes on X-ray at the hands (Gevers et al., 1989a) (Table 2). The latter finding confirms the Bone Mechanical Properties Mechanical properties of iliac crest trabecular bones have been tested in 21 specimens. A significant difference in stiffness elastic modulus as measured by compression tests was found, the osteoarthritic trabecular bone being significantly stiffer than non-osteoarthritic trabecular bone (Fig. 9). A significant correlation was found between elastic modulus, apparent density, the 368 J. DEQUEKER ET AL. A B Fig. 8. Microradiographs of iliac crest sections (310). A: Osteoarthritis grade 0; B: Osteoarthritis grade II–IV. From Gevers et al. (1989a). percentage pure to crude bone volume, and cortical thickness and trabecular width, the latter two evaluated by quantitative histomorphometry. Compressive strength was higher in osteoarthritic bone but not significantly. A significant correlation between trabecular bone volume and trabecular width and maximum compressive strength was established (Gevers et al., 1989b). Biochemical Composition—Growth Factors in Bone Iliac crest bone matrix composition in the osteoarthritis group differs significantly in several aspects from the control group without osteoarthritis at the hand joints (Gevers and Dequeker, 1987). In the EDTA, significantly less glycoproteins and proteoglycans could be extracted, while the osteocalcin content was found to be significantly increased in the osteoarthritis group (Gevers and Dequeker, 1987). We also found that IGF-I and IGF-II and TGF-b concentrations are increased in the matrix of the iliac crest (Dequeker et al., 1993b). IGF-I concentrations were measured by radioimmunoassay (RIA) using a recombinant human IGF-I (Ciba-Geigy Basel, Switserland). IGF-II concentrations were determined by radioreceptor assay (RRA) using H-35 rat hepatoma cells (American Type Culture Collection CRL 1548) as the receptor source. TGF-b concentrations were determined by bioassay, measuring the inhibition of DNA synthesis by mink lung epithelial cells (Mv. ILU American type Culture Collection CL64). Osteocalcin was measured in the supernatants of the EDTA extracts by RIA. Because these locally produced growth factors are resistant to an acidic milieu, they can be released and activated during bone resorption and act on osteoblast and osteoclast progenitor cells in a paracrine or autocrine fashion. It has been suggested that the IGFs and TGF-b in bone may act to link bone formation to resorption in the remodeling cycle (Baylink et al., 1990; Canalis et al., 1988; Farley et al., 1987). Many biologic properties attributed to the IGFs and TGF-b make them ideal ‘‘coupling factors.’’ In addition to their well-recognized anabolic activity on bone cells, these growth factors are deposited in large quantities in human bone, where they would be available for action after osteoclastic resorption. Because the IGFs (Mohan et al., 1986) and TGF-b (Jennings, personal communication) are stable at acidic pH, they may be released during resorption in a bioactive form to act on either osteoblasts or osteoblast precursor cells. The higher concentrations of these growth factors in bone in osteoarthritis would predict the observed findings that, in general, patients with osteoarthritis would maintain a greater bone mass. The increased density of bone in patients with osteoarthritis may have a negative impact. This increased density may impair the resilience of the bone immediately adjacent to the articular space, reducing the shock-absorbing capacity and contributing to the pathogenesis of osteoarthritis and further degeneration of the cartilage. These quantitative and qualitative biochemical differences according to the presence of osteoarthritis of the hands confirm our earlier findings and suggest that bone matrix alterations may play a role in biomechanical behavior of subchondral bone. CONCLUSION The multiplicity of significant alterations found in patients with primary osteoarthritis and in osteoarthritic bone far away from the articular lesion indicates that osteoarthritis is part of a more generalised bone disease. These findings support the hypothesis that quantitative and qualitative differences in bone may explain the inverse relationship between osteoarthritis and osteoporosis. These quantitative and qualitative differences in bone may produce disease by increasing subchondral bone stiffness and by making subchondral bone less deformable to impact loads. This stiff bone transmits more force to overlying cartilage, making it more vulnerable. In contrast to Radin’s hypothesis (Radin et al., 1970; Radin and Rose, 1986), we hypothesize that subchondral bone stiffness is part of a more general bone alteration instead of healing of local microfractures. In line with our observations of a generalized change in bone in osteoarthritis, Doherty et al. (1983) demonstrated that the predisposition to ‘‘primary’’ osteoarthritis influences the development and severity of ‘‘secondary’’ osteoarthritis. Patients who had undergone unilateral meniscectomy 19 years or more before and who had X-ray changes in the hands consistent with generalised osteoarthritis showed more frequently a BONE DENSITY AND LOCAL GROWTH FACTORS IN OA 369 Fig. 9. Mechanical properties of iliac crest trabecular bone in women 56–80 years according to osteoarthritis grade of the hands. From Gevers et al. (1989b). more severe knee osteoarthritis, judged clinically and radiologically, in both unoperated and operated knees. Furthermore, if osteoporosis spares osteoarthritis, osteopetrosis should be a risk factor for osteoarthritis. In the few patients with osteopetrosis who reach adulthood, osteoarthritis has been reported (Milgram and Jasty, 1982). The higher body weight and associated increased subcutaneous fat in the osteoarthritis group, in addition to increased force on cartilage and bone, may also affect the bone density by preserving a better postmenopausal oestrogen status due to peripheral conversion of androstenedione to oestrone in subcutaneous fat. The inverse relationship between osteoarthritis and osteoporosis is not only of academic interest in explaining pathophysiological mechanisms of both diseases, it is also of interest in clinical practice and in decisionmaking on preventive measures. General osteoarthritis usually becomes manifest around the age of the menopause, before major bone loss occurs. Therefore, osteoarthritis could be a good negative indicator for selecting patients at risk for osteoporosis. The anthropometric, biological, and biochemical differences between primary osteoarthritis and osteoporosis reflect that systemic and metabolic factors are involved in the pathophysiology of this disease and that these common, crippling diseases are not a simple consequence of ageing. Osteoarthritis is a physiologic imbalance, a ‘‘joint failure’’ similar to ‘‘heart failure’’ in which mechanical as well as constitutional factors play a role. The initiation and progression of cartilage damage are distinct phenomena. One of the mechanisms of initiation of osteoarthritis is subchondral bone stiffness, often a part of generalized (heridited) increased bone density. Once cartilage damage is initiated, the stiffness of the subchondral bone may contribute further to progression and chondrocyte dysfunction because cartilage repair cannot follow the repeated sequence of events. From then onwards, a cascade of catabolic factors interferes because wear debris excites a synovial inflammation with protease, collagenase, interleukin 1, and prostaglandin E2 production. REFERENCES Amir, G., Pirie, C.J., Rashad, S., and Revell, P.A. (1992) Remodeling of subchondral bone in osteoarthritis: A histomorphometric study. J. Clin. Pathol., 45:990–992. Aström, J. and Beertema, J. (1992) Reduced risk of hip fracture in the mothers of patients with osteoarthritis of the hip. J. Bone Joint Surg., 74B:270–271. Batra, H.C. and Charnley, J. (1969) Existence and incidence of osteoid in osteoarthritic femoral heads. J. Bone Joint Surg., 51B:360. Baylink, D.J., Linkhart, T.A., Farley, J.R., Mohan, S., Linkhart, S.G., and Fitzsimmons, R.J. (1990) The potential role(s) of bone-derived growth factors as determinants of local bone formation, Turner syndrome. R.G. Rosenfeld, and M.M. Grumbach, eds. Marcel Dekker, New York. Baylink, D.J., Mohan, S., Linkhart, S., Linkhart, T., Fitzsimmons, R., and Farley, J. (1989) The potential role(s) of bone-derived growth factors as determinants of local bone formation. In: Turner Syndrome. R. Rosenfeld, and M. Grumbach, eds. Marcel Dekker, New York. 22:267–280. Belmonte-Serrano, M.A., Bloch, D.A., Lane, N.E., Michel, B.E., and Fries, J.F. (1993) The relationship between spinal and peripheral osteoarthritis and bone density measurements. J. Rheumatol., 20:1005–1013. Bluestone, R., Bywaters, E.G.L., Hartog, M., Holt, P.J.L., and Hyde, S. (1972) Acro-megalic arthropathy. Ann. Rheum. Dis., 30:243–258. Broom, N.D., and Poole, C.A. (1982) A functional-morphological study of the tidemark region of articular cartilage maintained in a non-viable physiological condition. J. Anat., 135:65–82. Bullough, P.G. and Jagannath, A. (1983) The morphology of the calcification front in articular cartilage. Its significance in joint function. J. Bone Joint Surg., 65:72–78. Cameron, H.U., and Fornasier, V.L. (1975) Trabecular stress fractures. Clin. Orthop., 111:266–268. Cameron, H.U. and Fornasier, V.L. (1979) Fine detail radiography of the femoral head in osteoarthritis. J. Rheumatol., 6:178–184. Canalis, E., McCarthy, T., and Centrella, M. (1988) Growth factors and the regulation of bone remodeling. J. Clin. Invest., 81:277–281. Carlsson, A., Nillson, B.E., and Westlin, N.E. (1979) Bone mass in primary cox-arthrosis. Acta Orthop. Scand., 50:187–189. Christiansen, C., Christiansen, M.S., and Transbol, I. (1981) Bone mass in postmenopausal women after withdrawal of oestrogen/ gestagen replacement therapy. Lancet, i:459. Cooper, C., Cook, P.L., Osmond, C., Fisher, L., and Cawley, M.I.D. (1991) Osteoarthritis of the hip and osteoporosis of the proximal femur. Ann. Rheum. Dis., 50:540–542. Dedrick, D.K., Goldstein, S.A., Brandt, K.D., O’Connor, B.L., Goulet, R.W., and Albrecht, M. (1993) A longitudinal study of subchondral plate and trabecular bone in cruciate-deficient dogs with osteoarthritis followed up for 54 months. Arthritis Rheum., 36:1460–1467. Denko, C.W., Boja, B., and Moskowitz, R.X. (1990) Growth promoting peptides in osteoarthritis: Insulin, insulin-like growth factor-1, growth hormone. J. Rheumatol., 17:1217–1221. 370 J. DEQUEKER ET AL. Dequeker, J., Remans, J., Franssen, R., and Waes, J. (1971) Ageing patterns of trabecular and cortical bone and their relationship. Calcif. Tissue Int., 7:23–30. Dequeker, J., Burssens, A., Creytens, G., and Bouillon, R. (1975) Ageing of bone: Its relation to osteoporosis and osteoarthrosis in postmenopausal women. Front. Horm. Res., 3:116–130. Dequeker, J., Burssens, A., and Bouillon, R. (1982) Dynamics of growth hormone secretion in patients with osteoporosis and in patients with osteoarthrosis. Hormone Res., 16:353–356. Dequeker, J., Goris, P., and Uytterhoeven, R. (1983) Osteoporosis and osteoarthritis (osteoarthrosis): Anthropometric distinctions. JAMA, 249:1448–1451. Dequeker, J., Johnell, O. and the MEDOS study group (1993a) Osteoarthritis protects against femoral neck fracture: The MEDOS study experience. Bone, 14:S51–S56. Dequeker, J., Mohan, S., Finkelman, R.D., Aerssens, J., and Baylink, D.J. (1993b) Generalized osteoarthritis associated with increased insulin-like growth factor types I and II and transforming growth factor b in cortical bone from the iliac crest. Arthritis Rheum., 36:1702–1708. Dequeker, J., Lenaerts, J., and Bouillon, R. (1994) Alteration growth hormone/IGF axis in osteoarthritis. Clin. Rheumatol., 13:163. de Sèze, S., Renier, J.C., and Rakic, P. (1962) Arthrose vertébrale et ostéoporose. Fréquences comparées de l’arthrose disco-vertébrale chez deux groupes de sujets d’âge comparable: Ostéoporotiques et non ostéoporotiques. Rev. Rhum., 29:237. Doherty, M., Watt, I., and Dieppe, P. (1983) Influence of primary generalized osteoarthritis on development of secondary osteoarthritis. Lancet, II:8–11. Duncan, H., Riddle, J.M., Jundt, J.W., and Pitchford, W. (1985) Osteoarthritis and the subchondral plate. In: Degenerative Joints. G. Verbruggen and E.M. Veys, eds. Elsevier (Biomedical Division) 2, pp. 181–191. Ekholm, R., and Norbäck, B. (1951) On the relationship between articular changes and function. Acta Orthop. Scand., 21:81. Ewald, F.C., Poss, R., Pugh, J., Schiller, A.I., and Sledge, C.B. (1982) Hip cartilage supported by methacrylate in canine arthroplasy. Clin. Orthop., 171:273–279. Farley, J.R., Tarbaux, N., Murphy, L.A., Masuda, T., and Baylink, D.J. (1987) In vitro evidence that bone formation may be coupled to resorption by release of mitogen(s) from resorbing bone. Metabolism, 36:314–321. Fazzalari, N.L., Darracoti, J., and Vernon-Roberts, R. (1985) Histomorphometric changes in the trabecular structure of a selected stress region in the femur in patients with osteoarthritis and fracture of the femoral neck. Bone, 6:125–133. Fazzalari, N.L., Vernon-Roberts, B., and Darracott, J. (1987) Osteoarthritis of the hip: Possible protective and causative roles of trabecular microfractures in the head of the femur. Clin. Orthop., 216:224– 233. Fazzalari, N.L., Moore, R.J., Manthey, B.A., and Vernon-Roberts, B. (1992) Comparative study of iliac crest and subchondral femoral bone in osteoarthritis patients. Bone, 13:331–335. Foss, M.V.L. and Byers, P.D. (1972) Bone density, osteoarthrosis of the hip and fracture of the upper end of the femur. Ann Rheum Dis., 31:259–264. Gannon, J.M., Walker, G., Fischer, M., Carpenter, R., Thompson, R.C., and Oegema, T.R. (1991) Localization of type X collagen in canine growth plate and adult canine articular cartilage. J. Orthop. Res., 9:485–494. Geusens, P., Dequeker, J., and Verstraeten, A. (1983) Age-related blood changes in hip osteoarthritis patients: A possible indicator of bone quality. Ann. Rheum. Dis., 42:112–113. Geusens, P., Vanderschueren, D., Verstraeten, A., Dequeker, J., Devos, P., and Bouillon, R. (1991) Short-term course of 1.25(OH)2D3 stimulates osteoblasts but not osteoclasts in osteoporosis and osteoarthritis. Calcif. Tissue Int., 49:168–173. Gevers, G. and Dequeker, J. (1987) Collagen and non-collagenous protein content (osteocalcin, sialoprotein, proteoglycan) in the iliac crest bone and serum osteocalcin in women with and without hand osteoarthritis. Collagen. Rel. Res., 7:435–442. Gevers, G., Dequeker, J., Geusens, P., Devos, P., and De Roo, M. (1988) Comparison of osteocalcin levels and of bone mineral content at the radius and the spine in primary osteoporosis and primary osteoarthrosis. J. Orthop. Rheumatol., 1:21–27. Gevers, G., Dequeker, J., Geusens, P., Nyssen-Behets, and C., Dhem, A. (1989a) Physical and histomorphological characteristics of iliac crest bone, according to osteoarthritis at the hand joints. Bone, 10:173–178. Gevers, G., Dequeker, J., Martens, M., and Van Audekercke, R. (1989b) Biomechanical characteristics of iliac crest bone in elderly women, according to osteoarthritis grade at the hand joints. J. Rheumatol., 16:660–663. Gotfriedsen, A., Riis, B.J., Christiansen, C., and Rodbro, P. (1990) Does a single local absorptiometric measurement indicate the overall skeletal status? Implications for osteoporosis and osteoarthritis of the hip. Clin. Rheumatol., 9:193–203. Green, W.T., Martin, G.N., Eanes, E.D., and Sokoloff, L. (1970) Microradiographic study of the calcified layer of articular cartilage. Arch. Pathol., 90:151–158. Grynpas, M.D., Alfert, B., Katz, I., Lieberman, I., and Pritzker, K.P.H. (1991) Subchondral bone in osteoarthritis. Calcif. Tissue Int., 49:20– 26. Hannan, M.T., Anderson, J.J., Zhang, Y., Levy, D., and Felson, D.T. (1993) Bone mineral density and knee osteoarthritis in elderly men and women. The Framingham Study. Arthritis Rheum., 12:1671– 1680. Hart, D.J., Mootoosamy, I., Doyle, D.V., and Spector, T.D. (1994) The relationship between osteoarthritis and osteoporosis in the general population. The Chingford Study. Ann. Rheum. Dis., 53:158–162. Havdrup, T., Hulth, A., and Telhay, H. (1976) The subchondral bone in osteoarthritis and rheumatoid arthritis of the knee. Acta Orthop. Scand., 47:345–350. Healy, J.H., Vigorita, V.J., and Lane, J.M. (1985) The coexistence and characteristics of osteoarthritis and osteoporosis. J. Bone Joint. Surg., 67A:586–592. Hochberg, M.C., Lethbridge-Cejku, M., Scott, W.W., Reichle, R., Plato, C.C., and Tobin, J.D. (1994) Serum levels of insulin-like growth factor 1 in subjects with osteoarthritis of the knee. Arthritis Rheum., 37:1177–1180. Hordon, L.D., Stewart, S.P., Troughton, P.R., Wright, V., Horsman, A., and Smith, M.A. (1993) Primary generalized osteoarthritis and bone mass. Br. J. Rheumatol., 32:1059–1061. Ingemark, B.E. (1950) The nutritive supply and nutritional value of synovial fluid. Acta Orthop. Scand., 20:144. Inove, H. (1981) Alterations in the collagen framework of osteoarthritic cartilage and subchondral bone. Int. Orthop., 5:47–52. Kellgren, J.H. (1963) Atlas of Standard Radiographs of Arthritis. In: The Epidemiology of Chronic Rheumatism. J. H. Kellgren, M. R. Jeffrey and J. Ball, eds. Blackwell, Oxford, 2, pp. 1–23. Koszyca, B., Fazzalari, N.L., and Vernon-Roberts, B. (1989) Trabecular microfractures. Nature and distribution in the proximal femur. Clin. Orthop., 244:208–216. Kusakabe, A. (1977) Subchondral cancellous bone in osteoarthritis and rheumatoid arthritis of the femoral head. Arch. Orthop. Unfall. Chir., 88:185–197. Lane, L.B. and Bullough, P.G. (1980) Age-related changes in the thickness of the calcified zone and the number of tidemark in adult human articular cartilage. J. Bone Joint Surg., 62:372–375. Lane, N.E. and Nevitt, M.C. (1994) Osteoarthritis and bone mass. J. Rheumatol., 21:1393–1396. Lane, L.B., Aquiles, V., and Bullough, P.G. (1977) The vascularity and remodeling of subchondral bone and calcified cartilage in adult human femoral and humeral heads. J. Bone Joint Surg., 59:272– 278. Layton, M.W., Goldstein, S.A., Goulet, R.W., Feldkamp, L.A., Kubinski, D.J., and Bole, G.G. (1988) Examination of subchondral bone architecture in experimental osteoarthritis by microscopic computed axial tomography. Arthritis Rheum., 31:1400–1405. Lemperg, R. (1971) The subchondral bone plate of the femoral head in adult rabbits. I: Spontaneous remodeling studied by microradiography and tetracycline labelling. Virchows Arch., 352:1–13. Littlejohn, G.O., Hall, S., Brand, C.A., and Davidson, A. (1986) New bone formation in acromegaly: Pathogenetic implications for diffuse idiopathic skeletal hyperostosis. Clin. Exp. Rheumatol., 4:99–104. Madsen, O.R., Schaadt, O., Bliddal, H., Egsmose, C., and Sylvest, J. (1994) Bone mineral distribution of the proximal tibia in gonarthrosis assessed in vivo by photon absorption. Osteoarthritis Cartilage, 2:141–147. McAlindon, T.E., Teale, D., and Dieppe, P.A. (1992) Insulin-like growth factor 1: Effect of age accounts for apparent correlation with sclerosis and osteophytosis in osteoarthritis of the knee (abstract). Br. J. Rheumatol., 31:213. Milgram, J.W. and Jasty, M. (1982) Osteopetrosis—A morphological study of twenty-one cases. J. Bone Joint Surg., 64A:912–929. Minns, R.J. and Stevens, F.S. (1977) The collagen fibril organization in human articular cartilage. J. Anat., 123:437–457. Mohan, S., Jennings, J.C., Linkhart, T.A., and Baylink, D.J. (1986) Chemical and biological characterization of low molecular weight human skeletal growth factor. Biochim. Biophys. Acta, 884:234–242. Mokassa Bakumobatane, L., Dequeker, J., Raymaekers, G., and Aerssens, J. (1993) Effect of osteoarthritis (OA) and body weight on BONE DENSITY AND LOCAL GROWTH FACTORS IN OA subchondral cancellous bone quality of proximal tibia. Osteoarthritis Cartilage, 1:55–56. Moore, R.J., Fazzalari, N.L., Mantey, B.A., and Vernon-Roberts, R. (1994) The relationship between head-neck-shaft angle, calcar width, articular cartilage thickness and bone volume in arthrosis of the hip. Br. J. Rheumatol., 33:432–436. Nevitt, M.C., Scott, J.C., and Lane, N.E. (1992) Hip osteoarthritis and bone mineral density in older white women (Abstract). Arthritis Rheum., 35:S42. Oettmeier, R., Abendroth, K., Oettmeier, S. (1989) Analyses of the tidemark on human femoral heads. Acta Morphol. Hung., 37:169– 180. Pogrund, H., Rutemberg, M., Makin, M., Robin, G., Menczel, J., and Steinberg, R. (1982) Osteoarthritis of the hip joint and osteoporosis: A radiological study in a random population sample in Jerusalem. Clin. Orthop., 164:130–135. Pugh, J.W., Radin, E.L., and Rose, R.M. (1974) Quantitative studies of human subchondral cancellous bone: Its relationship to the state of its overlying cartilage. J. Bone Joint Surg., 56A:313–321. Radin, E.L. and Paul, I.L. (1970) Does cartilage compliance reduce skeletal impact loads? Arthritis Rheum., 13:139–144. Radin, E.L. and Rose, R.M. (1986) Role of subchondral bone in the initiation and progression of cartilage damage. Clin. Orthop. Rel. Res., 213:34–40. Radin, E.L., Paul, I.L., and Tolkoff, M.J. (1970) Subchondral bone changes in patients with early degenerative joint disease. Arthritis Rheum., 12:400–405. Radin, E.L., Parker, H.G., Pugh, J.W., Steinberg, R.S., Paul, I.L., and Rose, R.M. (1973) Response of joints to impact loading. III Relationship between trabecular microfractures and cartilage degeneration. J. Biomech., 6:51–57. Raymaekers, G., Aerssens, J., Van den Eynde, R., Peeters, J., Geusens, P., Devos, P., and Dequeker, J. (1992) Alterations of the mineralization profile and osteocalcin concentrations in osteoarthritic cortical iliac crest bone. Calcif. Tissue Int., 51:269–275. Rees, J.A. and Ali, S.Y. (1988) Ultrastructural localisation of alkaline phosphatase activity in osteoarthritic human articular cartilage. Ann. Rheum. Dis., 47:747–753. Reinmann, I. and Christensen, S.B. (1979) A histochemical study of alkaline and acid phosphatase activity in subchondral bone from osteoarthritic human hips. Clin. Orthop., 140:85–91. Reinmann, I., Mankin, H.K., and Trahan, C. (1977) Quantitative histologic analysis of articular cartilage and subchondral bone from osteoarthritic and normal human hips. Acta Orthop. Scand., 48: 63–73. Revell, P.A., Pirie, C., Amir, G., Rashad, S., and Walker, F. (1990) Metabolic activity in the calcified zone of cartilage: Observations on 371 tetracycline labelled articular cartilage in human osteoarthritic hips. Rheumatol. Int., 10:143–147. Roh, Y.S., Dequeker, J., and Mulier, J.C. (1974) Bone mass in osteoarthrosis, measured in vivo by photon absorption. J. Bone Joint Surg., 56A:587–591. Schouten, J.S.A.G., Van den Ouweland, F.A., Valkenburg, H.A., and Lamberts, S.W. (1993) Insulin-like growth factor-1: A prognostic factor of knee osteoarthritis. Br. J. Rheumatol., 32:274–280. Schubert, M., Hamerman, D., et al. (1968) Primer in Connective Tissue Biochemistry. Lea and Febinger, Philadelphia. Simkin, P.A., Graney, D.O., and Fiechtner, J.J. (1980) Roman arches, human joints, and disease: Differences between convex and concave sides of joints. Arthritis Rheum., 23:1308–1311. Simon, S.R., Radin, E.L., and Paul, I.L. (1972) The response of joints to impact loading. II. In vivo behaviour of subchondral bone. J. Biomech., 5:267–272. Smith, R.W. and Rizek, J. (1966) Epidemiologic studies of osteoporosis in women of Puerto Rico and Southeastern Michigan with special reference to age, race, national origin and to other related or associated findings. Clin. Orthop., 45:31–48. Sokoloff, L. (1969) The Biology of Degenerative Joint Disease. University of Chicago Press, Chicago, pp. 162. Stougård, J. (1974) The calcified cartilage and the subchondral bone under normal and abnormal conditions. Acta Pathol. Microbiol. Scand., 82:182. Todd, R.C., Freeman, M.A.R., and Pirie, C.J. (1972) Isolated trabecular fatigue fractures in the femoral head. J. Bone Joint Surg., 54B:723. Trueta, J. and Harrison, M.H.M. (1953) The normal vascular anatomy of the femoral head in adult man. J. Bone Joint Surg., 35B:442. Urist, M.R. (1960) Observations bearing on the problem of osteoporosis. In: Bone as a Tissue. K. Bodahl, ed. McGraw-Hill, New York, pp. 18–23. Vandermeersch, S., Geusens, P., Nijs, J., and Dequeker, J. (1990) Total body mineral measurement in osteoarthritis, osteoporosis and normal controls. In: Current Research in Osteoporosis and Bone Mineral Measurement. E.F. Ring, ed. British Institute of Radiology, London, pp. 49. Verstraeten, A., Van Ermen, H., Haghebaert, G., Nijs, J., Geusens, P., and Dequeker, J. (1991) Osteoarthrosis retards the development of osteoporosis. Clin. Orthop., 264:169–177. Weintroub, S., Papo, J., Ashkenazi, M., Tardiman, R., Weissman, S.L., and Salama, R. (1982) Osteoarthritis of the hip and fracture of the proximal end of the femur. Acta Orthop. Scand., 53:261–264. Woods, C.G., Greenwald, A.S., and Haynes, D.W. (1970) Subchondral vascularity in the human femoral head. Ann. Rheum. Dis., 29:138– 142.