Trabecular bone microarchitecture: A review La microarchitecture de l os trabéculaire : une revue

Transcription

1 Morphologie (2008) 92, GENERAL REVIEW Trabecular bone microarchitecture: A review La microarchitecture de l os trabéculaire : une revue D. Chappard, M.-F. Baslé, E. Legrand, M. Audran Inserm, U922, LHEA, Faculty of Medicine, 1, rue Haute-de-Reculée, Angers cedex, France Available online 18 November 2008 KEYWORDS Trabecular bone; Microarchitecture; Histomorphometry; MicroCT; Remodeling MOTS CLÉS Os trabéculaire ; Microarchitecture ; Histomorphométrie ; Microtomographie ; Remodelage Summary The bone mass is constituted during the life by the modeling and remodeling mechanisms. Trabecular bone consists in a network of trabeculae (plates and rods) whose distribution is highly anisotropic: trabeculae are disposed parallel to the resultant of stress lines (Wolff s law). Trabecular microarchitecture appears conditioned by mechanical strains, which are exerted on the bones of the skeleton. However, few methods are currently clinically validated to appreciate and follow the evolution of microarchitecture in bone diseases. The most developed studies relate to microarchitectural measurements obtained by bone histomorphometry with the use of new algorithms, which can appreciate 2D various characteristics of the trabeculae, such as thickness and connectivity. Several works have shown that microarchitecture parameters should be obtained by using several independent techniques. X-ray microtomography (microct), micro- RMI, synchrotron also allow the measurement in 3D of the trabecular microarchitecture in a nondestructive way on bone specimens. This review describes the evolution of our knowledge on bone microarchitecture, its role in bone diseases, such as osteoporosis and the various methods of histological evaluation in 2D and 3D Elsevier Masson SAS. All rights reserved. Résumé Le capital osseux se constitue au cours de la vie par les mécanismes de modelage et de remodelage. Le tissu trabéculaire est constitué par un ensemble de travées (plaques et piliers), dont la répartition est hautement anisotrope : les travées se disposent parallèlement à la résultante des lignes de contraintes (loi de Wolff). La microarchitecture trabéculaire apparaît conditionnée par les contraintes mécaniques qui s exercent sur les pièces squelettiques. Cependant, peu de méthodes sont actuellement validées cliniquement pour apprécier et suivre l évolution de la microarchitecture dans les ostéopathies. Les études les plus développées portent sur l appréciation microarchitecturale par histomorphométrie osseuse grâce à l utilisation de nouveaux algorithmes permettant d apprécier en 2D différentes caractéristiques trabéculaires, dont la connectivité. Plusieurs travaux ont montré que l appréciation de la microarchitecture devait utiliser plusieurs techniques indépendantes. La microtomographie X (microct), la micro-irm, le synchrotron permettent aussi de mesurer en 3D l architecture Corresponding author. address: daniel.chappard@univ-angers.fr (D. Chappard) /$ see front matter 2008 Elsevier Masson SAS. All rights reserved. doi: /j.morpho

2 Trabecular bone microarchitecture: A review 163 trabéculaire de façon non destructive sur des prélèvements osseux. Cette revue décrit l évolution des connaissances sur la microarchitecture osseuse, son rôle dans les maladies osseuses, comme l ostéoporose et les différentes méthodes d évaluation histologique en 2D et en 3D Elsevier Masson SAS. All rights reserved. Introduction Bone, as a connective tissue, is in constant adaptation during all stages of life and two fundamental processes control the bone mass. Modeling allows the acquisition of the form and the mass of bone during childhood and adolescence. It ensures the growth of the skeletal pieces in length. The osteoblastic activity is prevalent on the osteoclastic activity (and can occur even in the absence of a resorption step). The modeling activity occurs in the primary spongiosa, under the growth plates when spicules of calcified cartilage are covered with bone. In membranous bones (such as the skull), osteoblast directly elaborate bone without a preliminary cartilaginous anlagen. The modeling phase allows to reach the peak bone mass towards 25 to 30 years. Modeling can persist in the adult in some circumstances: fracture healing, benign or malignant bone metaplasia and during the primary phase of adaptation of bone to biomaterials. However, areas of minimodeling can be observed occasionally in adults and are due to a direct bone apposition without preliminary intervention of osteoclasts. Remodeling allows the adaptation of bone to local variations of strains, to hormonal and metabolic changes during all the rest of the life. In this case, remodeling is accomplished by the coordinated action of osteoclasts and osteoblasts acting in concert in the form of basic multicellular remodeling units (BMU). Bone remodeling is characterized by a phase of resorption of a piece of bone tissue that had become mechanically (or metabolically) less competent. This phase precedes the formation of a new bone structure unit (BSU), which is better adapted to the new conditions. Osteoblasts elaborate BSU differently in cortical bone (where they are in the form of cylindrical osteones) and in trabecular bone (where these bone packets form curved structures, alike incomplete osteones). BSU themselves are arranged and packed in a different way in cortical bone and in trabecular bone (where they form trabeculae). This level of organization is known as the bone microarchitecture. A higher degree of organization is represented by macroarchitecture of bones themselves: they possess angulations, curvatures, which enable them to adapt to strains and gravity. The human skeleton is composed of approximately 20% of trabecular bone and 80% of cortical bone. Trabecular bone has a remodeling level higher than that of cortical bone. Its 3D microarchitecture is directly conditioned by the mechanical strains, which are exerted on it. During childhood, trabecular bone is primarily made of a dense network of plates with, often, an isotropic 3D repartition (i.e., uniformity in all directions). In adults, a preferential orientation of the trabecular plates is observed along the direction of the strains, which are exerted on the bone. It is noticeable that until the 1980s, bone microarchitecture was not taken into account in the comprehension of the pathophysiological mechanisms of metabolic bone diseases; osteoporosis was only considered as a disease associated with a reduction in the bone mass. At the present time, definition of the osteoporosis, according to the WHO and International Osteoporosis Fundation (IOF), is a systemic skeletal disease characterized by low-bone mass and microarchitectural deterioration of bone tissue, leading to enhanced bone fragility and a consequent increase in fracture risk [1]. The role of bone microarchitecture History Trabecular bone microarchitecture is now recognized as an important component of bone quality. Other factors are the remodeling level, the size of hydroxyapatite crystals, the quality of collagen [15]. The biomechanical value of bone as a biomaterial depends on both its volume and also on its adapted distribution in the 3D space, that is, its microarchitecture. The Scottish naturalist, John Hunter ( ), discovered by examining animal mandibles, there coexist, at the same time, areas of bone destruction and areas of bone apposition and also that bone destruction preceded bone apposition. In 1867, the German anatomist Hermann von Meyer received a grant of the Prussian government to study skeletal posture. During his research, he studied the orientation of bone trabeculae in the upper femoral extremity and found it similar in all subjects. At the time von Meyer presented his scientific results, a Swiss engineer (Karl Culmann) realized that the drawings representing the spatial distribution of bone trabeculae in the femoral epiphysis had an astonishing resemblance to the strain lines, which he had used for designing a crane. Culmann and von Meyer postulated together a theory saying that the direction of bone trabeculae coincides with the strain lines. Julius Wolf, a German anatomist, confirmed that not only bone trabeculae were aligned with the directions of the strains, but that their orientations are modified if the strains are changed. For the first time, Wolf realized that bone microarchitecture represented an adaptative answer to mechanical variations. Such adaptability of bone microarchitecture allows a considerable reduction of the bony-material mass necessary to support the

3 164 D. Chappard et al. Figure 1 Microarchitecture of cortical bone, aspect in scanning electron microscopy. A. The cortice is porous and made of osteones compacted together, the Haversian canals are disposed parallel to the long axis of the bone, according to the maximum strains direction. The lamellae have been made visible by acid etching (the bars stands for 100 m). B. The Haversian canals are evidenced by microct in a semitransparent reconstruction of a piece of femoral-neck cortex. C. Microarchitecture of trabecular bone, in a high-definition radiograph of a human calcaneus showing the arrangement in arches of the trabeculae. Microarchitecture de l os cortical, aspect en microscopie électronique à balayage. A. L os cortical est poreux et composé d osteons compactés ; les canaux de Havers sont disposés parallèlement au grand axe de l os long (fémur), selon la direction des contraintes maximales. Les lamelles ont mises en évidence par une gravure à l acide (la barre représente weight of the individual. The theory of adaptation of bone microarchitecture to strains is now known as the Wolf s law. The German surgeon Wilhem Roux suggested for the first time that cells contained inside bone could perceive and provide an answer to the mechanical stresses by controlling bone resorption and apposition. The American anatomist John C. Koch showed, in 1917, that bone microarchitecture was optimized: he showed that bone density was higher in the areas where strains are maximum and he suggested that bone mass is distributed in the best way to obtain a maximum biomechanical adaptation for a minimal material mass. Then, these concepts were frequently under-recognized, although some reports by scanning electron microscopy were published from time to time [40,41]. The orthopedic surgeon Harold M. Frost, of the Henry Ford s Hospital in Detroit, showed that adaptation of bone microarchitecture is due to bone remodeling. He proposed the first cellular theories implied in bone remodeling and showed the differences between bone modeling and remodeling. He determined that osteoclastic and osteoblastic activities were coupled in time and space and proposed the BMU theory [18]. He showed how strain intensity controlled the nativity and life of BMU (structural adaptations to mechanical use: SATMU) according to the mechanostat theory [16,17]. However, the interdependence of strains and bone remained under-recognized during decades; a rediscovery of the importance of bone microarchitecture occurred in the 1980s, in particular on the impulse of Michael A. Parfitt. He was the first to propose a mathematical model for bone microarchitecture by recognizing the importance of anisotropy, that is, the oriented distribution in space. In dense cortical bone, osteones are packed such that axes of all Haversian canals are parallel to the main strain exerted on the bone shaft (Fig. 1A and B). In trabecular bone, the bony material realize a network composed of plates parallel to the strain lines and connected by transverse rods or pillars that ensure the cohesion of the whole system. The trabecular microarchitecture is evidenced on X-ray films: plates are vertical in bone submitted to uniaxial strains (i.e., in the vertebrae, in the tibial plateau) and create arches when there are strains in various directions (i.e., femoral head, calcaneus) (Fig. 1C). Parfitt et al. also proposed a set of stereological techniques to measure trabecular microarchitecture on bone biopsies [33]. However, these techniques are based on the assumption that all trabeculae in the different bones are either in the form of plates or rods ( the plate and rod model). One should note that the distribution of trabeculae finside the various bones of the skeleton was recognized to be different [35]. 100 m). B. Les canaux de Havers sont visualisés en microct dans un fragment d os cortical provenant d un col fémoral. La matrice osseuse a été rendu semi-transparente de façon à bien mettre en évidence les canaux de Havers. C. Microarchitecture de l os trabéculaire dans une radiographie d un calcaneus humain montrant l arrangement en voûte des travées.

4 Trabecular bone microarchitecture: A review 165 Parameters currently used to measure trabecular bone microarchitecture The most classical histomorphometric parameters currently used for the description of trabecular bone microarchitecture are all based on Parfitt s principles of the plate and rod model. These microarchitecture descriptors are: trabecular thickness (Tb.Th, in microns), trabecular number (or more exactly trabecular density) (Tb.N in per millimetre) and the trabecular separation (Tb.Sp, in microns). These parameters are derived from a combination of measurements from trabecular surfaces and perimeters and can be measured with microscopes equipped with oculars containing specific reticules or with image analyzers [32]. With the development of microcomputers and the growing interest for stereology (a branch of mathematics concerned with the relationships of 2D 3D), several robust methods have been proposed to study trabecular microarchitecture by using methods independent of surface and perimeter measurements (i.e., independent of the plate and rod model ) on digitized images of bone (Fig. 2A). The trabecular bone pattern factor (Tb.P f ) This method was proposed by Hahn et al. and is based on the use of mathematical morphology [21]. The principle of the method rests on the fact that, in a perfectly connected structure, concave surfaces are abundant, whereas convex surfaces are very few. Conversely, in a disconnected structure, concave surfaces become less abundant and convex surfaces are then more numerous. Practically, this method is done by image analysis, performing a dilation of the original image of the trabecular network and measuring perimeters and surface areas before and after dilation. Thus, when a trabecular network is highly disconnected, the convex surfaces are increased and the dilatation process increases the perimeter drastically, while the area is only moderately affected. This provides low values of Tb.P f in a well-connected network and high values when marked disconnection of trabeculae is present. Interconnectivity index (ICI) The method was originally proposed by Le et al. to describe the connectivity of porous biomaterials, such as corals [24]. Corals were proposed as substitute biomaterials for bone grafting since some of them have an exocellular calcium carbonate matrix that mimics trabecular bone [20]. When applied to trabecular bone, connectivity of the marrow cavities can be appreciated on digitized images by taking the skeletons of their profiles (i.e., the watershed line). Since the skeletonization process is very sensitive to local variations of the boundaries, which produce undesirable dendrites, the skeleton must be pruned to remove the aberrant terminal ends. On the pruned skeleton, the total number of nodes, node-to-node branches, node-to-free end branches are determined. Also, the number of trees is obtained, a tree being the structure composed of interconnected node(s) with node-to-node and/or node-to-free end branch(es). The ICI of the bone-marrow cavities combines these parameters as previously shown [8]. So, an increased connectivity of marrow cavities (given by a high number of nodes and segmental branches associated with a low number Figure 2 Measurement of trabecular bone microarchitecture by histomorphometric methods based on mathematical morphology. A. Digitized image of a bone biopsy showing the cortices and the trabecular network. B. The strut analysis simplifies the trabeculae to the watershed, trabeculae are reduced to a line and connection nodes are figures by a dot. C. Star volume obtained by sending a seed in the marrow cavities and the rays of the star stop when touching a bone surface. Mesure de la microarchitecture de l os trabéculaire par des méthodes histomorphométriques basées sur la morphologie mathématique. A. Image digitalisée d une biopsie osseuse montrant les corticales et le réseau trabéculaire. B. Analyse des travées après réduction des travées en une «ligne de partage des eaux», les nœuds de raccordement sont figurés par un point. C. Méthode du star volume obtenue en plaçant un point de façon aléatoire dans les cavités médullaire et en traçant des rayons qui s arrêtent lorsqu ils touchent une surface trabéculaire ou endostéale.

5 166 D. Chappard et al. of trees) increases the ICI and corresponds to a fragmentation of the trabecular-bone network. Characterization of the trabecular network (strut analysis) This technique is based on work of Compston and Dempster [11,13]. A method of image skeletonization is applied to the trabeculae with the same algorithms as described for ICI determination. This allows an identification of the various types of trabeculae and a measurement of their connectivity. Anastomoses or nodes between trabeculae are identified; the trabeculae ending with a free terminus are measured, as well as those connecting two nodes or trabeculae connected with the cortices (Fig. 2B). Each type of strut is allocated to a different color, thus, providing a visual characterization of the whole trabecular network. In order to obtain a single parameter for easy handling, the node to free-end ratio is determined (N/F) [12]. The star volume This method was largely exploited for the measurement of porous materials, in particular of cements and rocks [28]. It is based on the analysis of the marrow spaces: starting from a randomly placed seed in the marrow cavity, one can project rays in all the directions of space (Fig. 2C) [38]. The rays stop as soon as they meet a trabecular surface, a cortice or edges of the section. This constitutes a kind of star and the measurement of the length of each ray of star is done. One sees that the more the network will be disconnected, the more the length of the rays will be important. If a great number of stars is generated, small perforations are highlighted inside the network trabecular. However, this method is very time consuming and the grid technique is preferred (this method is sometimes referred as the maximum intercept length [MIL]): a series of grids are computed with parallel lines running with various angles running from 0to2. Each grid is stored on the hard disk computer and is intersected with the image of the marrow cavities. This provides linear segments (called chords) superimposed on the marrow spaces. The cubed length of each chord is then computed with each grid, so that all directions from 0 to 360 are explored very rapidly. The star volume can be determined on the marrow spaces (a high star volume indicates a highly fragmented-bone network); it can also be used on the trabeculae to know their average size. Euler-Poincaré s number The method consists in counting the number of particles n present in space trabecular and the number of marrow cavities circumscribed by the trabeculae m. The number of Euler, E = n m. Plus the network trabecular is connected, plus E is weak (even negative). E must be adjusted according to the surface of space trabecular. The more connected is the trabecular network, the less is E; in this way, negative values are obtained in highly connected systems. The maximum traveling pathway within the network The trabecular network can be seen as a labyrinth in which one can move in all directions. The more perforations are disconnecting the trabecular network (thus, merging the marrow cavities), the longer is the pathway [37]. The fractal dimension of the trabecular network Biological objects have an irregularity and a complexity, which are often difficult to quantify by the Euclidean geometry. The fractal analysis makes it possible to approach the complexity of structures and curves. On a 2D section, perimeter of the trabeculae can occupy more or less the section surface according to the complexity of branching, connection and more or less regular disposition in space... The fractal dimension D can be measured by the boxes counting method, which consists in superimposing on an image of the trabecular boundaries, a series of grids made of similar squares of side (and mimicking a chessboard) [14]. The number of squares which intercept the trabecular boundaries is measured N( ) and one starts again with a new grid with a larger. D is obtained by determining the slope of the straight regression line between log N( ) and log (this is called a Richardson s plot). The slope D of the regression line corresponds to the Kolmogorov fractal dimension. One can measure fractal dimension by other methods: a progressive dilation of the trabecular perimeters provides the Minskowski-Bouligand D, the mass radius/lacunarity is obtained by positioning circles of growing diameter and counting the number of pixels of the boundaries that are covered [7]. All these methods showed are not self-exclusive and none of them is sufficient to provide a unique parameter that fully describes microarchitecture. We found that it is always necessary to use several techniques in association to characterize a network trabecular because these different techniques do not explore the same components of microarchitecture [7,8]. We found, by using a hierarchical cluster analysis that three groups of clusters can be identified: one describing the size of the trabeculae, one describing the medullar cavities and the later corresponding to the branchings of trabeculae. It becomes easy to see that, according to pathophysiological mechanisms, which are implied in the genesis of the various types of bone loss, certain parameters can appear more sensitive than others at earlier times [3,8]. In animal models, measurements of bone microarchitecture were found more early altered than bone volume (appreciated by histomorphometry, ash weight and densitometry) [29]. Nonlinear correlations (e.g., hyperbolic or exponential) are frequently met when studying the relationships between bone mass (BV/TV) and the various architectural parameters. However, certain parameters (D, E, Tb.Th) are linearly correlated with bone volume. The inflexion point of the hyperbolic curves corresponds to a mean Bone Volume of 11%, a threshold previously assigned by histomorphometric studies to represent a value associated with spontaneous fractures [30]. We found that this threshold corresponds to a network whose trabeculae have a mean thickness of 70 m. Osteoclasts are able to resorb bone at an average depth of 40 m: so when osteoclasts erode a thinned trabecula circumferentially, a perforation can occur and a pillar is sectioned. This mechanism was evidenced in corticosteroidinduced osteoporosis [6] and was also reported in male osteoporosis [26]. In males, it is estimated that bone resorption leads to a progressive reduction in the trabecular thickness upon aging, but the trabecular microarchitecture and connectivity remain roughly preserved. As soon as

6 Trabecular bone microarchitecture: A review 167 perforations can occur, they have considerable deleterious consequences on the biomechanical value of the trabecular network. With the development of microcomputers, new measuring instruments have recently appeared allowing the 3D analysis of bone. Micro-MRI are now under development and X-ray microtomography (microct), which allow a fast exploration and 3D measurement of bone, samples are available. MicroCT have also been developed for in vivo analysis of small animals [39] and for analysis of peripheral bones, such as the radius and tibia in man [36]. The synchrotron is also an interesting tool, but it is a very limited technique due to high costs, difficult accessibility and the necessity to evaluate small samples [34]. The standard radiographs, which are used as a routine clinical exam, represent a 2D projection of the trabecular bone on a silver film or an X-ray sensitive sensor. Several groups have tried to appreciate microarchitectural disorders by describing methods based on image analysis. Although the 3D 2D relationships are not fully understood at the present time, it is likely that there is an interdependence between 2D histomorphometric parameters, 3D microct data and analysis of plain radiographs [4,19]. Texture analysis of the trabecular network can be appreciated on radiographic images and appears an interesting and cheap method of appreciation of bone microarchitecture [27]. Techniques of images analysis proposed are based on the study of run lengths, co-occurrence matrix, heterogeneity of the pixels grey levels and algorithms based on fractal geometry (skyscrapers method, blankets, Hurst...). MicroCT allows the measurement of trabecular volume and trabecular characteristics directly in 3D. In addition, new parameters have been developed: structure model index (SMI) is close to 0 if the trabecular network is mainly composed of plates, near 3 if rods are dominating; the degree of anisotropy evaluates the orientation of trabeculae in space; the distribution frequency of the trabecular thicknesses and marrow cavities width is also of the utmost interest; new indices of connectivity have also been proposed (connectivity density, MIL). However, microct is only at its beginnings and there is yet no standardization of parameters between manufacturers; furthermore, the validity of some algorithms is Figure 3 MicroCT imaging of iliac bone. A. In a young subject: the cortices are thick, the trabecular network is dense. B. In a postmenopausal osteoporosis, note the holes inside the network corresponding to areas of loss of connectivity. C. In a male with idiopathic osteoporosis, note the conversion of plates into rods, although the connectivity is rather well even if trabeculae are thin. D. In a male with osteoporosis due to multiple risk factors (alcoholism and glucocorticoid treatment). Note the thinning of the cortices, the considerable disorganization of the trabecular network with area without trabeculae. Aspect en microct de l os iliaque. A. Chez un jeune sujet : les corticales sont épaisses et le réseau trabéculaire est dense. B. Dans une ostéoporose postménopausique, notez les perforations à l intérieur du réseau correspondant aux zones de perte de connectivité. C. Dans une ostéoporose masculine idiopathique, notez la conversion des plaques en piliers et le maintient d une bonne connectivité, même si les travées sont amincies. D. Dans une ostéoporose masculine idiopathique, due à plusieurs facteurs de risque (alcoolisme et traitement par glucocorticoïdes). Notez l amincissement considérable des corticales, la désorganisation maximale du réseau trabéculaire et la disparition complète des travées dans certaines zones.

7 168 D. Chappard et al. questionable. In particular, it appears that algorithms used for thickness measurements are strongly influenced by the shape of the measured objects themselves: nodes (zones of connectivity between trabeculae) have a considerable influence on thickness values and can lead to erroneous conclusions [5,9]. MicroCT has also the disadvantage of providing no information on cellular activities, but offers the advantage of being nondestructive for bone samples, which can thus be processed by conventional histological methods. Alterations of trabecular microarchitecture During ageing In the young, the trabecular network is dense and plexiform. During ageing, bone trabeculae are thinned because of a constant osteoblastic depression. Like any connective cell, osteoblasts encounter a reduced capacity for matrix synthesis (collagen and non-collagenic proteins regularly decrease), leading to a progressive conversion of the plates into rods (Fig. 3A and B). In female At the menopause, the estrogenic deficiency leads to an increase in several cytokines (IL-6, IL-7, TNF...), leading to a stimulation of the osteoclastic activity [10]. The number of activated BMU increases and osteoclasts can erode trabeculae since their lifespan is increased (inhibition of apoptosis) [22]. The consequence in an increased in the number of trabecular perforations (Fig. 3C), which disorganize the 3D microarchitecture [23]. Some authors have proposed the term killer osteoclasts to indicate this phenomenon responsible for an acceleration of bone remodeling with a bone loss, which can reach 2% per year during the menopausal period leading to a 20 to 30% reduction of the initial bone mass. In male In male, etiologic factors which condition bone loss are multiple, intricate and the diagnosis is often much more complex than in female. The peak bone mass is acquired just at the beginning of the third decade and is approximately 30% higher than that of females because of the largest diameter of the bones. Prepubertal growth, which largely conditions the size of bones, is on average two years longer in man. The long bones diameter and vertebral surfaces are thus approximately 20 to 25% higher, contributing to an increased biomechanical resistance. Tomodensitometric studies have shown that trabecular-bone loss during ageing is similar in both genders, particularly in the spine (if one excepts the menopausal period). Evolution of long bones is, on the other hand, different: cortical porosity and endosteal resorption, which occur in both genders, are reduced in man while periosteal apposition is more important. So the cross section of long bones, a factor of biomechanical resistance, increases with the age in man, particularly at the femoral neck. Evolution with the age of bone trabecular microarchitecture seems also different in normal man as evidenced using histomorphometric methods. Perforations of bone trabeculae are less frequent in man so that connectivity is better preserved in the males [31]. The study of the bone microarchitecture by histomorphometry has brought important data to knowledge of the pathophysiology of the disease. During primitive or secondary osteoporosis, there is a marked reduction of the trabecular-bone volume, in the thickness and number of trabeculae, but also a marked disorganization of the trabecular microarchitecture: a reduction in the number of nodes (on the strut analysis), an increase in the number of free ends (bone trabeculae with end termini), an increase of ICI and star volume of the bone-marrow spaces. An important and sudden rise in these architectural indices occurs when the trabecular bone volume becomes lower than 11% or the thickness of the trabeculae lower than 70 m [6]. A logistic regression study, done in 108 osteoporotic males, confirmed that the relative risk of vertebral fractures was strongly influenced by the quality of bone microarchitecture [26]. After adjustment for age, body mass index and femoral density, the occurrence of fractures remained associated with microarchitectural indices, with relative risks varying between 1.7 and 3.2 for one S.D. variation of the connectivity parameters. According to etiologic factors, it seems that microarchitectural abnormalities of the trabecular network are different [2]. In glucocorticoid-induced osteoporosis, trabeculae become thinner, but the overall connectivity is preserved. However, specific aspects of plate thinning can be observed in microct with multiple minute perforations occurring in their center [5]. In hypogonadic patients, the trabeculae remain thick, but disorganized by holes in the network. In male osteoporosis, the deleterious effect of risk factors is additive on bone microarchitecture. For example, a subject presenting a hypogonadism and alcoholism will have a microarchitecture much more altered than a subject presenting only one of these risk factors. The number of fractures parallels the number of fracture risks as well [25]. Conclusion Trabecular microarchitecture is a very important aspect of bone fragility. It is now recognized as one of the bone-quality factors that can explain the occurrence of fractures. A survey of the scientific literature in the Medline database indicates that the number of papers including bone microarchitecture or bone structure is growing exponentially. Seven papers included one of these key words in 1970, thirty in 1987 (when the first papers dealing with microarchitecture were published) to 515 in Aside to microarchitecture, other factors are now recognized and assembled under the umbrella term of bone quality [15]. Bone quality includes the remodeling rate, the mineralization degree and its heterogeneity, the size of hydroxyapatite crystals, the collagen and non-collagenous proteins composition, the osteocyte viability and the microor nanomechanical resistance. It is likely that abnormalities in any of these factors can provoke bone fragility by acting either at tissular, cellular or molecular level.

8 Trabecular bone microarchitecture: A review 169 References [1] Anonymous. Consensus development conference. Prophylaxis and treatment of osteoporosis. Am J Med 1993;94: [2] Audran M, Chappard D, Legrand E, Libouban H, Baslé MF. Bone microarchitecture and bone fragility in men: DXA and histomorphometry in humans and in the orchidectomized rat model. Calcif Tissue Int 2001;69: [3] Chappard D, Alexandre C, Riffat G. Spatial distribution of trabeculae in iliac bone from 145 osteoporotic females. Acta Anat (Basel) 1988;132: [4] Chappard D, Guggenbuhl P, Legrand E, Baslé MF, Audran M. Texture analysis of X-ray radiographs is correlated with bone histomorphometry. J Bone Miner Metab 2005; 23:24 9. [5] Chappard D, Josselin N, Rougé-Maillart C, Legrand E, Baslé MF, Audran M. Bone microarchitecture in males with corticosteroid-induced osteoporosis. Osteoporos Int 2007;18: [6] Chappard D, Legrand E, Baslé MF, Fromont P, Racineux JL, Rebel A, et al. Altered trabecular architecture induced by corticosteroids: a bone histomorphometric study. J Bone Miner Res 1996;11: [7] Chappard D, Legrand E, Haettich B, Chalès G, Auvinet B, Eschard JP, et al. Fractal dimension of trabecular bone: comparison of three histomorphometric computed techniques for measuring the architectural two-dimensional complexity. J Pathol 2001;195: [8] Chappard D, Legrand E, Pascaretti C, Audran M, Baslé MF. Comparison of eight histomophometric methods for measuring trabecular bone architecture by image analysis on histological sections. Microsc Res Tech 1999;45: [9] Chappard D, Retailleau-Gaborit N, Legrand E, Baslé MF, Audran M. Comparison insight bone measurements by histomorphometry and microct. J Bone Miner Res 2005;20: [10] Cohen-Solal ME, Boitte F, Bernard-Poenaru O, Denne MA, Graulet AM, Brazier M, et al. Increased bone resorbing activity of peripheral monocyte culture supernatants in elderly women. J Clin Endocrinol Metab 1998;83: [11] Compston JE. Connectivity of cancellous bone: assessment and mechanical implications. Bone 1994;15: [12] Compston JE, Mellish RWE, Garrahan NJ. Age-related changes in iliac crest trabecular micro-anatomic bone in man. Bone 1987;8: [13] Dempster DW. Bone histomorphometry in glucocorticoidinduced osteoporosis. J Bone Miner Res 1989;4: [14] Fazzalari NL, Parkinson IH. Fractal dimension and architecture of trabecular bone. J Pathol 1996;178: [15] Felsenberg D, Boonen S. The bone quality framework: determinants of bone strength and their interrelationships, and implications for osteoporosis management. Clin Ther 2005;27:1 11. [16] Frost HM. Bone mass and the mechanostat : a proposal. Anat Rec 1987;219:1 9. [17] Frost HM. Skeletal structural adaptations to mechanical usage (SATMU): 2. Redefining Wolff s law: the remodeling problem. Anat Rec 1990;226: [18] Frost HM. Some ABCs of skeletal pathophysiology. III: bone balance and the delta B. BMU. Calcif Tissue Int 1989;45: [19] Guggenbuhl P, Bodic F, Hamel L, Baslé MF, Chappard D. Texture analysis of X-ray radiographs of iliac bone is correlated with bone micro-ct. Osteoporos Int 2006;17: [20] Guillemin G, Meunier A, Dallant P, Christel P, Pouliquen JC, Sedel L. Comparison of coral resorption and bone apposition with two natural corals of different porosities. J Biomed Mater Res 1989;23: [21] Hahn M, Vogel M, Pompesius-Kempa M, Delling G. Trabecular bone pattern factor a new parameter for simple quantification of bone microarchitecture. Bone 1992;13: [22] Hernandez CJ, Beaupré GS, Carter DR. A theoretical analysis of the changes in basic multicellular unit activity at menopause. Bone 2003;32: [23] Kleerekoper M, Villanueva AR, Stanciu J, Sudhaker RD, Parfitt AM. The role of three-dimensional trabecular microstructure in the pathogenesis of vertebral compression fractures. Calcif Tissue Int 1985;37: [24] Le HM, Holmes RE, Shors E, Rosenstein DA. Computerized quantitative analysis of the interconnectivity of porous biomaterials. Acta Stereol 1992;11S1: [25] Legrand E, Audran M, Guggenbuhl P, Levasseur R, Chales G, Baslé MF, et al. Trabecular bone microarchitecture is related to the number of risk factors and etiology in osteoporotic men. Microsc Res Tech 2007;70: [26] Legrand E, Chappard D, Pascaretti C, Duquenne M, Krebs S, Simon Y, et al. Trabecular bone microarchitecture, bone mineral density and vertebral fractures in male osteoporosis. J Bone Miner Res 2000;15:13 9. [27] Lespessailles E, Chappard C, Bonnet N, Benhamou CL. Imaging techniques for evaluating bone microarchitecture. Joint Bone Spine 2006;73: [28] Levitz P, Tchoubar D. Disordered porous solids: from chord distributions to small angle scattering. J Phys 1992;2: [29] Libouban H, Moreau MF, Legrand E, Audran M, Baslé MF, Chappard D. Comparison of histomorphometric descriptors of bone architecture with dual-energy X-ray absorptiometry for assessing bone loss in the orchidectomized rat. Osteoporos Int 2002;13: [30] Meunier PJ, Courpron P, Edouard C, Bernard J, Bringuier JP. Physiological senile involution and pathological rarefaction of bone. Quantitative and comparative histological data. Br Med J 1973;2: [31] Mosekilde L. Sex differences in age-related loss of vertebral trabecular bone mass and structure biomechanical consequences. Bone 1989;10: [32] Parfitt AM, Drezner MK, Glorieux FG, Kanis JA, Malluche H, Meunier PJ, et al. Bone histomorphometry: standardization of nomenclature, symbols, and units. Report of the ASBMR histomorphometry nomenclature committee. J Bone Miner Res 1987;2: [33] Parfitt AM, Matthews CHE, Villanueva AR, Kleerekoper M, Frame B, Rao DS. Relationships between surface, volume and thickness of iliac trabecular bone in aging and in osteoporosis. Implications for the microanatomic and cellular mechanisms of bone loss. J Clin Invest 1983;72: [34] Peyrin F, Muller C, Carillon Y, Nuzzo S, Bonnassie A, Briguet A. Synchrotron radiation microct: a reference tool for the characterization of bone samples. Adv Exp Med Biol 2001;496: [35] Singh I. The architecture of cancellous bone. J Anat 1978;127: [36] Sornay-Rendu E, Boutroy S, Munoz F, Delmas PD. Alterations of cortical and trabecular architecture are associated with fractures in postmenopausal women, partially independent of decreased BMD measured by DXA: the OFELY study. J Bone Miner Res 2007;22: [37] Tabor Z. Analysis of the influence of image resolution on the discriminating power of trabecular bone architectural parameters. Bone 2004;34: [38] Vesterby A. Star volume in bone research: a histomorphometric analysis of trabecular bone structure using vertical sections. Anat Rec 1993;235:

9 170 D. Chappard et al. [39] Waarsing JH, Day JS, van der Linden JC, Ederveen AG, Spanjers C, De Clerck N, et al. Detecting and tracking local changes in the tibiae of individual rats: a novel method to analyse longitudinal in vivo micro-ct data. Bone 2004;34: [40] Whitehouse WJ. The quantitative morphology of anisotropic trabecular bone. J Microsc 1974;101: [41] Whitehouse WJ, Dyson ED, Jackson CK. The scanning electron microscope in studies of trabecular bone from a human vertebral body. J Anat 1971;108: