The ultimate strength of bone is determined by the bone mineral density (BMD) of the cortical and cancellous bone as well as the trabecular and cortical microarchitecture. In human medicine, determination of these variables is critical for the early detection of osteoporosis and other bone diseases. For years only BMD was determined, but recently three-dimensional microarchitecture was added to the list of criteria [1–3]. This is important because the BMD of healthy and osteoporotic bones can overlap [4, 5]. Bone mineral density can be measured in vivo using dual energy x-ray absorptiometry (DEXA), quantitative ultrasonography (QUS) and peripheral quantitative computed tomography (pQCT) [1, 6–9]. The three-dimensional bone structure can only be assessed using histomorphometric methods [6, 9–12]. However, because trabecular bone consists of a three-dimensional network, even stereological techniques are not sufficient to produce an exact three-dimensional definition of the bone micro-structure based on histomorphological findings. The introduction of micro-computed tomography (micro-CT) allowed, for the first time, the stereological in-vivo examination of bone [8]. The decreasing number and thickness of bone trabeculae, but also other variables of bone architecture, can be detected earlier with micro-CT than via histomorphometry. In other studies age-related variations in the microstructure, the structure model type and trabecular thickness of human cancellous bone were investigated [10, 13]; these studies showed that the Structure Model Index (SMI) increased and the trabecular thickness decreased with age. The SMI is defined as a value between 0 and 3 and is calculated from the relative amount of the number of trabecelae, which have the shape of rods or plates. Xtreme computed tomography (XtremeCT, Scanco Medical, Auenring 6–8, 8303 Bassersdorf, Switzerland), which is a high-resolution pQCT, provides micro-CT for the clinical diagnosis of osteoporosis and other bone diseases in human medicine. The resolution is limited to about 100 μm because of the low x-ray dose allowed for patients. The Xtreme-CT-generated structural data were verified by comparing them to values obtained using 28-micron-resolution micro-CT [14]. The correlation coefficients for the values obtained by the two methods ranged from 0.81 to 0.98. Laib and co-workers (1998) evaluated healthy post-menopausal women for microstructural changes using a prototype of Xtreme-CT (Fig. 1) and found that the extent as well as the localization of bone loss varies greatly among individuals [15].

Although equine bone diseases cannot be compared directly with osteoporosis in humans, they affect the use and longevity of the horse and therefore require adequate diagnosis and treatment. The most commonly-used diagnostic methods in horses include macroscopic and radiographic evaluation [16, 17], DEXA and QUS. Dual energy x-ray absorptiometry has been used by Firth and co-workers (1999) to assess the effect of age, exercise and growth rate on osteochondrosis in foals [18]. McClure and co-workers (2001) compared BMD obtained via DEXA with the density that was measured using Archimedes' principle [19]. Donabedian and co-workers (2005) were the first to make DEXA measurements in live horses [20]. Quantitative ultrasonography is another commonly used diagnostic technique, and has the advantage of providing information other than BMD. This is because the attenuation of the sound waves (broadband ultrasound attenuation [BUA]) and the speed of sound through the bone (speed of sound [SOS]), two variables which are measured by QUS, are greatly affected by the mechanical properties and structure of the medium [21, 22]. Quantitative ultrasonography has been used by Buckingham and Jeffcott (1991) to investigate the osteopenic effects of limb immobilization in horses [23, 24]. Carstanjen and co-workers (2002) concluded that QUS was useful for the assessment of the metacarpus, radius and tibia in horses [25]. Another study investigated DEXA and QUS in horses and found that the two methods provided different information; QUS measured not only BMD but also the microstructure and composition of the bone [26].

Peripheral quantitative computed tomography is another non-invasive and very exact method of bone evaluation, which allows, to a certain extent, separate assessment of cortical and cancellous bone. The results of BMD determined by pQCT are in good agreement with other methods of measurement [27]. Cornelissen and co-workers (1999) used pQCT (XCT 960A, Stratec, Germany) to determine the influence of exercise on BMD of immature bone tissue in horses. Using a manually determined cut-off value, those authors assigned BMD values to trabecular bone. However, this did not constitute a direct measurement as is obtained via micro-CT. Computed tomography was also used by Waite and others [28] and Mäule and Gerhards (2004) to examine the distal limb of horses [29]. They concluded that the tissue density cannot be used as the sole criterion for the assessment of bone, because individual variations were too large. In addition to all the imaging techniques, biochemical markers can also be used to make an overall assessment of skeletal condition [21, 30]. However, the methods described thus far allow only a general survey of the entire skeleton rather than the evaluation of individual bones. Furthermore, it has not been feasible to adequately and directly assess the microstructure of trabecular bone. The technique required to achieve this is micro-CT, which has rarely been used in horses [31, 32].

In the present study, micro-CT (Xtreme CT) was used to obtain detailed data on cortical and trabecular bone, including microstructure, in the radii and tibiae of 15 horses. Special emphasis was given to the effects of age on the BMD and microarchitecture of the bones. These variables are likely to affect the mechanical properties of bones and may influence the fracture tendency in human bones [33]. Our investigations, together with more biomechanical testing, could provide better information as to whether changes in the microarchitecture influence the susceptibility to fractures. Likewise, the micro-CT data could provide the basis for using Finite Element Methods, which serve to determine the effects of various loads on bone and to simulate possible fracture configurations via a computer. The objective of this study was to investigate age-related changes in the BMD of the radius and tibia of the horse.

Both ends of the bones were embedded in epoxy resin (Biresin^® G28 Harz and Biresin^® G26 Härter, 1:1 mixture, Sika Germany GmbH, Bad Urach, Germany) to a level above the widest part of the epiphysis, and the bones were labeled. The bones were re-wrapped in saline-soaked cloth, placed individually in plastic bags to prevent loss of moisture and stored at 4°C until further use.

A total of 56 bones were examined using XtremeCT (Fig. 1), which is a high resolution (41–246 μm nominal isotropic resolution) peripheral computed tomography unit for in-vivo measurement of BMD and microstructure in humans. In human medicine, this recently introduced imaging modality is used mainly for the early detection and monitoring of osteoporosis and other bone diseases in the distal radius and distal tibia. XtremeCT has a microfocused x-ray beam with a maximum scanning length of 150 mm. The scan-time for 9 mm (110 slices) is 3 minutes. The bones were scanned at the 50% mark (73 slices, resolution 123 μm) and at the 80% mark (110 slices, resolution 82 μm; Fig. 2). A computer software program (HP AlphaStation) was used to extract the BMD and microstructure data of the bones from the images.

XtremeCT measures a large number of geometric, densitometric and physical variables, which are displayed in Excel data sheets. All BMDs were expressed as mg hydroxyapatite/cm³ (HA/cm³). The following variables were determined:

In the present study, XtremeCT, a type of peripheral computed tomography, was used to evaluate the microstructure of trabecular and cortical bone of the horse. This technique was originally designed to evaluate the human tibia and radius [15]. In contrast to human medicine, this technique is not suited for in-vivo studies in the horse because of certain restrictions. Extremities to be examined must remain in the Xtreme computed tomographic scanner without moving for six minutes. It would therefore be feasible theoretically to examine the third phalangeal bone in the anaesthetized horse. Soft tissue was removed from the bones in this study to allow solid fixation of the bones in the measuring apparatus (Fig. 1) and to identify the two predetermined measuring sites (Fig. 2). To date, there have only been a limited number of studies on microtomography in which a prototype of Xtreme-CT was used, and these studies dealt exclusively with human bone [11, 15].

Our results showed that the equine tibia had a larger mean cross-sectional surface area and a thicker diaphyseal and metaphyseal cortex than the radius. However, there were no significant differences between the radius and the tibia with respect to the trabecular structure. A possible reason for this is that in the horse, but not in man, both the radius and tibia are weight-bearing bones. In humans, the microstructure of cancellous bone varies, particularly among long bones, vertebrae and the flat pelvic bones [36].

There were very few apparent differences between the bones of mares and geldings. However, studies using a larger number of horses, including stallions and mares of all age groups are needed to confirm our results. The effects of exercise and different riding disciplines on BMD and bone architecture should also be investigated. Training has a positive effect on bone micro-architecture [24, 37, 38]. In humans, the differences between the bone structure of young men and women are also minimal. Only the peak bone mass, which is the maximum bone mass, is markedly higher in men than in women [39]. This is because men are generally taller than women and thus have larger bones. There are distinct differences in the loss of bone mass of men and women as they age. Although an age-related loss of bone mass is seen in both genders, the loss in women is more rapid and marked in the first few years after menopause [40]. This predisposes older women to a higher incidence of fractures. The rapid loss in bone mass is attributable to a sudden decrease in estrogen concentration during menopause. Unlike women, aged mares do not undergo a precipitous decrease in estrogen concentration because they maintain follicular activity [41], and therefore, a significantly different bone microstructure from that seen in older geldings or stallions would not be expected. More studies are needed to investigate whether geldings have poorer bone quality than stallions because of lower levels of male sex hormones.

With respect to the effect of age on bone quality, the trabecular number (Fig. 3 and 4) and trabecular BMD (Fig. 5) decreased, and the trabecular separation and cortical BMD of the metaphysis increased in the radius and tibia with increasing age. Based on these findings, we assume that the bone trabeculae are not replaced but rather progressively decrease in number with age. The trabecular thickness and volume also tended to decrease, although the changes were not significant. Evaluation of a larger number of bones may yield more age-related variables. Evaluation of the polar moment of inertia of the diaphysis and metaphysis of the radius and tibia revealed no age-related changes. Although in one study, horses with fractures were older than the overall equine patient population [42], there are no more indications that older horses have a higher incidence of fractures than younger horses. There is no scientific evidence of age-related osteoporosis in horses. Osteoporosis is defined as a severe loss in BMD predisposing the individual to spontaneous fractures [43]. Osteopenia, on the other hand, describes a decrease in BMD that is not associated with spontaneous fractures [44]. According to these definitions, the changes in trabecular BMD, number and separation that we observed in the older horses do not constitute osteoporosis (Fig. 6 and 7). Because of this, osteopenia commonly goes unnoticed and is probably more common than osteoporosis. However, other studies involving mostly cannon bones suggest that a decrease in the number and thickness of the trabeculae increases the risk of fracture [38–40]. Several studies have evaluated the mechanical properties of bone using micro-computed tomography and related these properties to bone strength [33, 45–47]. Nevertheless, because they have been used with success in elderly people, locking compression plates, which increase the stability of fracture repair, might be advantageous in horses as well [48].

In humans, the effect of age varies with the type of bone; using histomorphometry, scanning electron microscopy and biomechanical testing, Mosekilde (2000) determined that the weight-bearing ability of the human vertebrae was approximately 1,000 kg in young adults, but only 150 to 250 kg in the elderly [40]. This change is attributable to a decrease in the trabecular BMD, trabecular bone volume ratio, ash density and cortical thickness. However, the trabecular thickness decreases only in certain bones, for example in the femur [10]. These changes are first seen in people over 80 years of age. The structure model index also changes from a plate-like to a rod-like structure [10, 13], and the bone volume/total volume ratio as well as the bone strength decrease; however, these changes occur slightly earlier than thinning of the trabeculae and may be seen at the age of 60 [13].

Our study investigated BMD and microarchitecture of the radius and tibia because we have been interested in the frequency and configurations of fractures of the tibia and radius for some time [49]. Other studies have dealt with metacarpal and metatarsal bones of young racehorses [27] using DEXA, QUS or QCT. In another study, age-related changes of the mechanical properties of the metacarpal bones in horses were investigated based on changes in BMD, mechanical strength and fragility [50]. The BMD was not affected by the gender and the age of the horses, although it increased numerically until the age of six and then remained unchanged or decreased again. The strength of the bone peaked at about 4.5 years of age and correlated well with the BMD. The fragility of the bone peaked at about six years of age and thereafter was the only variable that decreased significantly with age.

To our knowledge, there are no high-resolution imaging studies on the micro-architecture of equine long bones and how it is affected by the gender and age of the horse. The results of the present study show that bone microstructure undergoes age-related changes, which may predispose to fractures. Further investigations are necessary to determine the effect of microarchitectural changes on the strength of bones and their susceptibility to fracture. Such studies should include very old mares to investigate the effect of ovarian senescence on bone micro-architecture.