Musculoskeletal responses of 2-year-old Thoroughbred horses to early training. Conclusions

This paper summarises and presents in context the main findings of an extensive series of studies of early training lasting 13 weeks in which the tissue responses of 2-year-old Thoroughbred horses were assessed using a combination of methods. Negligible clinical injury was detected and thus the study fulfilled the intention of investigating adaptive change rather than injury. Cancellous and cortical bone, some digital tendons, and articular cartilage responded to early training exercise to a greater or lesser degree. Clinical examination and ancillary diagnostic aids currently in veterinary clinical use are not sufficient to detect early abnormalities in metacarpo-phalangeal joint (MCPJ) cartilage found in both trained and untrained horses. Future work should centre on detection of such changes, on the precise registration of training workload, and on the manipulation of the responses of musculoskeletal tissues by careful investigation of the effects of introducing conditioning exercise at a young age.     

Musculoskeletal responses of 2-year-old Thoroughbred horses to early training. 1. Study design, and clinical, nutritional, radiological and histological observations

AIMS: This is the first in a series of papers reporting studies in 2-year-old Thoroughbred racehorses that aimed to determine the response of musculoskeletal tissues to early training on grass and sand racetracks. In this paper, the experimental set-up of the whole study is described, and nutritional, workload, and clinical, radiographic and pathological outcomes are reported, including semi-quantitative assessment of macroscopic changes in articular cartilage. METHODS: The study group comprised 14 two-year-old Thoroughbred fillies reared entirely at pasture. Of these, seven were selected by a licensed racehorse trainer to undergo a 4-week period of initial training in which they were taught to accept saddle and rider, followed by a 13-week period of flatrace training at a racetrack (Weeks 1-13); the other seven fillies were confined to large grass enclosures and were not trained. Nutrient, including macro- and trace-element intakes were estimated. Distances cantered or galloped and average velocities were quantified for the trained horses. All horses were observed daily, weighed approximately weekly, and underwent a clinical lameness examination at Weeks 5, 9 and 13. Distal forelimbs were radiographed prior to Week 1, during Weeks 7-8, and again at the end of the study, when macroscopic changes in articular cartilage of the proximal surface of the proximal phalanx were also scored after staining with India ink. RESULTS: Dietary intakes met or exceeded recommended requirements for all nutrients except sodium, which was low in the trained horses. Bodyweight increased throughout the study in the untrained horses, and increased until Week 7 and then decreased slightly in the trained horses. Mean velocity data were used to define three stages of the training programme: Stage 1 comprised canter in Weeks 1-4; Stage 2 comprised canter in Weeks 5-8; and Stage 3 comprised canter in Weeks 9-13 and galloping twice weekly. Four of seven horses completed training. These covered a mean distance of 179.2 km at mean velocities (excluding gallops) of 7.63 m/sec (SD 0.58), 8.99 m/sec (SD 0.56), and 8.43 m/sec (SD 0.74) for Stages 1-3, respectively, and galloped 4.45 km at 14.4 (SD 0.1) m/sec. The three horses that did not complete training became lame in Weeks 9, 10 and 11, and covered 147.9 km at velocities of 7.38 m/sec (SD 0.44), 8.88 m/sec (SD 0.33) and 8.43 m/sec (SD 0.59) and galloped 2.1 km. Overall, slight or intermittent lameness in trained horses was noted on 76/655 (12%) of horse observation days. Swelling was evident on 284/655 (41%) of horse observation days in the metacarpophalangeal (MCPJ) and metatarsophalangeal (MTPJ) joints (92%), palmar metacarpal tendon region (7%) or carpus (1%). Swelling of the MCPJ or MTPJ was not associated with obvious lameness. Radiographic changes were minor and no gross lesions in bone or tendon tissue were evident except for one case of dorsal metacarpal disease. Post mortem, the cartilage of some MCPJ and MTPJ had obvious wear lines and high lesion scores, which were not consistently related to clinical evidence of pain, lameness or joint swelling. Mean lesion scores were not significantly different between the MCPJ and MTPJ, or between trained and untrained horses. CONCLUSIONS: Workload can be readily quantified in racehorses under semi-commercial training conditions. Obvious lesions in cartilage of the MCPJ or MTPJ were present in some trained and some untrained horses and not consistently associated with clinical evidence of lameness, joint swelling or change in other connective tissues. CLINICAL RELEVANCE: Workload data in racing horses are likely to be highly relevant for studying the pathogenesis of changes in bone, tendon and cartilage during training, for training management and for risk analysis in racehorse populations. Although obvious cartilage lesions produced little clinical effect, such lesions have previously been shown to be progressive and to prejudice athletic capability. Detection of such occult lesions in young horses will require more sophisticated detection methods.     

Musculoskeletal responses of 2-year-old Thoroughbred horses to early training. 7. Bone and articular cartilage response in the carpus

AIM: To describe features of the morphology of the carpus, quantify the thickness of hyaline and calcified cartilage, and to describe the morphology and density of subchondral bone in the third carpal bone (C3) of young Thoroughbred horses in early training.

METHODS: C3 of seven 2-year-old horses in training and seven untrained horses matched for age, sex and breed were assessed by gross appearance, computed tomography, fine-structure radiography, image analysis of high-resolution photographs, and histology.

RESULTS: Macroscopic lesions in cartilage were few and mild, and not significantly different between groups. High bone mineral density (BMD), in some cases typical of cortical bone, was confined to the dorsal load path, and was significantly higher in trained than in untrained horses (p<0.01). In the most dorsoproximal aspect of the radial articular facet, apparently outside the dorsal load path, the BMD in both trained and untrained horses was significantly less than in other regions of interest (ROIs). Adaptive increase in density was associated with thickening of the (junctions of) trabeculae oriented proximo-distally. Hyaline cartilage was thicker (p<0.001) in the concavity of the radial articular facet than dorsal or palmar to it, and was thicker in the trained than untrained group (p=0.007). No such differences were detected in the thickness of articular calcified cartilage (ACC).

CONCLUSIONS: The rapid response of bone in C3 to relatively small amounts of high-speed exercise was confirmed. A previously unreported increase in thickness of hyaline cartilage was evident, perhaps indicating that this tissue may be more responsive than hitherto thought, at least to particular types of exercise at particular times. These changes occurred with little evidence of abnormality, and thus appeared to be adaptive to the exercise regimen. The model developed should be used for further definition of the exercise stimulus required to produce adaptive, protective changes in sites susceptible to athletic injury.

CLINICAL RELEVANCE: The data will serve as reference for use in subsequent imaging studies in which sophisticated aids such as magnetic resonance imaging (MRI) may be used to predict carpal lesions.     

Musculoskeletal responses of 2-year-old Thoroughbred horses to early training. 4. Morphometric,microscopic and biomechanical properties of the digital tendons of the forelimb

AIM: To determine the weight, volume, density and cross-sectional area (CSA) of the digital flexor tendons, common digital extensor tendon (CDET) and suspensory ligament (SL) of the forelimb of young Thoroughbred horses in early training, and to assess the response to a training programme of known parameters of superficial digital flexor tendon (SDFT) tissue at mid-metacarpal level.

METHODS: The tendons of seven 2-year-old Thoroughbred horses in training were inspected, transected into segments of known length, and compared with those from seven untrained horses matched for age, sex and breed. The weight, volume, density and CSA of each segment, and the crimp angle, histological features, and biomechanical ultimate stress and stiffness of tendon samples from the mid-metacarpal region of the SDFT were determined.

RESULTS: There was no macroscopic evidence of swelling or discolouration in any of the tendon segments or cut surfaces. The volumes of SDFT and CDET segments of horses in the trained group were significantly greater than those in the untrained group (p=0.036 and p=0.039, respectively). A greater increase in volume than weight resulted in a lower density in the SDFT but not CDET in trained compared with untrained horses (p=0.038). CSA of these two tendons was significantly greater in the trained group (p=0.002 and 0.036, respectively), the percentage increase being greater in the CDET than the SDFT. The number of tenocytes at four sites in the mid-metacarpal SDFT region was less in trained than untrained horses (p=0.025). There was no histological evidence of inflammation, and no difference in crimp angle between groups. There was no significant between-group difference in stiffness or ultimate stress of tendon strips.

CONCLUSIONS: Volume and CSA of the SDFT and CDET were larger in trained than untrained horses. The SDFT was less dense in the trained group. Because no evidence of tendonitis was detected and training appeared to have no significant effect on crimp angle or biomechanical properties of tendon strips, the size and density changes were presumed to be adaptive and induced by the training.

CLINICAL RELEVANCE: Although evident in this in vitro study, the detection of adaptive from initial pathological increase in size of the SDFT is likely to be difficult in vivo.     

Musculoskeletal responses of 2-year-old Thoroughbred horses to early training. 3. In vivo ultrasonographic assessment of the cross-sectional area and echogenicity of the superficial digital flexor tendon

AIM: To determine if the superficial digital flexor tendon (SDFT) of young Thoroughbred horses changed in size and echogenicity in association with early race training.

METHODS: Cross-sectional area (CSA) and echogenicity were determined ultrasonographically at five levels of the SDFT of the forelimbs of 2-year-old fillies (n=14), corresponding to 4, 8, 12, 16 and 20 cm distal to the accessory carpal bone (DACB). Measurements were made before and after a 13-week period in which a trained group of seven horses was compared with another group of seven untrained horses.

RESULTS: Level below the accessory carpal bone had a significant effect on CSA and Level 8 was smaller than all other levels except Level 12, while Level 12 was smaller than Levels 4 and 20 but not different from Levels 8 and 16. There was a significant interaction between level and time due to effects observed at Level 8. The CSA at Level 8 measured pre-training was different from that of Levels 4 and 20 in both pre- and post-training groups (p<0.05), but when measured post-training was not different from any other measurement. There was also a significant interaction between treatment group and time. There was no difference between CSA for the untrained and trained groups at the pre-training observation (p=0.9), but post-training the CSA (pooled over all levels) in trained horses was significantly larger than that of the untrained horses both post-training (p=0.019) and pre-training (p=0.034), and was not different from the pretraining CSA recorded in the trained group (p=0.29). Treatment group had no effect on echogenicity (p=0.43), while echogenicity was less at the end of the trial in both trained and untrained horses (p<0.001).

CONCLUSIONS: Early training for racing was associated with an increase in mean CSA of the SDFT. Other factors such as age and maturity may play a role in limiting this increase.     

Musculoskeletal responses of 2-year-old Thoroughbred horses to early training. 6. Bone parameters in the third metacarpal and third metatarsal bones

AIM: To determine the effect of a known training regimen on the size and mineral content of the third metacarpal (Mc3) and third metatarsal (Mt3) bones of 2-year-old Thoroughbred horses trained on racetracks.

METHODS: Mc3 and Mt3 of seven horses trained on grass and sand tracks were scanned at several sites using conventional quantitative and peripheral computed tomography (CT). Bone dimensions and density in the diaphysis and epiphysis were compared with those from seven untrained horses. Calcein label was injected in two clusters, during Weeks 9 and 12. The extent and rate of diaphyseal modelling was determined by confocal fl uorescent microscopic examination of thin plane parallel sections of the mid-metacarpal region.

RESULTS: Volumetric bone mineral density (BMDv) of the epiphysis was markedly higher and of the diaphysis was slightly higher in trained compared with untrained horses, but greater bone size in the trained horses had the greatest effect on an index of bone strength. Active osteons, defined as Haversian systems containing calcein label, were fewer, of smaller diameter at the time of calcein injection, and had a greater bone apposition rate in trained than in untrained horses.

CONCLUSIONS: Conventional training of 2-year-old Thoroughbred racehorses over a 13-week period had a significant effect on bone size, density and strength index when compared to untrained horses.

CLINICAL RELEVANCE: Bone responded rapidly to early training. The data provide reference values and sites for use in longitudinal studies of commercial training regimens.     

Musculoskeletal responses of 2-year-old Thoroughbred horses to early training. Conclusions

This paper summarises and presents in context the main findings of an extensive series of studies of early training lasting 13 weeks in which the tissue responses of 2-year-old Thoroughbred horses were assessed using a combination of methods. Negligible clinical injury was detected and thus the study fulfilled the intention of investigating adaptive change rather than injury. Cancellous and cortical bone, some digital tendons, and articular cartilage responded to early training exercise to a greater or lesser degree. Clinical examination and ancillary diagnostic aids currently in veterinary clinical use are not sufficient to detect early abnormalities in metacarpo-phalangeal joint (MCPJ) cartilage found in both trained and untrained horses. Future work should centre on detection of such changes, on the precise registration of training workload, and on the manipulation of the responses of musculoskeletal tissues by careful investigation of the effects of introducing conditioning exercise at a young age.     

Musculoskeletal responses of 2-year-old Thoroughbred horses to early training. 5. Kinematic effects

AIM: To examine the effect of race training on kinematic parameters of the trot in a sample of 2-year-old Thoroughbred fillies. METHODS: Two-year-old Thoroughbred fillies that had been either trained for racing (n=7) or left untrained in small paddocks (n=7) were examined for lameness, and linear and temporal kinematic data were collected from video recordings of each horse trotted in-hand, 5, 9 and 13 weeks after the beginning of training. Horses were graded as: sound (Grade 0), suspected but not obviously lame (Grade 1), or definitely lame at the trot (Grade 2). RESULTS: Mean velocity at the trot did not vary with lameness grade. Compared with horses classified as sound (Grade 0), horses classified as suspected or definitely lame (Grades 1 and 2) had shorter (2.55 (SE 0.02) vs 2.41 (SE 0.02) and 2.44 (SE 0.03) m; p<0.001) and faster (671 (SE 3) vs 645 (SE 1) and 647 (SE 4) msec; p<0.001) strides. None of the linear or temporal parameters measured varied significantly between the successive observation periods. When data from sound horses were pooled, the training group trotted at a higher mean velocity (4.22 (SE 0.08) vs 3.22 (SE 0.06) m/sec; p<0.001) and with a longer stride length (2.79 (SE 0.05) vs 2.18 (SE 0.06) m; p<0.001) than the untrained group. The stride duration was longer (669 (SE 7) vs 662 (SE 9) msec; p=0.01), stance period was shorter (34.1 (SE 0.6) vs 39.3 (SE 0.7) %; p<0.001) and mid-stance was achieved earlier in the stride (12.1 (SE 0.6) vs 13.1 (SE 0.2) %; p=0.001) in the trained than the untrained group. A longer swing phase in the trained group was associated with an increase in retraction time (9.2 (SE 0.2) vs 7.6 (SE 0.3) %; p=0.001). CONCLUSIONS: These data indicate that early race training in young Thoroughbreds was associated with quantifiable changes in linear and temporal kinematic parameters of the trot, which were related to the race-training objective of improving the horse's ability to work at higher velocities. CLINICAL RELEVANCE: Race training resulted in changes in the gait of non-lame horses at the trot, which should be considered during veterinary lameness examinations.     

Musculoskeletal responses of 2-year-old Thoroughbred horses to early training. 8. Quantitative back-scattered electron scanning electron microscopy and confocal fluorescence microscopy of the epiphysis of the third metacarpal bone

AIM: To characterise and explain the increase in density evident by computerised tomography (CT) and radiography in companion studies as a response to training, in bone in the palmar and dorsal regions of the condyles of the third metacarpal bone (Mc3) of 2-year-old Thoroughbred horses.

METHODS: Compositional back-scattered electron (BSE) imaging in scanning electron microscopy (SEM) and confocal scanning laser microscopy (CSLM) were conducted on polymethyl methacrylate (PMMA)-embedded mediolateral slices of the right distal Mc3 from seven 2-year-old Thoroughbred horses trained on a racetrack and seven untrained horses kept at pasture. One left Mc3 from each group was studied in transverse section planes. This study focussed on regions of Mc3 found to differ in density between the trained and untrained horses in companion studies using CT and radiography.

RESULTS: The increase of bone density in the condyles of Mc3 in trained horses compared with untrained horses occurred, without prior osteoclastic resorption, via the deposition of new bone on pre-existing internal surfaces. Within prior marrow spaces of cancellous bone, there was also rapid formation of immature strands and fronds of bone which were more cellular and mineralised, and more lamellar bone tissue was deposited on these new scaffolding elements in the trained horses. Both resulted in increased bone volume fraction (BVF). The microscopic mineralisation density of the bulk of the new tissue was lower than in pre-existing bone, and CT and radiography underestimated the increase in BVF. The new tissue was thus probably less stiff at the microscopic scale than pre-existing bone, though its addition would stiffen the global structure.

CONCLUSIONS: In Mc3 of all the trained horses, there were obvious differences in microscopic structure compared with those from the untrained horses. Moderate, industry-standard levels of exercise used to prepare young horses for racing induced the formation of new bone in non-bone spaces in bone tissue, such that the bone organ should better withstand later increased levels of exercise.     

Musculoskeletal responses of 2-year-old Thoroughbred horses to early training. 7. Bone and articular cartilage response in the carpus

AIM: To describe features of the morphology of the carpus, quantify the thickness of hyaline and calcified cartilage, and to describe the morphology and density of subchondral bone in the third carpal bone (C3) of young Thoroughbred horses in early training. METHODS: C3 of seven 2-year-old horses in training and seven untrained horses matched for age, sex and breed were assessed by gross appearance, computed tomography, fine-structure radiography, image analysis of high-resolution photographs, and histology. RESULTS: Macroscopic lesions in cartilage were few and mild, and not significantly different between groups. High bone mineral density (BMD), in some cases typical of cortical bone, was confined to the dorsal load path, and was significantly higher in trained than in untrained horses (p<0.01). In the most dorsoproximal aspect of the radial articular facet, apparently outside the dorsal load path, the BMD in both trained and untrained horses was significantly less than in other regions of interest (ROIs). Adaptive increase in density was associated with thickening of the (junctions of) trabeculae oriented proximo-distally. Hyaline cartilage was thicker (p<0.001) in the concavity of the radial articular facet than dorsal or palmar to it, and was thicker in the trained than untrained group (p=0.007). No such differences were detected in the thickness of articular calcified cartilage (ACC). CONCLUSIONS: The rapid response of bone in C3 to relatively small amounts of high-speed exercise was confirmed. A previously unreported increase in thickness of hyaline cartilage was evident, perhaps indicating that this tissue may be more responsive than hitherto thought, at least to particular types of exercise at particular times. These changes occurred with little evidence of abnormality, and thus appeared to be adaptive to the exercise regimen. The model developed should be used for further definition of the exercise stimulus required to produce adaptive, protective changes in sites susceptible to athletic injury. CLINICAL RELEVANCE: The data will serve as reference for use in subsequent imaging studies in which sophisticated aids such as magnetic resonance imaging (MRI) may be used to predict carpal lesions.     

Musculoskeletal responses of 2-year-old Thoroughbred horses to early training. 6. Bone parameters in the third metacarpal and third metatarsal bones

AIM: To determine the effect of a known training regimen on the size and mineral content of the third metacarpal (Mc3) and third metatarsal (Mt3) bones of 2-year-old Thoroughbred horses trained on racetracks. METHODS: Mc3 and Mt3 of seven horses trained on grass and sand tracks were scanned at several sites using conventional quantitative and peripheral computed tomography (CT). Bone dimensions and density in the diaphysis and epiphysis were compared with those from seven untrained horses. Calcein label was injected in two clusters, during Weeks 9 and 12. The extent and rate of diaphyseal modelling was determined by confocal fluorescent microscopic examination of thin plane parallel sections of the mid-metacarpal region. RESULTS: Volumetric bone mineral density (BMDv) of the epiphysis was markedly higher and of the diaphysis was slightly higher in trained compared with untrained horses, but greater bone size in the trained horses had the greatest effect on an index of bone strength. Active osteons, defined as Haversian systems containing calcein label, were fewer, of smaller diameter at the time of calcein injection, and had a greater bone apposition rate in trained than in untrained horses. CONCLUSIONS: Conventional training of 2-year-old Thoroughbred racehorses over a 13-week period had a significant effect on bone size, density and strength index when compared to untrained horses. CLINICAL RELEVANCE: Bone responded rapidly to early training. The data provide reference values and sites for use in longitudinal studies of commercial training regimens.     

Musculoskeletal responses of 2-year-old Thoroughbred horses to early training. 4. Morphometric, microscopic and biomechanical properties of the digital tendons of the forelimb

AIM: To determine the weight, volume, density and cross-sectional area (CSA) of the digital flexor tendons, common digital extensor tendon (CDET) and suspensory ligament (SL) of the forelimb of young Thoroughbred horses in early training, and to assess the response to a training programme of known parameters of superficial digital flexor tendon (SDFT) tissue at mid-metacarpal level. METHODS: The tendons of seven 2-year-old Thoroughbred horses in training were inspected, transected into segments of known length, and compared with those from seven untrained horses matched for age, sex and breed. The weight, volume, density and CSA of each segment, and the crimp angle, histological features, and biomechanical ultimate stress and stiffness of tendon samples from the mid-metacarpal region of the SDFT were determined. RESULTS: There was no macroscopic evidence of swelling or discolouration in any of the tendon segments or cut surfaces. The volumes of SDFT and CDET segments of horses in the trained group were significantly greater than those in the untrained group (p=0.036 and p=0.039, respectively). A greater increase in volume than weight resulted in a lower density in the SDFT but not CDET in trained compared with untrained horses (p=0.038). CSA of these two tendons was significantly greater in the trained group (p=0.002 and 0.036, respectively), the percentage increase being greater in the CDET than the SDFT. The number of tenocytes at four sites in the mid-metacarpal SDFT region was less in trained than untrained horses (p=0.025). There was no histological evidence of inflammation, and no difference in crimp angle between groups. There was no significant between-group difference in stiffness or ultimate stress of tendon strips. CONCLUSIONS: Volume and CSA of the SDFT and CDET were larger in trained than untrained horses. The SDFT was less dense in the trained group. Because no evidence of tendonitis was detected and training appeared to have no significant effect on crimp angle or biomechanical properties of tendon strips, the size and density changes were presumed to be adaptive and induced by the training. CLINICAL RELEVANCE: Although evident in this in vitro study, the detection of adaptive from initial pathological increase in size of the SDFT is likely to be difficult in vivo.     

Musculoskeletal responses of 2-year-old Thoroughbred horses to early training. 3. In vivo ultrasonographic assessment of the cross-sectional area and echogenicity of the superficial digital flexor tendon

AIM: To determine if the superficial digital flexor tendon (SDFT) of young Thoroughbred horses changed in size and echogenicity in association with early race training. METHODS: Cross-sectional area (CSA) and echogenicity were determined ultrasonographically at five levels of the SDFT of the forelimbs of 2-year-old fillies (n=14), corresponding to 4, 8, 12, 16 and 20 cm distal to the accessory carpal bone (DACB). Measurements were made before and after a 13-week period in which a trained group of seven horses was compared with another group of seven untrained horses. RESULTS: Level below the accessory carpal bone had a significant effect on CSA and Level 8 was smaller than all other levels except Level 12, while Level 12 was smaller than Levels 4 and 20 but not different from Levels 8 and 16. There was a significant interaction between level and time due to effects observed at Level 8. The CSA at Level 8 measured pre-training was different from that of Levels 4 and 20 in both pre- and post-training groups (p<0.05), but when measured post-training was not different from any other measurement. There was also a significant interaction between treatment group and time. There was no difference between CSA for the untrained and trained groups at the pre-training observation (p=0.9), but post-training the CSA (pooled over all levels) in trained horses was significantly larger than that of the untrained horses both post-training (p=0.019) and pre-training (p=0.034), and was not different from the pre-training CSA recorded in the trained group (p=0.29). Treatment group had no effect on echogenicity (p=0.43), while echogenicity was less at the end of the trial in both trained and untrained horses (p<0.001). CONCLUSIONS: Early training for racing was associated with an increase in mean CSA of the SDFT. Other factors such as age and maturity may play a role in limiting this increase.     

Musculoskeletal responses of 2-year-old Thoroughbred horses to early training. 8. Quantitative back-scattered electron scanning electron microscopy and confocal fluorescence microscopy of the epiphysis of the third metacarpal bone

AIM: To characterise and explain the increase in density evident by computerised tomography (CT) and radiography in companion studies as a response to training, in bone in the palmar and dorsal regions of the condyles of the third metacarpal bone (Mc3) of 2-year-old Thoroughbred horses. METHODS: Compositional back-scattered electron (BSE) imaging in scanning electron microscopy (SEM) and confocal scanning laser microscopy (CSLM) were conducted on polymethyl methacrylate (PMMA)-embedded mediolateral slices of the right distal Mc3 from seven 2-year-old Thoroughbred horses trained on a racetrack and seven untrained horses kept at pasture. One left Mc3 from each group was studied in transverse section planes. This study focussed on regions of Mc3 found to differ in density between the trained and untrained horses in companion studies using CT and radiography. RESULTS: The increase of bone density in the condyles of Mc3 in trained horses compared with untrained horses occurred, without prior osteoclastic resorption, via the deposition of new bone on pre-existing internal surfaces. Within prior marrow spaces of cancellous bone, there was also rapid formation of immature strands and fronds of bone which were more cellular and mineralised, and more lamellar bone tissue was deposited on these new scaffolding elements in the trained horses. Both resulted in increased bone volume fraction (BVF). The microscopic mineralisation density of the bulk of the new tissue was lower than in pre-existing bone, and CT and radiography underestimated the increase in BVF. The new tissue was thus probably less stiff at the microscopic scale than pre-existing bone, though its addition would stiffen the global structure. CONCLUSIONS: In Mc3 of all the trained horses, there were obvious differences in microscopic structure compared with those from the untrained horses. Moderate, industry-standard levels of exercise used to prepare young horses for racing induced the formation of new bone in non-bone spaces in bone tissue, such that the bone organ should better withstand later increased levels of exercise.