Microdamage in the equine superficial digital flexor tendon.

The forelimb superficial digital flexor tendon (SDFT) is an energy-storing tendon that is highly susceptible to injury during activities such as galloping and jumping, such that it is one of the most commonly reported causes of lameness in the performance horse. This review outlines the biomechanical and biothermal effects of strain on the SDFT and how these contribute to accumulation of microdamage. The effect of age-related alterations on strain response and subsequent injury risk are also considered. Given that tendon is a slowly healing and poorly regenerative tissue, prompt detection of early stages of pathology in vivo and timely adaptations to training protocols are likely to have a greater outcome than advances in treatment. Early screening tools and detection protocols could subsequently be of benefit in identifying subclinical signs of degeneration during the training programme. This provides an opportunity for preventative strategies to be implemented to minimise incidences of SDFT injury and reduce recovery periods in elite performance horses. Therefore, this review will focus on the modalities available to implement early screening and prevention protocols as opposed to methods to diagnose and treat injuries.


| INTRODUC TI ON -THE TR AINING ENVIRONMENT-WHY IS THE S DF T MOS T AT RIS K?
The majority of superficial digital flexor tendon (SDFT) injuries occur during the fittening process of training regimes. 1 Previous studies have established that there is a lack of evidence-based practices in current equine conditioning programmes, 2,3 which may increase incidences of injury during training and limit the validity of previous research findings in the training environment.
Variations in training frequency, training intensity, training effects, surfaces and equipment have all been identified as potential risk factors for SDFT injury. 2,4,5 In addition to this, individual risk factors such as age, [6][7][8][9] breed 10 and sex, 11  The predominant function of all appendicular tendons is to transfer muscle-generated force to the skeleton. However, tendons can be classed as either positional or energy storing due to the differing in vivo load environments, with energy-storing tendons experiencing particularly high stresses or strains during exercise. 13,14 These tendons are stretched and recoil with each stride, storing and releasing energy which decreases the energetic cost of locomotion by as much as 26% in the human and 36% in the horse. 15,16 The principal energy-storing tendons are the forelimb SDFT in the horse, and the Achilles tendon in the human. 16 Due to such extreme loading environments in these energy-storing tendons, they are highly prone to overload-induced injury, with 75%-95% of tendon injuries in horses occurring in the forelimb SDFT, 17 and the majority of human tendon injuries occurring to the Achilles tendon. 18 There is also an age-associated increase in injury risk to the energy-storing tendons of both species, [19][20][21] which is discussed in a subsequent section.
In addition to the extreme loading environments experienced by energy-storing tendons, repeated loading of the tendon without sufficient recovery time between exercise sessions makes it more susceptible to failure, even when submaximal loads are applied. 22 Significant changes in the SDFT have been observed after racing, 23 with maximal changes observed 48 hours post-exercise, and returning to baseline values at 72 hours. This would suggest that exposing the tendon to high strains within a 72-hour period (ie before it can re-achieve a state of homeostasis) may further predispose the tendon to injury. Elevation of tendon temperature during high-intensity exercise is also a well-established result of mechanical overloading, and its effect on tenocyte function may have a significant impact on the onset of tendinopathy. 24 Several extraneous factors predispose the SDFT to injury during exercise. Guidance has been published on the effect of artificial surfaces on the musculoskeletal system, 4 possibly due to the rise in reported musculoskeletal injuries since the increased use of artificial surfaces. Additionally, increasing both fence height 25 and the frequency of jump training sessions within a training programme 26 have shown to increase chances of flexor tendon injuries.
Particularly with regard to the disciplines of National Hunt racing and the cross-country phase of eventing, abnormal mechanical loading resulting from varying terrain is likely to lead to erratic changes in magnitude, frequency, duration or direction of forces placed upon the limbs. 22 Due to the high incidence and multifaceted nature of SDFT injuries, this review summarises the current understanding of the structure-function relationships in the SDFT and identifies the biomechanical and biothermal effects of repeated high strains on the SDFT. Age-related alterations which result in increased injury risk are also reviewed, and we identify important areas for future study regarding the use of early screening tools and preventative strategies for minimising incidences and severity of SDFT injury.

| S TRUC TURE-FUN C TION REL ATI ON S HIPS AND THE S PECIALIS ED ROLE OF THE S DF T
The tendon is an integral part of joint movement and stability, as this intermediate tissue structure transmits the force generated from muscle contraction to bone, 27 facilitating movement around the joint. Tendons contain approximately 50%-60% water, and the remaining dry matter is composed of type I collagen (approximately 70%-85% DM), arranged in a hierarchical structure. 28 This structure is immersed in a noncollagenous matrix containing a small, but crucial, population of tenocytes which maintain the tendon's structural stability. As shown in Figure 1, collagen molecules accumulate to form subunits of increasing diameter, forming longitudinally aligned fibrils-the fundamental load-bearing material of structural tissues within the body. 28 Each fibril is composed of an arrangement of triple-helical collagen molecules, which are covalently cross-linked to one another, conferring strength to the fibril and overall stability to the molecules. These fibrils band together to form fibres and ultimately fascicles, 29 which occupy most of the tendon volume. As such, these tendon subunits are commonly analysed as linear elastic structures. 30 The interfascicular matrix (IFM) that surrounds and links the fascicles is a loose connective tissue that is comprised of type III collagen, elastin and proteoglycans, and is produced and maintained by a small population of interfascicular cells. 31 The composition and organisation of the extracellular matrix in different tendons provide optimised mechanical properties for different functions. 29,31,32 Understanding the complex interactions between varying proportions of tendon components to maintain a state of homeostasis is essential in understanding mechanisms by which injury occurs in different tendons. However, these differences in composition and mechanical properties at the different hierarchical scales have historically been challenging to determine due to high variability in the experimental data and lack of techniques to investigate differences, particularly at the lower levels of tendon hierarchy. 33 A significant body of research has more recently focused on elucidating the mechanisms by which tendon functional specialisation occurs and also identifying the age-related alterations which result in increased injury risk. A common approach to investigate tendon structure-function relationships is by the comparison of energy-storing tendons with their anatomically opposing positional counterparts (the equine common digital extensor tendon (CDET) for the SDFT, and human anterior tibialis tendon for the Achilles tendon). These positional tendons are ideal comparators as they experience much lower stresses and strains during locomotion, such that they are not susceptible to overload-induced injuries. 12 Functional differences between positional and energy-storing tendons can be explored by examining the structural composition of the whole tendon, macrostructural and microstructural structures. Due to ethical and technical constraints placed upon in vivo studies, the use of in vitro studies has helped to experimentally elucidate the biomechanical and biothermal thresholds for cell viability, 34 which have resulted in findings that are applicable to both equine and human studies. 35,36 By initially elucidating the compositional specialisation of these differing structures in vitro using the methods described below, researchers can subsequently begin to understand the varying effects of biomechanical and biothermal stresses induced during intense exercise on the SDFT in vivo and establish threshold values for injury risk. The quantification of a strain threshold value of irreversible damage at a microscopic and macroscopic level could subsequently help to predict extraneous factors that influence such mechanical parameters. 37 However, it is important to note that testing cells or tissues in an in vitro setting is not completely analogous to the environment in vivo. There are several important differences which may influence cell viability, which are well recognised in the literature. 38 Furthermore, previous reports investigating mechanisms of tissue physiology and damage often involve various laboratory species, which may not accurately reflect the equine system due to factors including genetic variation and metabolic differences. However, current research in vitro does provide insight into the physiological mechanisms that underpin general tendon function and the mechanism by which injury occurs in vivo.

| B IOMECHANIC AL /S TRUC TUR AL EFFEC TS OF S TR AIN
The effect of locomotory strain on the microstructure of tendons is measured in vitro through mechanical testing. As tendon injury relates to the straining/tearing of tendon tissues, tensile testing is used to determine the maximum stress (force per unit area, measured in MPa) and strain (percent elongation) that the tendon can withstand while being stretched before breaking. Tensile tests can also be used to measure elastic modulus, which is a method of quantifying 'stiffness' or the ability of the tendon to undergo reversible deformation during tensile loading. The elastic strain limit of the tendon is an important strain threshold at which maximum elastic modulus is reached, after which microdamage starts.
However, the majority of injuries are generally thought to occur as a combined result of continuous loading cycles above a certain intensity rather than due to a single overloading event. Matrix microdamage accumulates, overwhelming the capacity of cells to repair structural defects before subsequent loading cycles, 39 ultimately leading to clinical injury. Cyclic loading tests are subsequently used to apply continuous and repeated tensile loads to the tendon tissues to observe the progressive degradation of the material and ultimately tendon failure, and so would be more representative of the nature of stresses applied in a practical setting. 39 By using a combination of tensile and cyclic tests, studies have established how the SDFT is specialised for energy storage at different levels of the tendon hierarchy, and how these specialisations are affected by ageing.

| S PECIALISATI ON AT THE WHOLE TENDON LE VEL
When considering tendons as a whole, energy-storing tendons are more elastic than their positional counterparts, with lower elastic modulus (more compliant) and higher strain to failure, which allows F I G U R E 1 Schematic illustrating the hierarchical structure of tendon tissue, in which collagen molecules aggregate to form subunits of increasing diameter. Diameters shown apply to the equine superficial digital flexor tendon Epitenon Peritenon them to withstand their high strain environment. 40,41 During highspeed galloping, for instance, the equine flexor tendon has typically shown to stretch by 10%-15%. 13 These specialised mechanical properties are accompanied by several reported compositional differences between tendon types, including greater cellularity, glycosaminoglycan (GAG) and elastin content, and lower collagen content in the SDFT compared with the CDET. 12,40,[42][43][44] There are also differences in collagen cross-link profile and collagen fibril diameter between tendon types, with a lower mass average fibril diameter in the SDFT. 12 While it is difficult to directly relate the differences in matrix composition with specific alterations in mechanical properties, it is likely that the higher water and GAG content and lower collagen content in the SDFT result in a more compliant material, while the higher elastin content confers superior recoil. Indeed, it has been demonstrated that water content in the SDFT is negatively correlated with elastic modulus, while fibril diameter shows a positive correlation with elastic modulus. 12 Differences in protein turnover rate have also been identified between tendon types, with a longer collagen half-life in the SDFT, indicating slower turnover of this protein than in the CDET. By contrast, turnover of noncollagenous proteins occurs more rapidly in the SDFT than in the CDET, which may act to repair any damage to the noncollagenous matrix. 44 Similarly, it appears that collagen turnover in the human Achilles tendon is also slow or even absent in the adult, with studies indicating minimal collagen turnover in this tendon once maturity is reached. 45 However, comparative studies of protein turnover in functionally distinct human tendon are yet to be conducted, and turnover rates of specific proteins within tendons from large species remain unknown.

| S PECIALISATI ON AT THE M ACROSC A LE
Recent studies have further elucidated the structure-function relationships in the tendon at the macroscale by assessing mechanical and compositional specialisations of subunits of functionally distinct tendons. Unexpectedly, fascicle mechanical properties do not reflect those observed at the whole tendon level, with fascicles from the SDFT failing at lower strains than those from the CDET, despite the SDFT as a whole demonstrating significantly greater extensibility than the CDET. 41 This greater extensibility is provided by the IFM, which exhibits low stiffness behaviour in the SDFT, allowing sliding and recoil between fascicles. 29,41 Recent work demonstrates similar capability for interfascicular sliding in the human Achilles tendon, with the presence of discontinuities within the tendon observed by ultrasound indicating a complex strain environment within the IFM. 46 The IFM in energy-storing tendons also exhibit greater ability to resist and recover from cyclic loading, with less energy loss (hysteresis) and stress relaxation compared with the IFM in positional tendons, both in horses and humans, 29,47 as well as increased fatigue resistance in the SDFT compared with the CDET. 48 This behaviour is likely provided by the specialised composition of the IFM, which is rich in lubricin and elastin. 32,43,49 Lubricin is a large mucopolysaccharide, which enhances joint lubrication, and mouse knockout studies have demonstrated that lubricin promotes interfascicular sliding in the mouse tail IFM. 50 It is likely that elastin provides the IFM with efficient recoil. Indeed, the elastin content in the IFM is greater in the SDFT than in the CDET, 43 and enzymatic depletion of elastin in the subunits of the SDFT reduces IFM fatigue resistance and the ability to recover from loading, but does not affect fascicle mechanical properties. 51 Studies have also shown that protein turnover rate is greater in the IFM than within the fascicles, which is likely a mechanism to repair damage to this region. 52,53 While the fascicles in the SDFT do not appear to be specialised to enhance tendon extensibility, they are optimised for improved energy-storing capacity, with a helical structure providing greater compliance and enabling more efficient extension and recoil, as well as enhanced fatigue resistance of the fascicles, properties that are likely conveyed to the whole tendon. 48,54,55 By contrast, fascicles from the CDET, which do not have a helical structure, rely on sliding between collagen fibres to allow fascicle extension, and show a much poorer ability to recover from loading, resulting in decreased fatigue resistance. 48,55 While these structural differences in fascicles from functionally distinct tendons have been identified, any compositional differences in the fascicular compartment of functionally distinct tendons are yet to be determined.

| S PECIALISATI ON AT THE MI CROSC ALE
Very little is known about any differences in structure and/or composition between energy-storing and positional tendons at the microscale. However, recent work has shown that collagen fibrils from functionally distinct bovine tendons respond differently to applied elongation, with fibrils from energy-storing tendons demonstrating high strain stiffening and ability to resist molecular disruption, while fibrils from positional tendons are able to extend further, but suffer increased damage characterised by formation of kinks. 28,56 These properties may be conferred by differences in cross-link profile, with more thermally stable and greater total cross-link density in flexor tendons compared with extensors. 28 Differences in collagen crimping have also been observed, with shorter crimp and greater crimp angle in the SDFT compared with the CDET, which likely confer the superior recoil ability in the SDFT. 57

| EFFEC T OF AG EING ON FUN C TIONALLY DIS TIN C T TENDONS
Previous studies have found that the risk of SDFT injury increases with age. 8,9 As such, a significant body of work has been undertaken to establish how tendon properties alter as a function of age, and identify differences in ageing response in functionally distinct tendons. 43 A small number of studies have also investigated the effect of ageing on rodent tendons, demonstrating a decrease in tendon mechanical properties with age in the rat Achilles 67 and increased anterior tibialis stiffness in mice. 68 Gene and protein levels of lubricin and elastin in the rat Achilles and anterior tibialis tendons are also decreased with age; this is accompanied by a decrease in collagen expression at the gene, but not the protein level. 69 Interestingly, in this study, no differences were observed between different tendons with ageing; however, the degree to which the Achilles tendon in the rat is specialised for energy storing remains unknown. In addition, mechanisms of ageing are likely to differ between larger animals and rodents, in which growth plate closure does not occur and skeletal growth continues slowly after puberty, 70,71 such that care must be taken when translating these findings to larger animals.

| EFFEC T OF AG EING ON FA SCI CLE S AND INTERFA SCI CUL AR MATRIX
The effect of ageing on tendon subunits has also been investigated in the horse, with almost all of observed changes occurring to subunits of energy-storing tendons, and many of these occurring specifically to the IFM. The IFM in the aged SDFT exhibits alterations in a range of quasi-static and viscoelastic mechanical properties, including increased stiffness, decreased fatigue resistance and ability to recover from loading. 66 Age-related changes have also been observed in fascicles from energy-storing tendons, with those from the aged SDFT exhibiting increased stress relaxation, with altered response to loading leading to decreased fatigue resistance. 29,72 This is likely due to a loss of helix structure in these fascicles, meaning that when load is applied, extension occurs due to fibre sliding, leading to decreased efficiency and ability to recover from loading. 73  Fibromodulin and collagen III both act as regulators of collagen type I fibrillogenesis 74,75 ; therefore, an age-related decrease in these proteins may reduce the capacity for fibril formation and repair. In contrast to the IFM, there are no alterations in protein cleavage with increasing age in the fascicular matrix, indicating no change in protein turnover in healthy tendon. 52 However, the decreased levels of fibromodulin and collagen III may impair tendon healing, leading to chronic tendinopathy.

| B I OTHERMAL EFFEC TS OF S TR AIN
Adaptive processes succeeding biomechanical strains in vivo involve complex and intricate interactions between tendon tissue and vascular, nervous and immune systems, 76,77 and involve inflammatory agents which play a key role in the early stages of healing. It has been suggested that recurring short-term thermal insults may stimulate stress protein production, protecting tenocytes from damage in later exercise bouts. 35 These proteins are produced in response to sublethal chemical, metabolic or thermal insults and protect the cells from subsequent, otherwise lethal, conditions. 34 However, prolonged or repetitive exposure to hyperthermic conditions damages tendon cells that play an important role in maintaining the extracellular matrix. 38 In vitro tests can be used to measure the biothermal effects of strain by exposing ex-

| PATHOPHYS IOLOGY OF S DF T INJ URIE S
As previously explained, the viscoelastic properties of healthy tendon ideally allow it to extend without damage and recover its macroscale elastic properties after loading. Therefore, damage can be defined as a threshold value at which irreversible changes occur due to application of strain, 73,78 causing structural disruption to the site. This threshold value is referred to as the 'metabolic tipping point', 79 where the metabolic demands of the tendon during mechanical stress exceed the rate at which the nutrient  84 The first stage involves initial debridement of damaged materials at the site of injury due to the expression of catabolic and proinflammatory genes, 82 followed by scar tissue deposition. The inflammatory phase is important for proper tendon healing to occur, 59 and studies suggest that the inflammatory response is better modulated than completely suppressed in humans and canines. [85][86][87][88] Administration of nonsteroidal anti-inflammatory drugs (NSAIDs) immediately post-injury to decrease acute inflammation could otherwise significantly influence the quality of the subsequent healing response. 59,89 In the proliferative phase, a provisional matrix consisting predominantly of type III collagen is synthesised by fibroblasts, which differs structurally and functionally from normal tendon structure. 90 properties in mice. 81 In contrast, histologic studies of tendon removed from mature human patients suffering from tendinopathy have revealed alterations to collagen structure-in place of continuous, well-aligned, crimped collagen fibres, tendinopathic tissue features fragmented, disordered collagen matrix, often lacking clear fibre structure. 36 Mechanical changes, such as a decrease in elastic modulus, 22 are also characteristics of tendon injury.
While the preliminary formation of scar tissue between tendon ends provides basic functional continuity at the site of disruption, subsequent remodelling during the remodelling phase causes conversion of the provisional matrix to type I collagen (scar tissue) so that it is more capable of contracting as it matures, providing increased mechanical strength. However, studies have found comparatively disordered collagen structure at the fibre level, and a lack of compartmentalisation on a macroscale level. 93 Proliferation of the scar between the tendon and adjacent tissues is also undesirable because as these attachments later impede normal tendon gliding and function. 94 Adhesion formation results in loss of movement, and consequently higher risk of functional disability, 95  and disability. 92 This may explain (in part) why 23%-67% of horses treated using conservative methods re-injure tendons within 2 years of the initial injury. 96    performance, establish a source of pain or discover musculoskeletal overloads. 105,106 Interest is recently emerging for further application of IRT due to its noninvasive nature and ability to promptly assess skin temperature, both of which are particularly beneficial when monitoring athletes in training. 107 While IRT cannot reveal specific pathologies, it facilitates the localisation of increased heat (inflammation), 108 commonly referred to as 'hot spots'. The lack of insulating tissues overlying the tendon in the equine distal limb also provides a particularly accurate representation of internal temperature. In the acute phase, tendonitis is identified as a focal 'hot spot', later decreasing in temperature, but remaining elevated during the healing process. 102 This technique has already shown promise as an early screening tool in soccer players, whereby IRT detection protocols were implemented during preseason training. 109 Early detection of subclinical degenerative changes resulted in reduced incidences of injuries reported, as well as days lost to training due to injury.

| Monitoring biothermal effects of strain on the SDFT in vivo
Routine thermographic evaluation of performance horses may also assist in detecting/preventing injury during training and providing an opportunity for preventative action. 'Hot spots' have been detected thermographically up to 2 weeks prior to clinical evidence of any swelling or pain in the horse. 108 Most recently, the use of IRT as an early screening tool revealed significant temperature differences in racehorses before clinical manifestation of metacarpal conditions associated with training occurred. 110 However, to date, there is still very limited research conducted on the use of thermography in equine athletes and further research is required to establish reference baseline profiles. Thermographic evaluation of the distal limb is complicated by environmental factors, such as airflow 111 and ambient temperatures, 112,113 where surface temperature of distal limbs at rest correlates strongly with ambient temperature. 114 Differences of up to 5°C between SDFT core and skin temperature measurements during cryotherapy treatments also suggest a delayed response in skin temperature to extreme temperature changes, which need to be accounted for when using IRT. 115 In addition to the effects of ambient temperatures on the interpretation of results, individual factors, such as breed, coat colour, a moist/dirty coat and time spent alternating weight on limbs during rest, have previously affected interpretation of thermal images. 106 The generally high degree of symmetry between contralateral parts of the body in the healthy horse nonetheless provides a valuable asset in diagnosis of unilateral pathological conditions associated with inflammatory responses. Any thermal asymmetry could indicate abnormalities, although this would be absent in cases with bilateral pathology-the prevalence of which is 10%-35% in the SDFT. 99,116 A difference greater than 1°C over 25% of the compared distal parts of the limbs 105 or a threshold temperature variation of

| Early intervention strategies: Reducing biomechanical stresses
Guidance has been published on the effect of artificial surfaces on the musculoskeletal system, 4 possibly due to the rise in reported musculoskeletal injuries since the increased use of artificial surfaces.
Research on the effects of surface type on performance have been investigated in flat racing, 119,120 over hurdles, 121 high-speed trotting 122 and showjumping. 26 However, epidemiological data specifically related to injuries associated with various surfaces used across the equestrian disciplines still require further investigation. Studying hoof-surface interactions, in particular, can elucidate biomechanical stresses experienced by the limb that contribute to current knowledge to aetiology of musculoskeletal injuries in field conditions. 4,123 Arena surface assessment made between professional showjumping riders also demonstrated a substantial level of inter-rider variation, 124 which further suggests a lack of objective guidance available to inform riders and trainers of appropriate surface quality.
In an attempt to minimise biomechanical strains during exercise, a recent study showed that 85.6% of racehorse trainers in Australia use water-based exercise 125 and swimming, in particular, is emerging as an accepted part of training programmes for racehorses. 126 Other semi-flotation forms of water-based exercise include water walkers, ridden 'surge' training 125  water treadmills and water walkers may be suitable at an earlier stage of the fittening programme before ridden work commences.
The use of water treadmills specifically for the rehabilitation of musculoskeletal injuries in sport horses has been reviewed and includes recommendations and contraindications for use. 132

| Early intervention strategies: Reducing biothermal stresses
Understanding the factors that initiate inflammatory responses and managing the frequency of hyperthermic insults during training in preparation for elite levels of performance may assist in the prevention of SDFT injury. Adaptive processes in vivo after intense exercise involve complex and intricate interactions between tendon tissue and vascular, nervous and immune systems, 76 however, a standardised method of application that has been verified in the research is still yet to be established.

| CON CLUS ION
The SDFT is an energy-storing structure essential for efficient locomotion and thus performance, yet has very narrow mechanical margins for error, making it susceptible to injury. Our understanding of the pathophysiology of SDFT injuries has developed in recent years.
However, this has not translated to a reported reduction in injury rates in the performance industry. Regardless of the varying levels of knowledge and experience among industry stakeholders, the literature has strongly suggested that trainers are considerably more interested in measures to prevent SDFT injury rather than treatment due to the destructive nature of this injury. 97 Therefore, implementing evidence-based training practices may help to standardise protocols and decrease the risk of musculoskeletal injuries during training programmes. This review recommends considering some key areas that require further investigation by researchers, so that recommendations can be made for training programmes (in line with the principles of training 2 ) that specifically aim to prevent incidences of SDFT injury.

CO N FLI C T O F I NTE R E S T S
No competing interests have been declared.

E TH I C A L A N I M A L R E S E A RCH
Not applicable.

OWN E R I N FO R M E D CO N S E NT
Not applicable.

FU N D I N G I N FO R M ATI O N
This research did not receive any specific grant from funding agencies in the public, commercial or not-for-profit sectors. C. Thorpe is

DATA ACCE SS I B I LIT Y S TATE M E NT
Not applicable.

PE E R R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1111/evj.13331.