Science overview: Development of a structural and functional understanding of the equine back


  • P. R. Van WEEREN,

    Corresponding author
    1. Department of Equine Sciences, Faculty of Veterinary Medicine, Utrecht University, The Netherlands; School of Veterinary Science, Faculty of Health and Life Sciences, University of Liverpool, Leahurst, UK; and Veterinary Medical Center, Orthopaedic Research Center, College of Veterinary Medicine and Biomedical Sciences, Colorado State University, USA.
    Search for more papers by this author
  • C. McGOWAN,

    1. Department of Equine Sciences, Faculty of Veterinary Medicine, Utrecht University, The Netherlands; School of Veterinary Science, Faculty of Health and Life Sciences, University of Liverpool, Leahurst, UK; and Veterinary Medical Center, Orthopaedic Research Center, College of Veterinary Medicine and Biomedical Sciences, Colorado State University, USA.
    Search for more papers by this author

    1. Department of Equine Sciences, Faculty of Veterinary Medicine, Utrecht University, The Netherlands; School of Veterinary Science, Faculty of Health and Life Sciences, University of Liverpool, Leahurst, UK; and Veterinary Medical Center, Orthopaedic Research Center, College of Veterinary Medicine and Biomedical Sciences, Colorado State University, USA.
    Search for more papers by this author



The thoracolumbar region of the equine spine provides an important structural and functional component of the locomotor apparatus, especially as it relates to ridden exercise. Relative to the appendicular skeleton, the equine back has received little attention over time related to clinical diagnosis, imaging or treatment approaches. Fortunately, in the last few decades significant advances have been made in further understanding of the role that the thoracolumbar spine plays in athletic performance and in the transfer of biomechanical forces between the thoracic and pelvic limbs during normal locomotion and during select limb lameness or back pain conditions. The purpose of this article is to review the development of biomechanical modelling of the quadrupedal trunk and to provide insights into how future diagnostic and therapeutic modalities may be best focused on prevention, management and optimisation of spinal function in the ridden horse.

The biomechanical concept of the quadrupedal back in general

The biomechanical concept and perceived function of the quadrupedal back has changed over time. The Roman physician Galen (129–200 AD) described the first known concept of the thoracolumbar spine (cited by Slijper 1946). Galen refers to the prevailing architecture of his day and describes the quadrupedal back as a vaulted roof sustained by 4 pillars; the limbs. The next concept of the quadrupedal trunk dates from the middle of the 19th century and was again inspired by the technical advances in engineering of that era. In the so-called bridge concept of the equine back, the limbs are considered to form the land abutments of the bridge and the gap between these abutments is spanned by the bridge itself (Bergmann 1847; Zschokke 1892; Krüger 1939). The bridge concept of the quadrupedal trunk has dominated the veterinary and zoological literature for a long period, but it contains some inherent errors associated with specific spinal elements that are theorised to undergo compression or tension within the bridge model. More recently, the bridge model has been replaced by the ‘bow-and-string’ concept, in which the bow is represented dorsally by the relatively inflexible thoracolumbar vertebral column and the string is formed ventrally by soft tissues, consisting of the linea alba, the rectus abdominis muscle and related trunk structures. The bow-and-string model was first proposed by Barthez (1798), but was largely ignored until it was rediscovered by Slijper (1946) (Fig 1). Although the concept does not account for some specific anatomical aspects, most notably flexion-extension at the lumbosacral junction or pelvic rotation about the coxofemoral articulation (Fig 2), it is still the most widely accepted biomechanical model to date and its validity has also been confirmed by recent spinal research in the horse.

Figure 1.

‘Bow-and-string’ concept of the back according to Slijper. The vertebral column is the bow and the ventral musculature and sternum are the string. The ribs, lateral abdominal musculature, spinous processes and ligamentous connections are additional elements (Slijper 1946).

Figure 2.

Model of the equine back with induced extension without a) and with b) consideration of movement at the lumbosacral junction and coxofemoral articulations. Without motion at the lumbosacral junction, peak vertical displacements would be theorised to occur at the mid-point of the back. With motion at the lumbosacral junction and coxofemoral articulations, peak displacements would be expected to occur at the lumbosacral junction, which is supported by current spinal kinematic research (Haussler et al. 2007). (Nickel et al. 1986).

The beginning of equine biomechanical research related to the back

After the revival of interest in equine locomotion research in the early 1970s (Fredricson and Drevemo 1971) that formed the start of what has been called the ‘Second Golden Era of equine gait analysis’ (van Weeren 2001), it took a long time before the first publications reporting in vivo equine spinal kinematics appeared. The first study of this kind after the preWWII work by Berlin-based Wilhelm Krüger (1939), who fixed a camera in the branches of a tree to get a good view from above, was probably the work by Audigiéet al. (1999), who used the technique developed by Pourcelot et al. (1998). They used spherical markers fixed to the skin along the dorsal midline of the horse at more or less equal distances from the withers to the sacrum and were able to describe 2-dimensional or flexion-extension movements of the thoracolumbar spine in trotting horses. However, to fully describe and understand spinal movements, 3-dimensional vertebral motion needs to be evaluated: flexion-extension (i.e. rotation about the transverse or X-axis), left and right lateral bending (i.e. rotation about the vertical or Z-axis), and left and right axial rotation (i.e. rotation about the longitudinal or Y-axis that points in the direction of movement) (Fig 3). For a more extensive description of in vivo spinal kinematics see van Weeren (2009). Of these 3-dimensional movements, flexion-extension and axial rotation of the sacrum ascalculated from skin markers on the tuber coxae can be measured with acceptable accuracy when using skin markers as skin-movement artefacts are minor (Faber et al. 2001c). When skin markers move laterally, it cannot be determined to what extent the marker movement is due to lateral bending of the underlying vertebrae (or dorsal spinous processes) or due to axial rotation, since lateral bending and axial rotation are coupled movements within the thoracolumbar spine, which always occur in varying combinations due to articular characteristics and soft tissue connections between adjacent vertebrae.

Figure 3.

The 3 basic movements of the equine back include flexion-extension/FE a), left and right lateral bending/LB b) and left and right axial rotation/AR c).

Role of the epaxial musculature in the biomechanics of the back

Much of the biomechanics studied to date have been focused on the kinematics of the osseous vertebrae. However, in line with human back pain research, the biomechanics and function of the epaxial musculature, in particular the deep multifidus muscles in actively stabilising (or reducing unwanted motion of) the equine spine during movement, has also been studied in horses (Stubbs et al. 2006, 2010). Lumbo pelvic stability in man has been clearly shown to be dependent not only on the contribution of the passive elements (including the intervertebral discs, ligaments, joint capsules and facet joints), but also on the active elements (muscles) and the controller (nervous system), which involves a complex relationship between neural and muscular (motor) control of movement and is referred to as neuromotor control (Panjabi 1992).

In man, all of the many muscles of the back and trunk contribute to spinal control and support, but the deep lumbar multifidii have a major role in this respect, contributing two-thirds of the total increase in spinal stiffness of the lower back imparted by muscular action (Wilke et al. 1995). Stubbs et al. (2006) in a series of dissections showed in detail the 5 fascicles of the equine multifidii, with the direction of the fibres and orientation (force vectors) similar to the human anatomy, leading to the assumption that the equine multifidii have a similar role to that in man, providing caudal extension of their vertebra of origin as well as the deeper fascicles providing intervertebral compression within the thoracolumbar spine (Stubbs et al. 2006).

The horse, however, differs from man in several respects. Firstly, the horse has a tail and the multifidus complex continues right through the sacrum to the caudal vertebrae as the sacrocaudalis dorsalis medialis and lateralis muscles with fascicles that mirror that of the multifidus fascicles (Stubbs et al. 2006). Secondly, in the horse, the passive stability of the spine is high, due to the semi-rigid anatomy of most of the lumbar spine (Jeffcott and Dalin 1980). Therefore, the requirement for muscular control of the equine lumbar spine may well be more limited over much of its length. However, there is much evidence that the role of the multifidii is far from negligible. The bulk of the overlapping fascicles of mutifidii mirror the degree of motion found on biomechanical studies of the spine. In particular there is increased fascicle size associated with the increased motion at the lumbosacral junction (Stubbs et al. 2006). Furthermore, there is the presence of a bursa between the long tendinous cranial attachments of the sacrocaudalis dorsalis lateralis muscle and the last dorsal spinous process in the majority of horses. The bursa is present as the caudal extension of the multifidus complex, the sacrocaudalis dorsalis lateralis, crosses the lumbosacral junction, suggesting considerable motion and forces imparted on the spinous processes of this region (Stubbs et al. 2006). Thirdly, it has been found that the biomechanics of the lumbosacral region may vary considerably depending on underlying anatomical variations, particularly associated with divergence of the dorsal spinous processes located at the lumbosacral junction. Over a third of Thoroughbreds had variations in the lumbosacral junction associated with significant changes in the angle of the spinous process of L6 relative to the vertebral body, thus affecting the mechanical function (moment arm and force vector) of the multifidus muscle fascicles attaching to the last lumbar spinous process (Stubbs et al. 2006). The combination of anatomical variation and altered force generation by the multifidii may result in altered motion at the lumbosacral junction and potentially have an impact on performance or the risk for pathology and the development of back pain.

Further development of kinematic techniques to study back motion

Marjan Faber and co-workers were the first to comprehensively address the problem of skin displacement in relation to equine spinal kinematics. Firstly, they developed a method that allowed for the determination of 3-dimensional spinal kinematics without defining a local vertebral coordinate system (Faber et al. 1999). This method was applied to data generated by the motion capture of markers mounted on Steinmann pins inserted into the dorsal spinous processes of a number of thoracic, lumbar and sacral vertebrae and into both tuber coxae. In this way, a rigid connection between markers and underlying bony structure was created, thus avoiding artefacts due to skin displacement. Measurements were performed at walk, trot and canter, thus providing baseline spinal kinematic data for the 3 basic gaits of the horse (Faber et al. 2000, 2001a,b). These invasively-acquired data using markers attached to Steinmann pins formed the basis for the development of a clinically applicable method that used skin-fixated markers to measure 3-dimensional spinal kinematics (Faber et al. 2001c). The method was tested for repeatability by measuring the same horses on 5 consecutive days and for environmental influences by using the measurement system in 2 different laboratory settings with 2 different breeds of horses (i.e. Warmbloods and Standardbreds). Between-stride and day-to-day variability were low with inter-individual variability being considerably larger and differences between the labs and the breeds were small, which confirmed expectations. It was concluded that the skin-marker method provided reliable data with high degree of repeatability and hence was suitable for clinical use (Faber et al. 2002). The method calculates the range of motion (ROM) for the 3 basic rotations across several vertebral levels, which give an indication of overall spinal mobility. The method also determines ‘angular motion patterns’ (AMP), which describe the position in space of a given vertebra with respect to the horizontal (for flexion-extension) or the sagittal plane (for lateral bending). Therefore, AMPs provide information about the position of the vertebral column (flexed vs. extended; straight vs. scoliotic) at a given moment in time. This method has been optimised and eventually marketed by the manufacturers of the ProReflex1 kinematic gait analysis system as a user-friendly customised software programme. Other researchers have used a Steinmann pin and linear transducer kinematic system placed into adjacent dorsal spinous processes to provide an assessment of segmental kinematics within individual vertebral motion segments (Haussler et al. 2001). The abovementioned spinal kinematic methods have been used both for fundamental research into kinematics of the equine vertebral column and for more applied clinical research. These methods have been used for the kinematical characterisation of the health status of the back, for the evaluation of the effect of various therapies, for evaluation of the influence of tack and the assessment of the effects of a variety of equestrian activities and training techniques. The last area of research is of increasing importance, both from the viewpoint of quantification and improvement of athletic performance and from an animal welfare perspective.

Spinal kinematics and back pain

As a prelude to work in diseased animals, Johnston et al. (2002, 2004) studied normal horses and the influence of physiological factors such as conformation, use and gender on back motion. As expected, there was a relationship between conformation and spinal kinematics. Long-backed horses showed more lateral bending. Further, there was a negative correlation between the curvature of the mid-thoracic back and lateral bending at L1 and L3 and axial rotation at the pelvis, the (clinical) significance of which is unclear. Finally, horses with larger flexion-extension ranges of spinal motion had a greater stride length as is seen in other high-speed quadrupeds such as racing greyhounds or cheetahs (Johnston et al. 2002). Of possible physiological influences, age and gender had a certain influence. Flexion-extension range of motion (ROM) decreased with age, indicating that older horses had stiffer backs, which is perhaps not very surprising. Further, mares showed increased lateral bending at T10 compared to geldings, but less at L5, again a finding of which the clinical significance is as yet unclear. Dressage horses had significantly increased lateral bending than showjumpers, which would be important for lateral spinal movements and less for the flexion-extension needed during jumping (Johnston et al. 2004).

Back pain and the diagnosis thereof has been a controversial item for decades in equine orthopaedics because of the difficult accessibility of the equine back and the high level of subjectivity involved (Jeffcott 1979) and the perceived prevalence is high. In clinical practice the diagnosis of back pain is still largely made on subjective criteria, such as the response to palpation and the qualitative assessment of back motion, although recently pressure algometers have been introduced to quantify mechanical nociceptive thresholds, essentially based on secondary muscle change rather than on primary vertebral pathology, which may help in objectifying the diagnosis and treatment of back pain (Haussler and Erb 2006). Wennerstrand et al. (2004) have tried to use kinematics to discriminate between sound horses and horses with subjectively diagnosed back pain. They found a reduction in flexion-extension and axial rotation ROM in the affected horses, concomitant with a decrease in stride length. These findings agreed with earlier clinical observations (Jeffcott 1980) and spinal kinematic data from a single case report (Faber et al. 2003).

With the idea of working on cases with better defined back pain, Wennerstrand et al. (2009) used unilateral, lactic acid injections into the epaxial musculature as a model to induce controlled and transient back pain, a technique originally described by Jeffcott et al. (1982). The authors of both studies did not observe any effects on linear or temporal stride characteristics; however, back kinematics was clearly affected. A 2-phased response was evoked consisting initially of scoliosis that was concave towards the injected side indicating shortened, spasmed epaxial muscles, which was followed by lateral curvature toward the contralateral side. This phenomenon was explained by the difference between the acutely induced pain and the ensuing muscle stiffness or weakness during the following few days and could be considered to support the important role of the musculature in vertebral biomechanics (Wennerstrand et al. 2009). The mechanism by which the paraspinal injections influence back motion may be less straightforward than by the means of local chemical irritation alone. In a spinal kinematic investigation into the effect of infiltration with local anaesthetics into the interspinous spaces, which is commonly done to diagnose impinged or kissing dorsal spinous processes, an increase in both flexion-extension and lateral bending ROM was noted in both local anaesthetic and saline-control groups (Roethlisberger Holm et al. 2006). These effects may be mediated through alteration of proprioception evoked by both the volume of liquid and direct local anaesthetic effects administered within the multifidus muscle. This muscle is known to be of great importance for proprioception and postural control in man (San Juan et al. 2005) and there are strong indications that the same functions apply to horses (Stubbs et al. 2006).

The function of the multifidus muscle has been shown to be affected in low back pain in people, which is reflected in morphological changes in the muscle. The cross-sectional area (CSA) of multifidus measured using ultrasonography has been shown to decrease on the painful side, and at the clinically determined level of symptom provocation (Hides et al. 1994) and this reduction can be detected within 24 h of injury (Hides et al. 1996). It is notable that there is not an automatic resolution of these changes following the resolution of pain. Despite apparent recovery or resolution of pain following an episode of acute back pain, the dysfunction of multifidus persists (Hides et al. 1996), potentially leading to secondary altered spinal biomechanics and neuromotor control and the development of chronic back pain.

A recent study also published within this supplement has shown a similar correlation as described in man between multifidus CSA measured ultrasonographically with osseous pathology in Thoroughbred racehorses (Stubbs et al. 2010). Osseous lesions included fractures; spondylosis and lysis and/or periosteal new bone formation of the thoracolumbar vertebral body or end plate; facet joint or articular process; dorsal spinous process or lumbar transverse process; lumbosacral complex; and sacroiliac complex. Each lesion was graded as mild, moderate or severe where active new bone formation was required in order to grade a lesion as severe. In this study, there was a significant association between pathological grade and the degree of multifidus asymmetry with measurable left-to-right asymmetry in multifidus at or close to the level of osseous pathology, with the smaller multifidus CSA corresponding to a higher grade of pathology ipsilateral. Further, this asymmetry was correlated with severe lesions as opposed to lesions of a lower severity (Stubbs et al. 2010). These findings support both the functional importance of the multifidus muscles and that ultrasonography of the epaxial musculature may be a future adjunct diagnostic tool for equine back pain.

It is interesting to note that, where the rather drastic intervention of injecting strong lactic acid did influence spinal kinematics to only a modest extent and had no influence on stride length or stride time, even a very subtle lameness can readily affect back motion. The influence of lameness on back motion had been demonstrated earlier in horses with evident lameness (Buchner et al. 1996; Pourcelot et al. 1998), but it was shown more recently in studies on the effect of induced fore- and hindlimb lameness that this was true for very subtle and hardly perceptible lameness as well (Gómez Álvarez et al. 2007, 2008a). It was concluded from these studies that subclinical or mild lameness may well play a role in the pathogenesis of back pain, as earlier suspected based on clinical experience (Landman et al. 2004). However, the caveat should be made that in the experimental studies the induced lameness was acute; whereas, in most clinical cases lameness is often chronic. The interactions and pathogenesis of lameness-induced spinal dysfunction and back pain-induced altered gait need further study.

Therapy evaluation

Biomechanical analysis of spinal motion has to date only been used to evaluate the effect of various forms of manual manipulation. Haussler et al. (1999) used an invasive technique, in which Steinmann pins were implanted into a number of spinous processes that were connected by liquid metal strain gauges (Haussler et al. 2000), to show that chiropractic manipulation indeed could affect vertebral motion. In later work this research group demonstrated that chiropractic manipulation could increase flexion-extension spinal mobility in horses in which back pain was induced by the implantation of Steinmann pins (Haussler et al. 2007) and in ridden horses without experimentally induced back pain (Haussler et al. 2010). Chiropractic manipulation has also been shown to have a more pronounced beneficial effect on increasing mechanical nociceptive thresholds than massage or treatment with phenylbutazone (Sullivan et al. 2008).

The kinematical analysis method originally developed by Faber et al. (see above) was used by Gómez Álvarez et al. (2008b) to assess the effect of chiropractic manipulations on a number of horses with alleged back problems. The authors succeeded in demonstrating a decrease in extension of the thoracic back and a decrease in pelvic inclination, together with an increase in symmetry of the pelvic motion pattern. However, differences were small and not all of them were still detectable during a second measurement session 3 weeks later. The best documented study is a single case report in which a horse was treated twice and measured 5 times (from before the start of treatment to 7 months later). Clear and to a large extent lasting improvements in symmetry of movement were demonstrated, but a caveat was made that eventual clinical improvement seemed more related to a change in trainer than to the chiropractic treatment itself (Faber et al. 2003).

The influence of tack on back motion

The effect of a girth, an empty saddle and a saddle loaded with 75 kg of lead on equine spinal kinematics has been measured by De Cocq et al. (2004). The authors reported little influence from the girth and the saddle-only condition, but showed an increased extension of the back without a change in overall ROM with the weighted saddle. Horses counteract the increase in thoracolumbar spinal extension by increased retraction of the forelimbs. The forelimb retraction has, according to the bow-and-string model of the quadrupedal back as proposed by Slijper (1946), an effect of inducing a compensatory flexion moment on the vertebral column. There are many studies using a variety of pressure-sensing devices that aim at determining pressure patterns underneath a saddle on the equine back and possible influences thereon. These pressure patterns often do influence spinal motion patterns and vice versa, but a review of these studies does not fall within the scope of this paper. For a complete review of this topic see De Cocq and van Weeren (2011).

The effect of weighted boots on the movement of the back in asymptomatic riding horses has also been reported at the walk and trot (Wennerstrand et al. 2006). Weighted boots placed on the front feet affected the movement of the thoracic region of the spine; whereas, back boots placed on the hind feet increased the flexion-extension movement of the lumbar region, compared to when the horses did not wear boots. At a trot, the weighted hindlimb boots decreased the protraction and retraction of the hindlimbs, but did not significantly alter the stride duration.

Biomechanics of the equine back and performance

Hippotherapy is a method of treatment used by physiotherapists, which involves horses as treatment tools. Kinematic analysis of the horse's back has been used to assess the relationship between the walking speed of the horse and the vertical and horizontal displacements of the horse's back (Janura et al. 2010). The length of the step, vertical displacement of the pelvis and stride frequency were greater during the fast walk in comparison with the slow walk, which is of clinical relevance for hippotherapy patients.

Cassiat et al. (2004) demonstrated via spinal kinematic techniques developed by Pourcelot et al. (1998) that lower-level show jumpers had an increased flexion of the thoracolumbar and lumbosacral junction before take-off compared to higher-level jumpers, which might result in a less efficient strutting action of the thoracic limb as forward momentum is converted into upward movement. Santamaría et al. (2004) showed in a 5 year longitudinal study that jumping technique, including the use of the back, was well preserved from foal age to maturity. In that study, back motion in itself did not discriminate between good and bad jumpers, but retraction of the pelvic limbs relative to the back (i.e. coxofemoral joint extension) at the moment of clearing the fence did (Bobbert et al. 2005).

An interesting item was addressed (in part) by biomechanical research in dressage. Following the example of showjumping, many trainers and riders in dressage in the late 20th century had begun to include riding with a very low and strongly flexed head and neck position in their training protocols. This method, that became known as ‘Rollkur’ (Meyer 1992), ‘hyperflexion’ (Jeffcott et al. 2006), or ‘low, deep and round’ (LDR, Janssen 2003), became heavily criticised (strangely enough only in dressage; whereas the technique is widely used in showjumping) by animal rights activists, who claimed it was a significant equine welfare issue. Kinematic and kinetic research on a number of natural and non-natural head and neck positions made clear that, from a biomechanical perspective, the opposite position (i.e. an extremely extended head and neck position) indeed induced potentially deleterious effects such as higher peak forces on limbs (Weishaupt et al. 2006), but the very low head and neck position with extreme neck flexion did not produce any such effects (Gómez Álvarez et al. 2006; Weishaupt et al. 2006; Rhodin et al. 2009). In fact, the increased flexion-extension ROM in the entire back that was seen in the low head and neck position could possibly be explained as a better way of gymnastic training of the horse than when only ridden with more conventional head and neck postures (Gómez Álvarez et al. 2006). This research contributed, together with work on respiratory and behavioural aspects with regard to this exaggerated head and neck position, to the eventual decision by the Fédération Equestre Internationale (FEI) to henceforth discriminate between ‘Rollkur’ or ‘hyperflexion’ vs. ‘low, deep and round’ head and neck positions. The ‘Rollkur’ or ‘hyperflexion’ terms are meant to indicate an extremely flexed position achieved with undue force or cruel means, which is intolerable from a regulatory viewpoint. However, if the same head and neck position is achieved in a gentle way (Fig 4), the term ‘low, deep and round’ is used, which is an allowed training technique (FEI 2010).

Figure 4.

Application of the ‘low, deep and round’ (LDR) training technique; a strongly flexed neck and lowered head, achieved without excessive force. This head-neck position will induce flexion within the back (Photograph courtesy Sjef Janssen).

The future

Predicting the future has always been a hazardous undertaking. However, if the challenge is to predict the direction or focus of equine biomechanical studies, there are many indicators that may make the crystal ball a little less misty, as they all point toward the same direction. Technological advances have progressed tremendously during the last decades and can be expected to progress at even faster rates of speed in the years to come. Objective methods of assessing back pain and spinal dysfunction are beginning to be developed. The role of optimal spinal function in overall performance and interactions between limb lameness and back pain are becoming better understood. Interest in equestrian sports cannot be supposed to wane unless substantial economic hardships develop in the near future. Further, public interest in the welfare aspects of equestrian activities has increased steadily in the past years and this trend is expected to continue. All these factors support or facilitate research in the area of equine spinal biomechanics. Miniaturisation of measurement devices will, together with increased capacity for data logging and improvement in telemetric systems, enable the transition from the artificial surroundings of the equine biomechanics laboratory to the real-life situation. The use and future development of global positioning devices, magnetometers, gyroscopes and 3-dimensional accelerometers will make evaluation of spinal kinematics and the influence of the rider, tack and training methods easier to objectively assess for both researchers and horse owners/trainers (Valentin et al. 2010). Accurate, real-time measurements of the locomotor system during competitions or racing events will become every-day occurrences. On the other hand, biomechanical modelling and computer simulations of the structure and function of the equine back will most likely take a great flight too. Some preliminary work on modelling of the equine spine has been performed (Peham and Schobesberger 2004, 2006). Accurate validation of these models is still a real challenge; however, there is little doubt that more sophisticated and detailed models of the equine back will be presented in the coming years. The application of novel techniques such as pattern recognition (Schöllhorn et al. 2006) will increase and there will be a trend towards a more integrated approach where data on equine spinal biomechanics will be combined with data recorded from the rider and the horse-rider interfaces, such as saddle pressure data, force data from reins and stirrups, etc. and the horse-ground surface interactions to produce a more comprehensive understanding of the extremely complicated ground-horse-saddle-rider combinations. These interactions will need to be assessed individually during the widely-varied disciplines that constitute modern-day competitive horse riding. When it is possible to generate reliable and repeatable quantitative data on this horse-rider partnership, there will be little left of the once mysterious mists that used to surround the function and dysfunction of the equine back, opening the way to rational and evidence-based preventive and therapeutic interventions.

Conflicts of interest

The authors declare no potential conflicts.

Manufacturer's address

1 Qualisys AB, Gothenburg, Sweden.