Evaluation and treatment of horses with back pain

Back pain is commonly implicated by owners as a cause of poor performance but it is challenging to arrive at a secure diagnosis

01 May 2020, at 8:00am

If there is a clinical suspicion of back pain, the first step is to work out whether it is primary or secondary to another orthopaedic problem such as hindlimb lameness.


There is no single “gold standard” test for back pain. Intuitively, pain on palpation and bucking when ridden are the most frequent signs attributed to it, but horses with back pain may also present with reduced performance, poor transitions, making a poor bascule when jumping or refusing to jump, becoming disunited in canter, alongside a number of other nebulous signs. The author will always address any lameness first, as epaxial muscle pain secondary to lameness is common. Indeed, we know both fore- and hindlimb lameness affects thoracolumbar movement (Álvarez et al., 2007; Álvarez et al., 2008).

The same inertial measurement units (“sensors”) used in gait analysis can be used to quantify back movement in both the non-ridden and ridden horse (Figure 1) (Martin et al., 2016; Guire et al., 2018). This allows both objective lameness and back movement evaluation concurrently.

A number of kinematic studies have been performed which have assessed the response of the back to pain. In a study of horses adjudged to have natural back pain, reduced flexion-extension movement at or near the thoracolumbar junction was observed (Wennerstrand et al., 2004). Recent studies have shown that the thoracolumbosacral range of motion increases after successful diagnostic analgesia of hind limb lameness (Greve et al., 2017) and body posture is more equal between lunging directions after successful diagnostic analgesia (Greve et al., 2018). This supports the clinical observation of the co-existence of lameness and back problems (Landman et al., 2004).

If lameness is not present or has resolved, measurable asymmetric back movement may indicate the presence of back pain. The problem with interpretation of this data is making the link between “active”, relevant back pain and asymmetric movement or reduced range of motion. For the latter we have no reference to which we can compare back movement data. The author measures back movement before and after administration of phenylbutazone. Provided the horse is sound, any change in movement supports the existence of back pain.

Radiograph of a horse's back
FIGURE (2) Radiograph demonstrating insertion of needle ventral to site of bone remodelling

If the horse has a reproducible clinical sign, diagnostic analgesia is the most reliable way of attributing pain to a region. Infiltration of local anaesthetic between impinging spinous processes is commonly used; the author injects 10ml either side of the affected sites under radiographic guidance. Ultrasound guidance can be used but the aim is to inject ventral to the site of impingement to affect the dorsal spinal nerve within the interspinous ligament (Figure 2). If the signs exhibited are too inconsistent, then medication of the suspicious sites with corticosteroids is an option. The author sees this as a diagnostic technique rather than providing a long-term cure.


Radiographic assessment of the thoracolumbar spine should include latero-laterals for the spinous processes (Figure 3A) and latero-20° ventral-laterodorsal oblique projections for the articular process joints (APJs) (Figure 3B). Impinging spinous processes (ISP) have been found in 86 to 92 percent of racehorses at post-mortem and in 37 percent of normal horses, so their presence radiographically is not necessarily associated with thoracolumbar pain. Whilst a number of grading systems exist, the author believes the severity of impingement correlates poorly with the severity of clinical signs. Any remodelling or impingement of spinous processes should be considered a possible source of pain.

Osteoarthritis (OA) of the APJs is most commonly found between T15 and L1, and usually affects between two and five joints. Horses with OA of the APJs are more likely to have thoracolumbar back pain than horses with ISPs, and those with concurrent ISP and OA of the APJs are associated with the highest likelihood of pain. The presence of ventrospondylitis, discospondylitis and fractures should also be assessed.

Acquisition of high-quality ultrasound images of the thoracolumbar spine is now possible due to the continued development of ultrasound machines.

The following anatomical structures can be assessed depending on the probe used:

Linear probe (high frequency: 7.5 to 15MHz):

  • Supraspinous ligament
  • Summits of the spinous processes
  • Thoracolumbar fascia

Convex probe (low frequency: 3 to 6MHz):

  • Articular process (facet) joints
  • Lumbar transverse processes
  • Epaxial musculature

Rectal probe (5 to 7.5MHz):

  • Lumbar vertebral bodies (L4 to L6)
  • Intervertebral discs (L4 to 5 and L5 to 6)
  • Lumbosacral joint
  • Sacroiliac joint

Lameness results in reduced ability to use the thoracolumbar epaxial muscles normally (Greve and Dyson, 2015). Saddle-slip to one side occurs in approximately half of horses with hindlimb lameness (Greve and Dyson, 2013; Greve and Dyson, 2014), which is likely due to asymmetric thoracolumbar movement and/or asymmetric epaxial muscling. Osseous spinal pathology has been shown to cause measurable left/right asymmetry in multifidus at, or close to, the level of pathology in thoroughbred racehorses (Stubbs et al., 2010) and ipsilateral to symptoms in humans with acute/subacute lower back pain (Hides et al., 1994). This suggests that measurement of cross-sectional area of multifidus using ultrasound (Figure 4) could provide an objective indication of the existence of back pathology.


The aims of treatment are to alleviate pain and muscle spasm to allow muscle function and development, develop core muscle strength, improve flexibility and prevent recurrence.

Leo Jeffcott proposed the “bow and string” theory whereby the epaxial muscles – longissimus dorsi and multifidus – provide stability and locomotion and the abdominal muscles flex the back. Retraction of forelimbs and protraction of hindlimbs flex the spine (Van Weeren, 2009). Lowering the head tenses the nuchal ligament, cranially rotates thoracic vertebrae and flexes spine (Berner et al., 2012) whereas raising the head tenses the back. Rehabilitation should focus on flexing the spine (tensing the bow).

Back width increases in spring, in regular work and after a work session whereas it decreases in winter, with a heavy rider and in the presence of lameness (Greve and Dyson, 2014). Therefore, saddle fit should be checked several times a year and taking into consideration the above factors.

Large studies of back pain in people have found that strength/resistance and coordination/stabilisation exercise improves outcomes over other interventions (Searle et al., 2015), exercise improves pain relief and function compared to bed rest (Dahm et al., 2010) and stretching and strengthening results in the largest improvement over other exercise (Hayden et al., 2005).

The superficial longissimus dorsi spasms as it attempts to compensate for the loss of spinal stability (frequently evident on palpation). This is because it is poorly positioned to stabilise the spine. The deeper multifidus stabilises the spine and atrophy of this muscle results in micro-motion of the intervertebral joints thereby predisposing them to the development of degenerative joint disease. Therefore, therapeutic efforts should be made to build up multifidus muscle. One method is to perform dynamic mobilisation exercises (baited stretches).

Whole body vibration (WBV) has been shown to increase multifidus cross-sectional area and reduce left-right asymmetry (Halsberghe et al., 2017). Water treadmill (at fetlock or tarsal level) increases T18 and L3 flexion (tensions the bow) (Nankervis et al., 2016). Initial therapy for impinging spinous processes is usually conservative. Perilesional medication with corticosteroids is the author’s first line conservative treatment followed up with a robust six week programme (Box 1) of non-ridden exercise as we know that rider weight reduces back movement (Greve and Dyson, 2014; Martin et al., 2016).

BOX (1) An example of a non-ridden exercise programme (this should be altered according to the horse’s fitness and ability)
BOX (1) An example of a non-ridden exercise programme (this should be altered according to the horse’s fitness and ability)

Articular process joint OA treatment consists of intra/peri-synovial corticosteroids (injected under ultrasound guidance). Horses undergo a similar exercise programme (Box 1) as following ISP medication. Whilst medication of equine backs is common, systematic reviews of randomised controlled trials in human medicine show no strong evidence for or against the use of any type of injection ”therapy” (Staal et al., 2009). Injection of articular process joints with corticosteroids/local anaesthetic results in a transient improvement in pain but not disability, and are no better than placebo treatments (Chou et al., 2015). Nevertheless, in horses the author finds these injections provide valuable pain relief for a variable period.

Concurrent phenylbutazone medication can be administered if required. RAMP registered physiotherapist visits should be arranged fortnightly alongside dynamic mobilisation exercises. On completion of the programme ridden exercise resumes.

Pitcher plant is reported to provide pain relief without motor weakness via neural blockade. No significant differences in the pain relief or duration of significant relief was seen in human facet joint injections (Manchikanti et al., 2004) and no significant local anaesthetic action has been demonstrated (Campos et al., 2013). However, there are many anecdotal reports of its benefit. Be aware it is only licensed for external use.

Extracorporeal shockwave therapy initiates neovascularisation and upregulation of angiogenetic growth factors (Wang, 2012). It has been shown to be useful for treatment of impinging spinous processes and OA of APJs (Allen et al., 2010). Probe position and correct depth are critical: 35mm probe abaxially for spinous processes and 80mm probe abaxially for articular process joints. One thousand pulses are delivered each side, horses are rested for two days and returned to work over five days.

Mesotherapy is used by many practitioners and consists of multiple intradermal injections with combinations of lignocaine/dexamethasone/saline. It is hypothesised to stimulate type I and II nerve fibres that block pain transmission within the same spinal segment – the gait theory. There is some evidence supporting its use: in humans it was reported to be as effective at treating acute lower back pain as NSAIDs (Constantino et al., 2011). It can be combined with shockwave and exercise (“SME” therapy) for horses with impinging spinous processes (Turner, 2011).


Resection of impinging spinous processes (Figure 5) can be performed either standing or under general anaesthesia with success rates of 70 to 77 percent returning to work (Walmsley et al., 2002; Brink, 2014; Jacklin et al., 2014). Interspinous ligament desmotomy (ISLD) was first described in 2012 (Coomer et al., 2012) with a remarkable 95 percent success rate. A recent study found that 53 percent of horses are at equivalent or higher level of performance three years post ISLD (Prisk and García‐López, 2019).

In summary

The key to successful treatment of back pain is an accurate diagnosis and a diligent rehab regime combining pain relief (either medical or surgical), physiotherapy and a robust exercise regime activating the core stabilisation muscles.

Author Year Title
Allen, A.K., Johns, S., Hyman, S.S., Sislak, M.D., Davis, S. and Amory, J. 2010 How to diagnose and treat back pain in the horse. In: AAEP Proceedings. Pp. 384-388
Álvarez, C.G., Bobbert, M., Lamers, L., Johnston, C., Back, W. and Van Weeren, P. 2008 The effect of induced hindlimb lameness on thoracolumbar kinematics during treadmill locomotion. Equine Veterinary Journal, 40, 147-152
Álvarez, C.G., Wennerstrand, J., Bobbert, M., Lamers, L., Johnston, C., Back, W. and Weeren, P.V. 2007 The effect of induced forelimb lameness on thoracolumbar kinematics during treadmill locomotion. Equine Veterinary Journal, 39, 197-201
Berner, D., Winter, K., Brehm, W. and Gerlach, K. 2012 Influence of head and neck position on radiographic measurement of intervertebral distances between thoracic dorsal spinous processes in clinically sound horses. Equine Veterinary Journal, 44, 21-26
Brink, P. 2014 Subtotal ostectomy of impinging dorsal spinous processes in 23 standing horses. Veterinary Surgery, 43, 95-98
Campos, G.J., Chacón, T.C., Cova, F.J., Flores, S.A., Rojas, J.A., Risso, A.J. and González, H.A.Z. 2013 Evaluation of the local analgesic effects of a commercial aqueous extract of Sarracenia purpurea and ammonium sulfate in the equine abaxial sesamoid block model. Journal of Equine Veterinary Science, 33, 1004-1007
Chou, R., Hashimoto, R., Friedly, J., Fu, R., Dana, T., Sullivan, S., Bougatsos, C. and Jarvik, J. 2015 Pain management injection therapies for low back pain
Coomer, R.P., McKane, S.A., Smith, N. and Vandeweerd, J.M.E. 2012 A controlled study evaluating a novel surgical treatment for kissing spines in standing sedated horses. Veterinary Surgery, 41, 890-897

Andy Fiske-Jackson, BVSc, MVetMed, Dipl. ECVS, FHEA, MRCVS, is a senior lecturer, RCVS and European Specialist in Equine Surgery specialising in the use of objective gait analysis in both lameness and back pain cases. He has lectured both nationally and internationally on the subject.

More from this author