Veterinary Record 168:19 doi:10.1136/vr.c5992
  • Research
  • Paper

Use of a gonadotrophin-releasing hormone vaccine in headshaking horses

  1. J. E. Madigan, DVM, MA, DipACVIM1
  1. Veterinary Teaching Hospital, College of Sciences, Massey University, Palmerston North, New Zealand
  2. School of Veterinary Medicine, University of California, Davis, One Shields Avenue, Davis, CA 95616, USA
  1. E-mail for correspondence kirstiepickles{at}
  • Dr Pickles' and Dr Madigan's present address is School of Veterinary Medicine, University of California, Davis, One Shields Avenue, Davis, CA 95616, USA

The aim of this study was to investigate the use of a gonadotrophin-releasing hormone (GnRH) vaccine in the treatment of headshaking in horses. Fifteen geldings received two doses of the GnRH vaccine four weeks apart. Serum was collected before and after vaccination to measure concentrations of luteinising hormone (LH) (10 horses) and follicle-stimulating hormone (FSH) (six horses). Owners recorded the frequency of seven common headshaking behaviours using a visual analogue scale (VAS) before vaccination and at two, four, eight, 12, 16 and 20 weeks after vaccination. Serum LH (P=0.008) and FSH (P=0.03) concentrations decreased significantly following vaccination. Although approximately one-third of the owners reported a subjective improvement in headshaking, serial scoring did not indicate a reduction in headshaking behaviours following vaccination with a commercial GnRH vaccine. Vaccination reactions were observed in four of 15 horses (27 per cent), including one case of severe, presumed immune-mediated, myositis.

HEADSHAKING syndrome has been described for over 100 years and can affect horses of all breeds at any age (Williams 1899, Lane and Mair 1987, Mayhew 1992, Madigan and others 1995, Newton and others 2000, Madigan and Bell 2001, Mills and others 2002). Typical headshaking behaviours include spontaneous, violent, uncontrollable, vertical flicks of the head, excessive snorting of the nose, trying to rub the nose on the foreleg or ground and striking at the nose with a forelimb (Newton and others 2000, Madigan and Bell 2001, Mills and others 2002). The condition is usually worse at exercise (Lane and Mair 1987). There are a few treatment options for horses with headshaking, including nose nets, which have been reported to provide up to 70 per cent relief in 25 per cent of horses (Mills and Taylor 2003), and caudal compression of the infraorbital nerve with platinum embolisation coils, reported to result in 59 per cent success following one surgery (Roberts and others 2009).

Trigeminal neuralgia is regarded as the most likely explanation for the clinical signs observed in headshaking horses (Newton and others 2000, Madigan and Bell 2001). Indeed, human sufferers of trigeminal neuropathic pain describe burning, tingling, itching and electric shock-like sensations (Nurmikko and Eldridge 2001), which appear to fit well with the behaviours displayed by headshaking horses. Similar to trigeminal neuralgia syndromes in other species, studies of the trigeminal nerve, brain and head of horses with headshaking have revealed no pathology in the vast majority of cases (Newton 2001).

Recent studies have determined that headshaking is often seasonal, beginning in the spring and ceasing in the autumn in the majority of horses, and that geldings are over-represented (Lane and Mair 1987, Madigan and Bell 2001, Mills and others 2002). Geldings, like stallions, have a centrally mediated vernal increase in gonadotrophins; however, geldings lack gonadal testosterone and inhibin to provide negative feedback on gonadotrophin production, and indeed, higher seasonal concentrations of luteinising hormone (LH) and follicle-stimulating hormone (FSH) have been reported in geldings compared with stallions (Thompson and others 1986). Furthermore, the authors have also seen abnormally cycling mares (with small, inactive ovaries and anoestrus), showing headshaking, which have responded to manipulation of the reproductive cycle. A seasonal, gonadotrophin-mediated alteration in trigeminal nerve neurochemistry is therefore postulated as a pathophysiological cause of headshaking. Although LH receptors have not been identified in neural tissue, expression of LH receptors in non-reproductive tissue has been demonstrated in the ferret adrenal gland, and furthermore, these receptors appear functional such that administration of exogenous LH results in the release of sex steroids (Schoemaker and others 2002).

The authors hypothesise that prolonged elevated levels of gonadotrophins in geldings cause instability of the trigeminal ganglion and resultant neuropathic pain from inappropriate signals associated with various stimuli such as exercise, nasal air flow, light, sound, feed, smell and excitement. To test this hypothesis, a study was performed using a commercial gonadotrophin-releasing hormone (GnRH) vaccine with the aim of decreasing the serum gonadotrophin concentrations in headshaking horses.

Materials and methods

Clinical and diagnostic examination

Nineteen horses (12 in New Zealand and seven in California, USA) reported by their owners to be showing signs of headshaking were evaluated for inclusion in the study following completion of a questionnaire by the owner and assessment of live or video-recorded headshaking behaviour by the authors (KJP, JEM, JB), all of whom possess many years of experience in diagnosing headshaking. Horses were recruited prospectively into the study by advertising for cases in local veterinary publications and by word of mouth. The inclusion criteria of the study were based on a historical description and demonstration of characteristic signs of headshaking, as previously reported by Madigan and Bell (2001), such as violent, vertical head flicks, nasal rubbing and snorting, and an unremarkable physical and neurological examination.

This trial was approved by the Animal Ethics Committees at Massey University and the University of California, Davis.

Vaccination and blood sampling

The GnRH vaccine (Equity; Pfizer VMRD) used in the study is licensed in Australia for chemical castration and oestrus suppression of horses, and was imported under special licence into New Zealand and the USA for use in this trial. Elhay and others (2007) have previously reported this vaccine to contain a GnRH peptide conjugated to a protein carrier and admixed with an immunostimulating complex as adjuvant. The immunostimulating complexes of this vaccine are formed from Saponin Quil A (Brenntag Biosector), cholesterol (Avanti Polar Lipids) and dipalmitoylphosphatidylcholine (Avanti Polar Lipids). Each 1 ml dose contained 200 µg peptide conjugate, 300 µg immunostimulating complexes and 0.01 per cent thiomersal as preservative, and isotonic buffered solution to volume. Horses were vaccinated during the spring by intramuscular administration of two doses of GnRH vaccine, four weeks apart, as recommended by the manufacturer.

Blood was collected by jugular venepuncture for measurement of serum LH (Endolab, Christchurch Hospital, New Zealand; Department of Animal Science, University of California, Davis) and FSH concentrations (Department of Animal Science, University of California, Davis) before vaccination (week 0) and four weeks after the second vaccine dose (week 8).

Visual analogue scale recording of headshaking

Seven commonly reported headshaking behaviours were chosen for serial assessment of headshaking behaviour: vertical head flicking, snorting of the nose, dropping the nose to the ground, rubbing the nose on objects, rubbing the nose on the foreleg, striking at the nose, and flipping the nose or top lip (Madigan and Bell 2001, Mills and others 2002).

The frequency of these behaviours was documented on visual analogue scale (VAS) recording sheets (Fig 1) by the owners at weeks 0, 2, 4, 8, 12, 16 and 20. Each behaviour was depicted by a continuous 10 cm line (the VAS) on which the observer could mark anywhere along the line to indicate the frequency and severity of the behaviour, as appropriate. The descriptors ‘never’ and ‘continually’ marked the extremities of the line, ‘occasionally’ was marked above the second quartile and ‘frequently’ was marked above the third quartile. The scores for each behaviour were obtained by measuring from the start of the scale to the point marked on the scale by the owner in millimetres. To reduce variation, the owners were advised to perform each assessment under as similar as possible circumstances, for example, outdoors, early in the morning, under saddle, using the same bridle and bit for each evaluation, and with similar weather conditions if possible.

Fig 1

Visual analogue scale (VAS) recording sheet used by horse owners to evaluate seven common headshaking behaviours at intervals during the course of the study. Owners could mark anywhere along the horizontal line, which was 10 cm long in the recording sheet given to the owners, according to the frequency and severity of the headshaking behaviour

Statistical analyses

Anderson-Darling normality tests showed the data to be non-parametric. Prevaccination and postvaccination LH and FSH concentrations were compared using a Wilcoxon signed rank test. The VAS data were assessed graphically.


Horses completing study

Fifteen horses completed the study: all seven horses enrolled in the USA and eight horses enrolled in New Zealand. Four of the 12 horses that were enrolled in New Zealand were withdrawn from the study. Three of these horses were withdrawn by the authors due to lack of owner compliance in returning the VAS scoring sheets. One of these three owners indicated during a follow-up telephone call at four weeks that she was planning to have the horse euthanased due to continuing uncontrollable headshaking. The owner of the fourth withdrawn horse dropped out of the study after four weeks as she perceived no improvement in the horse's extremely violent headshaking. Serum samples and any VAS data from these horses were excluded from further analysis.

Vaccination and gonadotrophin concentrations

All 15 horses that completed the study completed the vaccination course. Serum samples were available from both weeks 0 and 8 for 10 horses for LH measurement and from six horses for FSH measurement (FSH could not be measured in the horses enrolled in New Zealand). Missing serum samples were due to lack of owner compliance or sample spoilage. The median prevaccination and postvaccination gonadotrophin concentrations were, 7.39 and 0.40 ng/ml for LH, and 17.56 and 1.95 ng/ml for FSH, respectively. Vaccination resulted in a significant decrease in LH (P=0.008) (Fig 2) and FSH (P=0.03) (Fig 3) concentrations indicating that vaccination was successful.

Fig 2

Serum luteinising hormone (LH) concentration in 10 headshaking horses before and after vaccination with a gonadotrophin-releasing hormone vaccine. Vaccination resulted in a significant reduction in LH concentration (P=0.008)

Fig 3

Serum follicle-stimulating hormone (FSH) concentration in six headshaking horses before and after vaccination with a gonadotrophin-releasing hormone vaccine. Vaccination resulted in a significant reduction in FSH concentration (P=0.03)

Three horses developed transient, localised muscle swelling and soreness, which resolved within five days, as a reaction to either the first or second vaccine dose. One horse suffered a severe, systemic inflammatory response following both the first and second vaccination doses. This horse showed pyrexia, leucocytosis and signs of depression, and suffered rapid muscle wastage, particularly over the epaxial and gluteal muscles, losing 30 kg over a two-week hospitalisation period. Despite endoscopy of the respiratory tract, transtracheal aspirate culture and cytology, gastroscopy, thoracic radiography, serial thoracic and abdominal ultrasonography, serial haematological and biochemical evaluations, including protein electrophoresis, urinalysis, faecal analysis and radiolabelled leucocyte scintigraphy, the cause of the systemic inflammation could not be determined. A muscle biopsy to investigate immune-mediated myositis was declined by the owner.

VAS and headshaking behaviour

Completed VAS sheets for all seven time points were available for seven of the 15 horses. For the other eight horses, VAS data were missing for one time point from five horses, for two time points from two horses and for three time points from one horse. Data were absent from one horse at two weeks, from two horses at 12 weeks, from four horses at 16 weeks and from five horses at 20 weeks.

The owners used a wide range of VAS scores throughout the study to describe the headshaking behaviour of their horses. Typical distributions of VAS measurements are depicted by the serial scoring for the headshaking behaviours ‘vertical head flicking’ (Fig 4) and ‘snorting the nose’ (Fig 5).

Fig 4

Visual analogue scale (VAS) data for vertical head flicking in horses before vaccination (week 0) and at intervals after vaccination with a gonadotrophin-releasing hormone vaccine. * Outlier

Fig 5

Visual analogue scale (VAS) data for snorting of the nose in horses before vaccination (week 0) and at intervals after vaccination with a gonadotrophin-releasing hormone vaccine. * Outlier

Two owners spontaneously reported an improvement in their horses' headshaking within the first two to four weeks following vaccination, which persisted for several months. A further three owners spontaneously reported a brief, temporary improvement following vaccination. However, there was no meaningful pattern in the serial VAS measurements for any headshaking behaviour.


Successful immunisation against GnRH was achieved, such that FSH and LH concentrations were significantly decreased four weeks following the second vaccine dose. The vaccine has previously been shown to induce production of antibodies to GnRH, decrease ovarian activity and serum concentrations of sex steroids and inhibit oestrous behaviour in mares (Elhay and others 2007). However, its use in geldings and its effect on gonadotrophin production have not previously been reported.

Reactions occurred following administration of the vaccine in several horses. Most of these reactions consisted of short-term, localised muscle swelling and soreness at the injection site; however, one horse had a severe systemic inflammatory response. Elhay and others (2007) reported localised injection-site inflammation in 10 of 24 (42 per cent) horses following the administration of primary or secondary doses of this vaccine. This frequency of reactions is considerably higher than that seen in the present study, although none of the reactions reported by Elhay and others (2007) were severe, and they included swellings that were palpable but not visible. Such minor reactions may have been missed by the owners in the present study. Interestingly, Elhay and others (2007) reported an injection-site reaction in one control horse that had been vaccinated with a placebo containing only adjuvant, and therefore it may have been the adjuvant and/or carrier vehicle, rather than the immunogenic GnRH peptide, that caused the local reactions.

Although some owners reported a perceived improvement in their horses' headshaking behaviour, analysis of the VAS data did not demonstrate any beneficial effect of the vaccine. The apparent improvement noticed by some owners may have been due to the placebo effect, and indeed, any benefit of vaccination should not have been evident until after the second vaccination. However, it is interesting to note that Elhay and others (2007) reported that nine of 24 (37.5 per cent) mares developed GnRH antibody titres following the first injection of this vaccine, although maximal titres were not achieved until two weeks after the second dose. Studies on mares have shown that secretion, and possibly production, of LH is highly dependent on GnRH, whereas FSH secretion is at least partially independent of GnRH stimulation (Garza and others 1986, Thompson and others 1986, Safir and others 1987). However, this does not appear to explain the lack of response of headshaking to vaccination, as the concentrations of FSH after vaccination determined in six of the 15 horses in the present study were very low and significantly reduced from the prevaccination concentrations.

The VAS system was used to score headshaking behaviour as these scales have previously been successfully used for serial monitoring of pain (Joyce and others 1975, Bijur and others 2001, To and others 2002), including trigeminal neuropathic pain (Schmidt and others 2003, Nguyen and others 2008, Lefaucheur and others 2009). Self-measurement of pain by VAS scoring has been shown to be a robust, sensitive and reproducible method of expressing pain severity (Joyce and others 1975, Langley and Sheppeard 1985), which correlates well with other methods of measuring pain (Bijur and others 2001). Additionally, VAS scales can be completed quickly, and thus are usually well accepted by people (Langley and Sheppeard 1985). However, for VAS scoring to be accurate, it is of paramount importance that users understand and know how to use the scales, and it is possible that the horse owners did not understand the system sufficiently and so returned inaccurate data. This may be particularly true if the owners did not follow the recommendation to carry out all serial evaluations under the same conditions. Owners were instructed on how to use the VAS forms, but it was not possible for the authors to be present at each scoring due to the widespread geographical distribution of the horses. It is also assumed that the horses' behaviour, as observed by the owners and marked on the VAS, was representative of the pain experienced by the horse, which may not be valid. Guidelines recommend the use of just one descriptor at each extremity of the 10 cm line, for example, ‘no pain’ and ‘worst pain imaginable’ (Joyce and others 1975), and thus the VAS used in the present study may have confused the owners by using too many descriptors and may have been used more as an interval fixed-point scale, encouraging clustering around the quartile descriptors (Scott and Huskisson 1976). The use of four descriptors may also have led the owners to assume that the VAS provides a linear measure of a behaviour rather than a logarithmic scale (Langley and Sheppeard 1985).

The small number of horses that completed the present study is likely to have hindered the ability to detect any positive effect of vaccination on headshaking behaviour. The initial number of horses recruited to the study was limited by the amount of vaccine the authors were able to import, and the study population was further reduced by horses being withdrawn from the study as a result of poor owner compliance and/or the owners' perception that vaccination had no effect. The lack of a control population receiving a placebo treatment is acknowledged as a severe limitation of the study; however, due to the distressing nature of headshaking, owners were willing to be recruited into the study only if their horses were to receive the vaccine. Absent VAS data, particularly at later time points, may also have influenced the ability to detect an effect of vaccination.

In conclusion, although some of the owners perceived a subjective improvement in their horses' headshaking behaviour, repeated VAS analysis did not demonstrate a beneficial effect of GnRH vaccination.


The authors thank Lisa Boden for advice on statistical analyses.


  • Provenance not commissioned; externally peer reviewed


Sign up for job alerts

Vet Record Careers features a comprehensive database of vet jobs.
Browse for your next job, or set personalised job alerts so you hear about vacancies as soon as they go online.
Find out more at Vet Record Careers

Free Sample Issue

This recent issue is free to all users to allow everyone the opportunity to see the full scope and typical content of the Veterinary Record.
View free sample issue >>