Internal neurolysis of the maxillary branch of the trigeminal nerve for the treatment of equine trigeminal mediated headshaking syndrome

Abstract

A 5-year-old Hannovarian warmblood gelding was presented for recurrent headshaking exacerbated with exercise. The horse displayed clinical signs of repetitive vertical head movements, face rubbing on the forelimbs and on the ground, repetitive sneezing, and striking the muzzle with his forelimbs. The clinical signs resulted in a horse that could not be ridden and was dangerous. Clinical signs were most persistent in direct sunlight, but occurred with excitement, exercise, or bridling indoors. A diagnosis of equine trigeminal mediated headshaking syndrome was made. Surgical treatment was performed with a supraorbital approach to the maxillary branch of the trigeminal nerve as it exits the round foramen, where an internal neurolysis (nerve combing) was conducted on both the left and right nerves. Severe headshaking behavior resolved after surgery. The horse displayed face rubbing of the muzzle which began 96 hours after surgery and resolved over 12 days with corticosteroid and vitamin E therapy. The horse became pasture sound and the clinical signs had resolved in the presence of sunlight, but repetitive vertical head movements persisted under saddle which left the horse unpleasant to ride.

Headshaking syndrome in horses is a condition in which the horse displays several repetitive movements or clinical signs that generally include vertical (up and down) (89%) or less commonly, horizontal (side to side) head movements, face rubbing, or rubbing of the muzzle on the ground or forelimbs (75%), excessive snorting or sneezing (64%), striking its muzzle with the forelimbs or hind limbs, curling of the upper lip (Flehmen response), and an anxious expression or demeanor (61%) (1–5). Headshaking syndrome is most commonly seen in geldings over 9 y of age (range: 1 to 30 y, mean: 9 y) (2). Mills et al (6) showed a seasonal incidence of headshaking among 254 horses, with 64% of the horses showing a seasonal predilection for the months with the longest daylight. The 36% of the group which displayed non-seasonal headshaking, unlike the seasonal group, did not change the frequency of headshaking clinical signs based on sunlight or rain conditions, when indoors or at night (6). In another study, 59% of cases were found to be seasonally associated (spring to fall) with a sex predominance for geldings and breed predilection for thoroughbreds (7). Bright or sunny days were associated with headshaking in 18 of 31 (58%) of the cases in a study by Madigan and Bell (2). The headshaking clinical signs occurred in 10 of 31 (32%) of cases at exercise only (2).

Equine headshaking syndrome is thought to be analogous to a condition in humans called trigeminal neuralgia, a rare condition affecting 0.01% of the population and with a predilection for women. Clinical signs can be elicited with talking, eating, shaving, or exposure to direct sunlight, or having the wind blow against the face. It may present as a burning or throbbing pain or a lancinating and electrical pain in the face (8–10).

The pathophysiology of equine trigeminal mediated headshaking is not well-understood and several pathways or pathologies have been suggested including involvement of the trigeminal nerve (cranial nerve V) with altered transmission patterns (2,11,12), as well as decreased nerve threshold activation potential, suggesting trigeminal nerve hyperexcitability compared to a group of normal horses (13,14). There is a sex-predilection for geldings that has lead to gonadotropin-releasing hormone (GnRH) being investigated as a primary or supporting cause for trigeminal neuropathy. This theory was examined in horses via challenge with GnRH vaccination; however, this caused no significant improvement in horses diagnosed with headshaking syndrome (12). Grossly, the trigeminal nerve has been examined and found to be normal or having only mild changes such as mild lymphocytic infiltrates at the trigeminal root (13,14). Latent equine herpes virus 1 (EHV-1) infection within the trigeminal nerve was examined for correlation with headshaking and no support for the theory of postherpetic pain was found (15). Surgery of the paranasal sinuses and intraoperative trauma to the infraorbital nerve or canal have also been associated with the development of trigeminal neuritis and headshaking (16).

A broad list of differential diagnoses has been proposed for cases of headshaking in horses (5,17–23). The final diagnosis of trigeminal mediated headshaking syndrome is a diagnosis of exclusion in many cases. Diagnosis relies on thorough examinations, including general physical, ocular, oral, dental, and lameness examinations; upper respiratory endoscopy including examination of the guttural pouches; skull radiographs; observation of the horse in a darkened environment as well as under full exposure to sunlight or a bright environment (8); advanced imaging such as computed tomography (CT), magnetic resonance imaging (MRI), or nuclear scintigraphy; positive contrast paranasal sinusography (24) and nerve blocks of the intraorbital, mental, mandibular, and/or maxillary nerves. Infraorbital nerve blocks have not been found to be specific, with improvement in only 14% (5) and 18% (25) of cases. Mair et al (25) found that infraorbital nerve block resulted in worsening of the headshaking behavior in 8/19 horses. Maxillary nerve block via a ventral zygomatic arch approach showed significant improvement in 11/17 (4) and 23/27 (26) horses. Accuracy of maxillary nerve block and risks associated with the nerve block in relation to the vital structures in close proximity were evaluated with 80% of experienced clinicians having successful outcome compared to 40% of inexperienced clinicians (27). Risks of periorbital injection with resultant bulging of the eye and possible prolapse of the globe as well as risk of vascular puncture and subsequent marked retrobulbar hematoma have been described (28). An ultrasound-guided technique has been developed to improve accuracy and minimize the risk of complications from the nerve block (28).

Treatment options available for horses with trigeminal mediated headshaking syndrome include medical management with cyproheptadine alone [0.3 mg/kg body weight (BW), PO, q8h], which showed moderate improvement in 70% of horses (2); a combination of cyproheptadine (0.2 to 0.5 mg/kg BW, q8h), and carbamazepine (4 mg/kg BW, q6h), which showed improvement of 80% to 100% in 80% of cases in the study (4); and nose nets or masks, for which showed 75% of owners reported improvement (6). Corticosteroids, mast cell stabilizers, non-steroidal anti-inflammatory drugs (NSAIDs), and antihistamines have all been used without success (3,29,30). Herbal supplements had no effect on headshaking in 1 study (31) and 2 horses benefited from permanent tracheostomy in another study (4). Surgical therapy options include infraorbital neurectomy with cryotherapy which resulted in 3/19 horses with long-term ablation of the clinical signs (25); chemical sclerosis of the infraorbital nerve within the rostral aspect of the infraorbital canal with phenol which resulted in 5/5 horses showing remission of clinical signs for 6 wk to 9 mo (4); caudal compression of the infraorbital nerve within the infraorbital canal with stacked platinum embolization coils which resulted in an initial successful outcome in 35/57 (63%) horses, but at 18-month follow-up success rate was 49% including horses that required more than 1 procedure or revision (26,32). Percutaneous electrical nerve stimulation (PENS) therapy has also been used in standing sedated horses with impulses delivered over the infraorbital nerve at 2 and 100 Hz alternating every 3 s for 25 min. Response to treatment was positive in 6/7 horses; however, repeated treatments were needed to maintain remission (33).

We postulated that a novel surgical approach to the trigeminal nerve at the level of the established maxillary nerve block location deep to the zygomatic process immediately cranial to the round foramen would allow for access to the most proximal portion of the trigeminal nerve without entering the calvarium. This would allow for ablation of the nerve’s sensory inputs from the face, muzzle, nose, sinuses, and maxillary tooth roots. This case report describes a novel approach to the maxillary branch (CN V2) of the trigeminal nerve at the level of the round foramen deep to the zygomatic process via a supraorbital approach to perform an internal neurolysis (nerve combing) and chemical ablation of the trigeminal nerves bilaterally as a salvage procedure to relieve the clinical signs of headshaking.

Case description

A 5-year-old Hannovarian gelding was presented with a 2-week history of acute onset of headshaking while being exercised, including movement of the head up and down, rubbing the face on the forelimbs, striking at the face with forelimbs, rubbing the muzzle on the ground when lunged, anxious demeanor, and periodic sneezing fits. The headshaking seemed to increase with intensity of exercise but was present even when bridling or at rest and the horse would strike dangerously at its own head and become very difficult to control. Sunlight also intensified the episodes of headshaking.

On examination, the horse had abrasions of the nose and lower lip from rubbing on the ground and inanimate surfaces. The clinical examination was otherwise unremarkable. The ophthalmic examination showed normal vision, pupillary light responses, menace response, anterior and posterior chambers. The horse was lunged in both left and right circles at a walk and trot. As the horse begun to lunge, he became agitated and shook his head in a highly repetitive vertical motion followed by repetitive face rubbing on the ground and his forelimbs. He could not be cantered as he became uncontrollable due to his repetitive headshaking. Taking him into sunlight further exacerbated the headshaking and he became uncontrollable. He could not be turned out in the sunlight and had to remain in the stable. Based on these findings, he displayed Grade 5/5 headshaking (4) (Table 1).

Table 1.

Clinical grading scale for severity of headshaking syndrome in horses (4).

Grade	Definition
1	Intermittent and mild clinical signs; facial muscle twitching; rideable
2	Moderate clinical signs; definable conditions under which signs occur/develop; rideable with some difficulty
3	Rideable, but unpleasant ride and difficult to control
4	Unrideable; uncontrollable
5	Dangerous, with bizarre behavior patterns

A dental examination revealed no abnormalities of the teeth, tongue, or oral cavity. Wolf teeth had been previously removed several years earlier and no remnants were seen. Upper airway endoscopy revealed no significant anatomical or functional abnormalities at rest. Radiographs of the skull showed no abnormalities of the tooth roots, sinuses, temporohyoid joint/stylohyoid bone, poll, or calvarium. No fractures or bony changes were detected. No lameness was recognized at a walk or trot in the straight line or in either circle on a lunge line.

Bilateral infraorbital nerve blocks, mental nerve blocks, and maxillary nerve blocks were carried out as previously described (4,25,26). The mental and infraorbital nerve blocks did not provide relief of the clinical signs; however, maxillary nerve blocks significantly improved the clinical signs and allowed the horse to be exercised with minimal headshaking.

Our diagnosis was equine trigeminal mediated headshaking syndrome. Treatment was as follows. The horse was initially confined to a darkened stall and a nose mask and fly mask, which shaded the eyes, was worn daily for 6 wk. There was no outdoor turnout and hand walking in the arena occurred only if the horse was controllable. The horse remained difficult to control and could not be ridden, with progressively bizarre behaviors such as running at the stall door and walls in the presence of light over the course of 6 wk of conservative management. Medical and surgical management options were discussed with the owner. Although conservative medical management with medication was advocated, the owner was not interested in life-long medical therapy. The horse did not improve until nerve blocks at the level of the maxillary nerve were performed. The current available surgical interventions would not have addressed pathology of the nerve at the maxillary nerve level; therefore, a novel procedure was proposed with consideration for the ethical and humane management of the horse in his current state. Euthanasia was considered.

The horse was admitted to hospital for surgery and anesthetized with xylazine (Rompun; Bayer Animal Health, Mississauga, Ontario), 0.5 mg/kg BW, IV, ketamine (Ketaset; Zoetis, Kalamazoo, Michigan, USA), 2.2 mg/kg BW, IV, butorphanol (Torbugesic; Mérial Canada, Baie d’Urfé, Québec), 0.02 mg/kg BW, IV, and guaifenesin 5% in lactated Ringers solution, IV, and maintained on isoflurane inhalation anesthesia. The horse was positioned on the surgical table in left lateral recumbency and the head was elevated with a positioning wedge pad to tilt the forehead upward at a 45° angle from the table surface. This allowed unrestricted access to both supraorbital spaces. The forehead from the level of the ears caudally to the level of the medial canthus cranially and abaxially to the level of the lateral canthus was clipped and aseptically prepared for surgery. A 5-cm semicircular incision along the axial margin of the supraorbital space was made with a #10 scalpel blade, beginning at the abaxial cranial margin of the zygomatic process and continuing axially along the margin of the supraorbital fossa to finish at the caudal abaxial margin of the zygomatic process. The skin and subcutaneous tissue was reflected laterally to allow access to the supraorbital fat pad. The supraorbital fat was resected and removed with electrocautery and blunt dissection until the supraorbital space was clear of adipose tissue. An operating rigid endoscope (Arthrex, Naples Florida, USA) was placed into the fossa and into the retrobulbar space to provide illumination and visualize the neurovascular bundles emerging from the round foramen ventral to the orbital fissure. The pterygoid crest of the sphenoid bone can be digitally palpated and the location of the maxillary nerve is ventral and slightly cranial to this crest. The ethmoidal foramen will be first encountered at the dorsal aspect of the pterygoid crest then the optic canal will be within the sheath of the periorbita and rectus muscles of the globe. A blunt probe and small retractor were used to dissect the maxillary branch of the trigeminal nerve from the surrounding nervous and vascular structures. The maxillary branch was identified as being continuous with the infraorbital nerve rostrally as it exited the maxillary foramen abaxial to the globe (Figure 1) and continued to the alar foramen caudally. Care was taken to isolate the maxillary nerve and avoid the maxillary artery, external ophthalmic artery, rostral deep temporal artery and caudal deep temporal artery which were all within the field of view (Figure 2). With the maxillary nerve isolated, a #11 scalpel was used to incise the nerve longitudinally over a 1-cm section (nerve combing) immediately cranial to the nerve exiting the round foramen in order to perform the internal neurolysis procedure. A chemical ablation of the nerve with 0.5 mL 70% ethanol was performed at the site of the internal neurolysis. The site was lavaged with sterile saline containing 10 million IU sodium penicillin (Pharmaceutical Parternship, Richmond Hill, Ontario) and checked to ensure hemostasis was maintained. The retrobulbar space was left and the deep subcutaneous tissue was closed with #2-0 Monocryl (Ethicon; Johnson & Johnson, Markham, Ontario) in a simple continuous pattern and the subcuticular tissue was closed with #2-0 Monocryl (Ethicon; Johnson & Johnson) in a simple continuous pattern. The skin was closed with skin staples and the procedure was repeated with the other maxillary nerve. No hemorrhage was found at either site.

Figure 1.

Bony topography of the pterygopalatine fossa. Open yellow arrow — surgical approach; Black arrow — optic foramen; Green arrow — Alar or Round foramen where cranial nerve (CN) 5 exits calvarium; Orange arrow — opening of the palatine canal; Blue arrow — combined opening of maxillary and sphenopalatine foramina where CN5 enters the infraorbital canal. Red hashed line — location of the maxillary branch (V2) of CN5.

Figure 2.

Identification of the maxillary branch (V2) of the trigeminal nerve (CN5) with a probe (black arrow). Cranial is to the left. A — Approach via supraorbital foramen looking ventrally (surgical approach). B — Course of CN5 with zygomatic process and tissues lateral to the nerve removed for reference.

The horse recovered uneventfully from anesthesia and was returned to a darkened stall. The horse was administered phenylbutazone (Phenylbutazone 20%; Rafter 8 Products, Calgary, Alberta), 2.2 mg/kg BW, IV, q24h for 5 d, gentamicin sulfate (Gentamax; Phoenix, St. Joseph, Missouri, USA), 6.6 mg/kg BW, IV, q24h for 3 d, Ceftiofur (Excenel; Zoetis), 2.2 mg/kg BW, IV, q24h for 3 d, and omeprazole (Gastrogard; Mérial), 4 mg/kg BW, PO, q24 h for 14 d.

At 96 h post-surgery, the horse began displaying face rubbing behavior and proprioceptive deficits with the muzzle in his inability to locate his feed bucket without bumping into it before prehending and eating his feed normally. This progressed to inappetence and apprehension when eating which lasted for 4 d before resolution. The face rubbing resolved within 12 d. Treatment for the face rubbing behavior consisted of vitamin E (WN Pharmaceuticals, Coquitlam, British Columbia), 12 IU/kg BW, PO, q24h for 30 d and dexamethasone (Dominion Veterinary Labs, Winnipeg, Manitoba), 0.05 mg/kg BW, IV, q24h for 14 d then 0.025 mg/kg BW, IV, q24h for 7 d.

With the resolution of the face rubbing behavior, the horse became calmer and the headshaking was not present during the first 60 d after surgery. The horse was maintained indoors for the first 30 d after which he was given 12 h outdoor turnout daily. Around 60 d, the horse displayed intermittent headshaking and a mild intermittent facial twitch on the left side of the face. When exercised, he displayed mild headshaking and was rideable with some difficulty. He was graded as a Grade 1 to 2 headshaking on Newton’s scale (4). He was able to tolerate sunlight without violent headshaking; however, he retained a loss of proprioception with his muzzle and frequently caused abrasions to the muzzle, upper lip, and nose by bumping into his feed pail/manger and stall door when awaiting food. The owner also noted that he seemed to have lost his normal flight zone around his nose and would frequently bump into the handler with his nose. At 6 mo after surgery, the clinical signs had stabilized and were the same as at the 60-day post-surgery follow-up with a Grade 1 to 2 headshaking syndrome (4); however, he was not able to be ridden as a performance horse in the hunter/jumper arena. The horse was euthanized at 6 mo after surgery due to financial constraints and the head was sent for postmortem evaluation. The pathologist reported no anatomical variations of the trigeminal nerves or the brain. Histological findings consisted of 2 tiny foci of gliosis immediately adjacent to a lateral ventricle, a small collection of lymphocytes in 1 small portion of the choroid plexus and 2 capillaries with swollen endothelial cells in different locations. No histological or gross pathological cause for the headshaking could be found.

Discussion

Idiopathic headshaking syndrome or equine trigeminal mediated headshaking syndrome is a poorly understood condition of the horse postulated to be due to inflammation or irregular nerve conduction of the trigeminal nerve(s). The horse in this report displayed classical signs of headshaking syndrome and was severely affected. This horse was younger than the mean age of onset of 9 y and the onset of the condition was reportedly acute. No trauma or other known cause triggered the onset of clinical signs and no gross anatomical or physiological abnormalities were found on the clinical examination or diagnostics to identify a cause for the onset of clinical signs.

Diagnostic nerve blocks have been cited in the literature as being useful in identifying the trigeminal nerve as the source of headshaking, which was the case with this horse as well. Infraorbital and mental nerve blocks did not improve the headshaking activity; however, anesthesia of both maxillary nerves via a ventral zygomatic process approach (28) ameliorated the headshaking. Medical treatment options have shown promise in remission of clinical signs in some horses but require life-long therapy with psychoactive medications such as carbemazepine and are associated with variable remission periods (3–5). In this case, the owner elected to forego medical management as the horse was desired to perform and restrictions do not allow the use of these medication in competing horses. Use of a nose mask resulted in increased display of the headshaking clinical signs. A full fly mask which darkened the eyes but did not cross the level of the infraorbital foramen in combination with a darkened stall was successful in lessening the clinical signs; however, the horse became dangerous again once exposed to sunlight, bridled, or agitated (presumably increases in blood pressure).

Surgical management of trigeminal mediated headshaking includes infraorbital neurectomy, chemical sclerosis, or cryotherapy of the infraorbital nerve within the canal or caudal compression/facilitated necrosis of the infraorbital nerve within the infraorbital canal and, most recently, percutaneous electrical nerve stimulation (PENS) (4,25–26,32–33). While phenol chemical sclerosis of the infraorbital nerve within the canal produced the highest success rate with 5/5 horses showing remission of clinical signs for 6 wk to 9 mo, repeated therapy would be required to maintain this remission and nerve damage may occur which could result in deafferentation nerve pain, although this was not reported (4). Caudal compression of the infraorbital nerve within the infraorbital canal with platinum embolectomy coils results in facilitated necrosis of the infraorbital nerve immediately rostral to the maxillary foramen within the infraorbital canal. During the study, several of the coils migrated and needed to be replaced and serious self-traumatizing face rubbing occurred in 63% of the horses after surgery as a significant complication resulting in 4 horses being euthanized (26). Infraorbital neurectomy results in a high complication rate with 85% of horses in the studies published having serious face rubbing lasting 3 to 8 wk and only 15% of horses achieving remission of clinical signs (25). Percutaneous electrical nerve stimuation appears to be a viable option for treatment, with 86% of horses achieving remission; however, repeated therapy at multiple intervals is required to maintain remission (33). Unlike medical management, PENS does not represent a competition medication exclusion (33).

In humans, the source of trigeminal neuropathy/neuritis/neuralgia is often vascular compression by the superior cerebellar artery against the trigeminal nerve root as it exits the brain stem. It is postulated that changes in blood pressure and compression of the vessel against the nerve results in neuritis or abnormal nerve transmission which causes the pain associated with trigeminal neuralgia in humans. Trigeminal neuritis tends to affect women more than men and occurs most commonly in patients 50 to 60 y old. It is also often unilateral with the right side of the face more commonly affected than the left (8–10). Although no direct comparison can be made in horses and other animals due to the lack of vascular brain base contrast studies, it stands to reason that given the blocking pattern in the horse, the source of the neuritis may not be in the infraorbital canal and may be located caudal to the round foramen at the level of the trigeminal ganglion or as in humans, at the level of the trigeminal nerve root as it exits the pons and before it enters Meckel’s cave. In humans, several surgical options are available such as microvascular decompression, percutaneous glycerol rhizotomy, percutaneous radiofrequency ablation, trigeminal balloon microcompression, stereotactic radiosurgery, and peripheral neurectomy (10). Mircovascular decompression has the highest success rate and directly addresses the pathology. The procedure involves a brain base approach and identification of the superior cerebellar artery where it is in intimate contact with the trigeminal nerve root. The artery is dissected from the nerve root and a Teflon pledget is place between the artery and nerve to insulate the nerve from the artery (10).

In the horse, this approach is complicated by the potential approaches to the brain base and the limited access to the brain base due to the location of the vertical ramus of the mandible, temporomandibular joint, guttural pouches, and parotid gland (34). The most proximal portion of the trigeminal nerve that can be accessed in the horse is at the site of the maxillary nerve immediately rostral to the nerve exiting the round foramen in the retrobulbar space. This is the location at which the maxillary nerve block is performed in the horse (28,35). Based on this anatomical and clinical information, an approach to the trigeminal nerve at this location would seem to be a logical location to surgically interrupt the nerve transmission thought to be associated with trigeminal mediated headshaking in the horse. In this case, a supraorbital approach to the maxillary branch of the trigeminal nerve (CN V2) was used. This approach avoids the large vessels present in a lateral approach, such as the maxillary artery and vein and the transverse artery and vein which lie directly over the location of the nerve under the zygomatic process (28,34). Using the supraorbital approach, the supraorbital and retrobulbar adipose tissue is removed and minimal vasculature and nervous tissue is encountered or disrupted with careful dissection. The pterygoid crest can be palpated digitally or visualized with a rigid endoscope and the neurovascular bundle emerging from the alar and round foramen can be seen and palpated. Great care must be taken to avoid damaging the neurovascular bundles including the optic nerve and associated vasculature within the periorbital tissue; however, the trigeminal nerve (maxillary branch of the trigeminal nerve) can be isolated from the surrounding vascular and nervous tissues and elevated with a probe to expose approximately 1 cm of nerve for neurolysis. The purpose of internal neurolysis is to not transect the nerve but rather incise and separate the nerve bundle along the long axis of the nerve which will decrease conduction intensity and speed and result in significant slowing of nerve transmission and hence nerve associated pain. The nerve bundle will scar and heal with a larger cross-sectional area which will maintain the decrease in nerve transmission and provide relief of the sensory associated nerve pain into the brainstem (9). Furthermore, we also performed a local chemical sclerosis of the trigeminal nerve at the most proximal aspect of the isolated nerve bundle with ethanol. This was done to decrease nerve sensory input by disruption of the axon and axon membrane via dehydration of the tissue with the ethanol (10). The nerve blocks required both left and right maxillary nerves to be anesthetized before the clinical signs ceased. For this reason, we elected to perform the procedure bilaterally.

The main complication after surgery in this case was face rubbing which began at 4 d after surgery and lasted approximately 12 d before resolution. This is similar, but shorter in duration, to the deafferent nerve pain response seen in previous studies (25–26,32). The face rubbing behavior was managed with vitamin E supplementation and dexamethasone corticosteroid therapy over 14 d, but this may have been self-limiting regardless of therapy. The horse also developed inappetence which lasted about 4 d, during which time the horse was reluctant to use the upper lip to prehend feed and seemed to have an adverse sensory reaction to contact of the muzzle with his feed and/or his feed or water pail. He ate his feed and drank water, but had an anxious expression and a mild facial twitch developed when doing so. We postulate that there was some deafferent nerve pain during this time which may have been associated with inflammation at the nerve surgery site. This resolved over 4 d and he resumed normal feed and water intake. In addition, between 45 and 60 d, the horse began to develop mild headshaking clinical signs again as well as a mild facial twitch on the left side of the face/muzzle. He seemed to also lack proprioception with his muzzle, nose, and upper lip. He bumped into inanimate objects with his nose and seemed to prehend his feed with his incisors rather than his upper lip before taking feed into his mouth. This is likely due to the lack of or delay in sensory transmission from the trigeminal nerve, as would be expected. His vision was assessed and found to be normal with normal pupillary light responses and menace responses in both eyes. The surgical sites healed without complication and no significant abnormalities of the supraorbital space were detected.

There was significant improvement in the clinical outcomes from a Newton Grade 5 to Grade 1 to 2 after surgery. The horse was rideable with difficulty due to the headshaking that persisted at 6 mo after surgery; however, the horse was markedly improved from the dangerous state he had presented and was now comfortable outdoors in sunlight as well as when being handled. Although this only represents a single case, we believe this surgical approach could be considered as a salvage procedure in horses with severe trigeminal mediated headshaking syndrome in which the clinical signs are best alleviated with a maxillary nerve block.