Clinical manifestations

Presentation and course

Hemifacial spasm is characterized by intermittent twitching of the muscles supplied by a single facial nerve, which differentiates this movement disorder from blepharospasm, typically a bilateral disorder, and other facial dyskinesias (76; 156; 97). Although the disorder is well-known to neurologists, the diagnosis may be delayed by several years. In fact, in a survey of 866 patients with dystonia or hemifacial spasm treated in 14 movement disorders centers in Canada, the mean lag time between symptom onset and diagnosis was 5.4 years (79).

Onset is generally between the second and eighth decades of life and averages between 45 and 51 years of age (41). About 7% of cases present at or before the age of 30 (136). Onset during infancy is atypical and usually indicates an underlying tumor (45; 17). Women account for about 60% of cases. Hemifacial spasm has been found to be more common in the Asian population than in other populations (154). Although familial cases have been reported (100), most are sporadic. In one series of 10 patients in five families and review of 13 historical familial cases, the likely pattern of inheritance was autosomal dominant with incomplete penetrance. No clinical differences were seen in familial cases compared to sporadic cases, including response to microvascular decompression, suggesting that inherited cases share the same relationship with vascular compression as sporadic cases (102). A mother and son were described to have young-onset familial hemifacial spasm secondary to a neurovascular contact between their right facial nerve and a branch of their right posterior inferior cerebellar artery (114).

Hemifacial spasm may be a sequela of facial nerve damage or reflect a structural lesion of the nerve or its surroundings (secondary hemifacial spasm). Approximately 5% of cases have a prior history of Bell palsy and a similar number have comorbid trigeminal neuralgia (146).

Primary hemifacial spasm begins in the periorbital region, usually the lower lid, in 60% to 90% of cases. Those with secondary hemifacial spasm often notice the spasms in perioral and periorbital muscles at onset (31). Wherever the onset, the movements gradually spread to involve other muscles supplied by the ipsilateral facial nerve. When multiple facial muscles are involved, the spasms are synchronous in all ipsilateral muscles (146). Some patients have spread to the contralateral side. In one report of 10 bilateral cases, the latency for contralateral development of spasms averaged 33.3 months. In his 1905 paper, Babinski described paradoxical synkinesis in hemifacial spasm, known as “the other Babinski sign”: “When orbicularis oculi contracts and the eye closes, the internal part of the frontalis contracts at the same time, the eyebrow rises during eye occlusion” (37). This “other Babinski sign” is typically present in hemifacial spasm but not in essential blepharospasm, and although it was found to have only 25.3% sensitivity, it differentiates hemifacial spasm from blepharospasm with 100% specificity (133). The disability that results from hemifacial spasm relates to multiple areas of function. More than 75% of patients are embarrassed by the spasms, and 15% to 65% complain of depressive symptoms (06). The spasms affect vision in about 60% of patients and interfere with work in about 36% (06).

The facial twitches in patients with hemifacial spasm may be clonic or tonic, and it is common for episodes to begin as multiple clonic spasms and culminate in sustained tonic contraction. Spasms occur spontaneously but are exacerbated by voluntary facial movement, stress, fatigue, anxiety, or change in head position (41). Although significantly decreased compared to wakefulness causing arousals, hemifacial spasm persists during sleep (68). The movements often persist during sleep. Stapedius muscle contraction frequently accompanies contraction of the muscles of facial expression and sometimes produces low-pitched tinnitus in the ipsilateral ear (149; 150). However, evidence of spontaneous activity in nonmotor functions, such as lacrimation and salivation, is not seen (53). Clinically apparent facial weakness may be seen in longstanding cases and is demonstrable with objective testing in the eyelid in 47% and other facial muscles in 77% of cases (48). Other cranial nerves, particularly the acoustic nerve, may be involved. As many as 50% of patients have abnormalities of the acoustic middle-ear reflex (105) and some degree of hearing loss can be found in nearly 15% (41). Concurrent trigeminal nerve dysfunction results in the syndrome named by Cushing, "painful tic convulsif." It has been estimated that 0.2% to 11% of patients with trigeminal neuralgia have associated hemifacial spasm (98). Coexistence with blepharospasm has also been reported (137). In a study of 95 patients, hand tremor accompanied hemifacial spasm in 40% of cases and 10% of control subjects (119).

Prognosis and complications

Hemifacial spasm is a chronic condition. Spontaneous resolution occurs in fewer than 10% of cases (41). Although early in the course the facial twitching may be no more than a nuisance, progression is the rule. Severe involvement can cause disability due to loss of binocular vision. There are no known complications of untreated hemifacial spasm.

Clinical vignette

A 44-year-old right-handed woman developed twitching of the right eyelid. She gradually noticed spread of the movements to the lower part of the face over the course of months. MRI brain was negative. She was treated with clonazepam and carbamazepine without relief. She had no significant medical history and no family history of movement disorders. Neurologic examination abnormalities were limited to the face. There were nearly continuous rapid synchronous contractions of the right side of the face. Injections of 25 units of botulinum toxin A at four sites over the orbicularis oculi significantly reduced the movements.

Biological basis

Etiology and pathogenesis

Hemifacial spasm is thought to result from compression of the facial nerve at its junction with the brainstem (97). Rarely, compression along the course of the nerve outside of the cranium has been reported to cause hemifacial spasm (117). In most cases, the compression is by blood vessels adjacent to the zone of exit of the facial nerve (149; 150). Contralateral fusiform aneurysms of the vertebral artery causing hemifacial spasm have also been reported (106). Digre reviewed 1688 cases of hemifacial spasm from the literature. In the 539 cases where etiology was specified, the cause of hemifacial spasm was compression by vascular structures in 94%, tumor in 4%, and bony or other abnormalities in 2%. In a series of 1642 hemifacial spasm cases, only nine resulted from a cause other than neurovascular compression (seven tumors, two vascular malformations) (60). Causative tumors include epidermoid neuroma, meningioma or astrocytoma of the cerebellopontine angle, petrous ridge meningioma, and parotid tumor. In a series of nine patients with cerebellopontine angle tumors, there were two vestibular schwannomas, five meningiomas, and two epidermoid tumors (92). Cerebropontine lipomas have also been shown to cause unilateral hemifacial spasm (39). Basilar impression, cranio-occipital malformation, and Paget disease were etiologic in a few cases (38). Other rare causes of hemifacial spasm are large posterior fossa masses, Chiari I malformation, and brainstem cysticercosis. Hemifacial spasm has been reported to result from parotid tumor (16). Transient hemifacial spasm has been reported in the context of diabetic ketoacidosis (13) and mandibular trauma (128). Hemifacial spasm has also been reported in patients with multiple sclerosis (99) and ocular myotonia (83). There have been a few cases of idiopathic intracranial hypertension associated with hemifacial spasm (116; 52).

Vascular compression or irritation of the facial nerve is the most common pathogenic mechanism of hemifacial spasm. Dolichoectatic vertebrobasilar arteries or compression of the facial nerve root exit zone can be demonstrated in up to 96% of cases using computed tomography scan, modified magnetic resonance imaging, 3-dimensional fast-spin magnetic resonance imaging, surface-rendered 3-dimensional magnetic resonance imaging, 3-dimensional phase contrast magnetic resonance angiography, or magnetic resonance tomographic angiography techniques (50; 111). In a study of 115 consecutive hemifacial spasm patients undergoing microvascular decompression, all patients had neurovascular compression (21). The offending vessel was the anterior inferior cerebellar artery in 43%, posterior inferior cerebellar artery in 31%, vertebral artery in 23%, and a large vein in 3%. Thirty-eight percent of patients had more than one compressing vessel. Along the course of the facial nerve, compression was most frequent at its attached segment and less common in the root detachment point, root exit point, and distal cisternal portion. A case report revealed 2 mm longitudinal penetration of the distal portion of the facial nerve root exit zone by the anterior inferior cerebellar artery as a cause of hemifacial spasm (112). In another case report anomalous, enlarged branch of the ascending pharyngeal artery was found to be the cause of hemifacial spasm (141).

Hypertension may play an important role in the genesis of hemifacial spasm. In a case-control study of patients with primary hemifacial spasm, hypertension was more common in cases than controls (odds ratio 2.76; 95% CI 1.43-5.33). Subjects with hemifacial spasm and hypertension are more likely to have imaging evidence of ventrolateral medullary compression than nonhypertensive subjects (36; 23). Neuroimaging of controls with hypertension reveals evidence of ventrolateral medullary compression in nearly 75% of cases (23; 26). It is important to recognize that among control subjects, neuroimaging studies suggest a sizeable number have areas of contact between the facial nerve and vascular structures (49%) or indentation of the ventrolateral medulla (41%), and this number is increased significantly in the presence of hypertension (72%) (23; 80). A prospective controlled study using 3-dimensional MR volumetric analysis showed that patients with hemifacial spasm have 11.4% smaller CSF volume in the posterior fossa (24). A study of voxel-based morphometry used to evaluate changes in gray matter by using T1-weighted imaging in 25 patients with hemifacial spasm and 25 demographically similar healthy volunteers found reductions in the thalamus, putamen, pallidum, dorsolateral prefrontal cortex, amygdala, and parahippocampal gyrus in patients with hemifacial spasm compared with healthy volunteers (14). Another voxel-based morphometry study found grey matter volume was reduced in the right inferior parietal lobule and increased in the cerebellar lobule in 8 of 42 patients with hemifacial spasm, when compared with 30 control subjects (144).

Support for the theory that compression of the nerve at the root exit zone incites the development of hemifacial spasm comes from several sources. Neuroimaging studies frequently show contact between the nerve and the offending structures. Biopsies of root exit zones of compressed nerves from patients with hemifacial spasm show demyelination of axis cylinders or nerve degeneration (29). Surgical relief of compression in the region is usually successful in relieving the disorder (75). However, the mechanism of production of spasm is unknown. The theory that focal demyelination or nerve degeneration causes aberrant regeneration similar to that which follows facial nerve injury is untenable as it fails to explain the spasms’ spontaneity and synchronous occurrence in muscles supplied by different nerve branches.

An alternative theory of "ectopic" or "ephaptic" transmission was proposed by Gardner and has been championed by Nielsen (53; 107; 108). According to this theory, nerve compression and the resulting demyelination cause an "ephapse" to form. An ephapse is a false synapse where ectopic activity may be triggered by mechanical irritation, changes in regional electrolyte concentration, or flow of extracellular current during passage of nerve impulses in adjacent fibers (108). This ectopic activity may then be conducted orthodromically or antidromically within the nerve fiber. Nielsen has studied electrical phenomena that seem to support this theory. When the zygomatic branch of the facial nerve is stimulated in hemifacial spasm, a direct response is obtained in the orbicularis oculi muscle just as in normal controls. However, in hemifacial spasm, there is also a response in the mentalis muscle, which is supplied by the mandibular branch muscle and not seen in the unaffected side or in normal controls (108). The authors concluded that lateral spread of current to branches of the nerve other than those stimulated occurs by means of crosstalk at the compressed area of the nerve. Following surgical relief of vascular compression of the facial nerve, these electrophysiologic abnormalities resolve (08). Sanders used single fiber EMG to determine whether nonsynaptic transmission occurred between branches of the facial nerve. "Jitter," variations in latency between the presynaptic stimulus and the muscle response, is a measure of synaptic delay. The demonstration in two patients with hemifacial spasm that jitter was low suggested that inappropriate spread to other muscles occurred by nonsynaptic means (122). This further supports the notion of ephaptic transmission. Motor unit potential analysis suggests the demyelination causes secondary axonal loss in hemifacial spasm (84). Reduction in the lateral spread response amplitudes in the orbicularis oculi, orbicularis oris, and mentalis muscles by desflurane suggests a central mechanism is involved in the lateral spread response (151).

However, additional studies suggest that the point of generation of abnormal activity is central to the area of vascular compression. During microvascular decompression operations, Moller recorded the conduction time from a stimulation site on the marginal mandibular branch of the facial nerve to the facial root exit zone (104). He then recorded the latency from stimulation at the root exit zone to the muscle response in the orbicularis oculi. The latency from marginal mandibular stimulation to orbicularis oculi contraction exceeded by 2 msec the sum of conduction times from the marginal mandibular branch to the root exit zone and from the root exit zone to the orbicularis oculi (104). The F wave, an indicator of motoneuron conduction and anterior horn excitability, is enhanced in patients with hemifacial spasm (71).

Patients with hemifacial spasm have a lower threshold to transcranial magnetic stimulation on the affected side (85), and paired transcranial magnetic stimulation shows a lessening of facilitation of the test motor evoked potential by the conditioning stimulus at certain interstimulus intervals, suggesting changes in facial motor neuron excitability (22). Patients with hemifacial spasm have enhanced F waves on the symptomatic side, another sign of increased excitability of the facial nucleus (110). Hyperexcitability of the blink reflex (40) and F waves (70; 69) in patients with hemifacial spasm further support this theory of pathogenesis. These studies suggest that the abnormal activity is generated central to the exit zone, probably at the facial nucleus. Moller proposed that aberrant afferent activity generated at the site of vascular compression caused reorganization of the facial nucleus by a mechanism similar to the "kindling" phenomenon described in the hippocampus (103). Following surgical decompression in 20 patients with hemifacial spasm, the excitability of the F wave decreased in patients who had only partial resolution of spasm postoperatively. In patients whose spasms disappeared completely after surgery, F-wave hyperexcitability completely resolved, though over a time course of up to two years (70). Successful botulinum toxin injection may also be associated with reductions in facial motor nucleus hyperexcitability as assessed with F waves (72). In patients, the electrophysiologic abnormalities of hemifacial spasm can be produced by repetitive simulation of the facial nucleus. It is also possible to induce facial nucleus hyperexcitability and the electrophysiologic abnormalities of hemifacial spasm by producing an area of demyelination and vascular compression at the facial nerve exit zone (89).

Epidemiology

Data from Olmsted County, Minnesota suggest that the average age-adjusted annual incidence of hemifacial spasm for all ages is 0.78 per 100,000 (0.81 per 100,000 in women and 0.74 per 100,000 in men). The prevalence is 14.5 per 100,000 in women and 7.4 per 100,000 in men (09).

Also see the article titled Epidemiology of movement disorders.

Differential diagnosis

Hemifacial spasm must be differentiated from other conditions that cause involuntary facial movements. These include conditions wherein the abnormal movement is organized at the lower motor neuron level as well as those wherein the brainstem or higher structures are involved. In a series of 215 patients referred to a tertiary movement disorders center for evaluation of involuntary facial movements, there were 38 (18%) patients with hemifacial spasm mimickers classified as psychogenic, tics, dystonia, myoclonus, and hemimasticatory spasm (156).

Postparalytic facial spasms. Postparalytic facial spasms generally reflect either fixed contraction of the facial muscles, synkinesis, or both. There is always a history of preceding facial weakness, and spontaneous spasms are generally absent.

Facial myokymia. Facial myokymia is characterized clinically by undulating movement of facial muscles. Electrophysiologically, there is rhythmic bursting of single muscle fascicles. The frequency of motor unit discharge ranges from 20 to 200 Hz. The bursts are asynchronous in various muscles of the face as well as in different motor units in the same muscle.

Spastic paretic facial contracture. Spastic paretic facial contracture may be confused with hemifacial spasm. The involved side is weak and contracted, unlike hemifacial spasm, wherein the facial muscles are relaxed between twitches. The presence of concomitant myokymia is also helpful in distinguishing the conditions. Spastic paretic facial contracture usually indicates focal pathology in the brainstem.

Blepharospasm and Meige syndrome. Blepharospasm and Meige syndrome are forms of focal dystonia. In blepharospasm, dystonic movements involve the orbicularis oculi muscle. Meige syndrome includes blepharospasm and dystonic movements of the lower face. Because Meige was not the first to describe the disorder, the term cranial dystonia is more appropriate for the combination of upper and lower face dystonia. Dystonic movements are generally bilateral and improve during sleep. Upper and lower facial involvement is generally not synchronous. When the lower face is involved, there is often clinically apparent involvement of muscles other than those innervated by the facial nerve.

Tics. These are rapid, stereotyped movements that resemble normal, coordinated movement. They vary in intensity and are arrhythmic. Tic sufferers are usually able to voluntarily reproduce and to transiently suppress the movements. Although tics frequently affect the orbicularis or facial muscles, most patients have additional tics in muscles not supplied by the facial nerve.

Tardive dyskinesia. This results from chronic treatment with dopamine receptor-blocking agents. Movements may be choreic or dystonic. Most often bilateral, the movements are irregular, and when upper and lower face are involved, are usually asynchronous and involve additional muscles other than those innervated by the facial nerve.

Psychogenic hemifacial spasm. These patients have more frequent tonic muscle contractions, bilateral asynchronous hemifacial involvement, isolated lower facial involvement, downward deviation of the mouth’s angle, and lack of the “other Babinski sign” compared to those with organic hemifacial spasm (12).

Diagnostic workup

The need for a diagnostic workup in patients with classic hemifacial spasm is debatable. Aberrant vessels have been identified by contrast-enhanced computed tomography (38), magnetic resonance imaging (140), and modified magnetic resonance angiography studies (01; 142) in 80% to 100% of patients with hemifacial spasm in some series. Using advanced magnetic resonance imaging techniques with 3-dimensional fast spin echo and the minimum intensity projection method enhances the ability to visualize vascular structures compressing the facial nerve (101). Three-dimensional MR cisternography was said to have 100% sensitivity and 94% specificity for vascular compression in a series of 14 patients with hemifacial spasm compared to 20 control subjects (91). Another high-yield magnetic resonance technique is 3-dimensional constructive interference in steady state (CISS). This technique showed 100% sensitivity for identifying vascular structures compressing the facial nerve in a study of 47 subjects with hemifacial spasm (139), but only 77.27% sensitivity and 75% specificity in another study of 44 patients with hemifacial spasm and 76 controls (56). Vertebral angiography adequately visualizes the offending vessels in a large number of cases and may increase the safety of surgical interventions (113). Tumors are rarely identified. In a survey of 190 investigators comprising treatment records of 1676 patients with hemifacial spasm, only nine tumors were found. Indeed, only 881 patients had neuroimaging studies (132). This reflects a change in patient management away from surgery and towards more conservative treatment strategies, such as botulinum toxin injection. Patients who are clinically atypical with signs of regional focal pathology should have neuroimaging studies. In this group, magnetic resonance imaging is the imaging study of choice. Thin-slice T2 MRI imaging of the facial nerve can guide clinicians in selecting patients who are good surgical candidates (126). A study of cerebral glucose metabolism using PET with (18) F-fluorodeoxyglucose demonstrated bilateral cerebral glucose hypermetabolism in the thalamus of patients with unilateral hemifacial spasm (127). Patients who do not improve with conservative management and who are candidates for surgical intervention should have neuroimaging at the discretion of the operating surgeon. Preoperative auditory screening may also be useful.

A study of inflammatory makers demonstrates that serum level of interleukin-6 correlated with the severity of hemifacial spasm (95).

Management

Several treatment options involving oral medications are available to the physician. Most trials of oral medications have been empiric and uncontrolled. The clinical similarities between trigeminal neuralgia and hemifacial spasm suggest that anticonvulsant therapy may be useful. Carbamazepine is often the first intervention. Alexander and Moses reviewed published reports of carbamazepine therapy of hemifacial spasm in 46 patients in seven series. Sustained improvement was seen in 35% and complete control in an additional 22% of cases (03). An open-label trial of gabapentin in 23 patients with hemifacial spasm showed significant reduction of spasm in 16 (34). Hughes studied orphenadrine and dimethylaminoethanol in a placebo-controlled trial in 13 patients with hemifacial spasm. He found that 62% of patients in this study had a beneficial response to anticholinergic medication (64). Sandyk and Gillman reported beneficial effects of baclofen in six patients in open trials (123). Other medications sometimes used in the treatment of hemifacial spasm include haloperidol and clonazepam. However, in 2015 a randomized double-blind placebo-controlled trial of clonazepam 0.5 mg twice a day versus placebo for four weeks did not show a statistically significant effect on hemifacial spasm (11). Levetiracetam may be beneficial in the treatment of hemifacial spasm (88). Anecdotal reports in hemifacial spasm associated with Paget disease suggest that treatment of the bone disorder with calcitonin or alendronate may improve the facial spasms (49).

There have been numerous extracranial neurosurgical approaches to the treatment of hemifacial spasm. Unilateral removal of the orbicularis oculi, corrugator superciliaris, and procerus muscles may be effective in controlling spasm around the eye, but offers no benefit for lower facial spasm. Adverse effects are mild (54). Early procedures included sectioning or injuring the peripheral nerve trunk or its branches. These procedures may relieve spasms but are accompanied by various degrees of facial weakness. Although facial strength often returns in 3 to 6 months, it is invariably accompanied by return of the facial spasms. A modification of this procedure wherein an anastomosis is made between the hypoglossal or spinal accessory nerve and the severed facial nerve in an attempt to prevent regrowth of the nerve does not significantly improve the results (149; 150). Attempts to injure the facial nerve using injections of alcohol or phenol afford only transient benefit. Other extracranial procedures include transtympanic puncture of the nerve (109) or insertion of tantalum wire into the nerve, decompression or longitudinal neurotomy in the facial canal, and radiofrequency thermocoagulation of the facial nerve. In all cases, facial weakness accompanies relief of spasms and both facial strength and spasms return in time (149; 150). Somewhat longer-lasting benefit has been achieved in patients treated by percutaneous puncture of the facial nerve at the stylomastoid foramen. Although facial palsy and relief of spasm are immediate, facial strength returns before the spasms. Extreme pain accompanies this procedure, and patients are often unwilling to undergo repeated treatments (74). Ito and colleagues described a technique of permanent selective facial nerve block using doxorubicin under local anesthesia in a 73-year-old woman with hemifacial spasms, but further investigation of this technique is required (73).

Intracranial surgical procedures have been used for the treatment of hemifacial spasm for more than 50 years (97). In 1962, Gardner reported the results of microvascular decompression of the facial nerve in the cerebellopontine angle in 19 patients. Janetta and others have subsequently reported large series of patients treated in this fashion. Review of 18 such series, comprising 4969 patients, reveals that 58% to 97% of patients markedly improved following one or more procedures. Overall, 84% had good to complete resolution of spasm, and 4% failed to respond. Recurrence rates varied from 0.4% to 10% (43; 152; 81; 86; 74; 44; 96; 07; 78; 149; 63; 02; 150; 15; 67; 129; 65). A prospective study of 1174 patients with hemifacial spasm who underwent microvascular decompression suggested a “cure” in 94.1% of cases (66). In another study, a metaanalysis of 39 studies including 6249 patients who underwent microvascular decompression surgery revealed that the overall spasm freedom rate was 90.5% at a follow-up of 1.25 ± 0.04 years (62).

As many as 50% of surgically treated patients have transient recurrence postoperatively beginning around four days postoperatively and improving gradually, with a median time to resolution of 28 days (69). The University of Pittsburgh investigators described the results of 41 repeat microvascular decompression procedures in 36 patients over a 3-year period; 5.9% of them failed the reoperation (42). Another study of 78 patients who had repeat microvascular decompression surgery revealed that 72 (98%) achieved complete resolution (155). Follow-up MRI imaging suggests that the degree of vascular depression afforded by the operative procedure does not predict the magnitude of subsequent benefit (25). In two studies that included 276 patients who were followed for 10 years, 90% had excellent recovery (57; 121). Follow-up of 25 cases of hemifacial spasm following decompression showed improvements in quality of life (61; 27). Some authors advocate intraoperative recording of facial nerve-innervated muscles as a prognostic indicator. A decrease in the lateral spread response (activity in muscles innervated by one or more branches of the facial nerve following electrical stimulation of another facial nerve branch) after relieving pressure on the facial nerve root entry zone predicts prognosis in some (87), although this remains controversial (35; 145). In another study, intraoperative facial nerve motor evoked potentials elicited by transcranial electrical stimulation were monitored (51). A reduction in facial nerve motor evoked potentials in the orbicularis oculi at the procedure end was associated with clinical improvement. In a study of 1335 patients who underwent microvascular decompression surgery, patients were classified as rapidly progressive or slowly progressive (90). The study found the following significant predisposing factors for rapid progression: patients were younger age at surgery, older age at symptom onset, and absence of intraoperative facial nerve indentation. The rapidly progressive group had worse outcomes than the slowly progressive group following microvascular decompression.

Complications of microvascular decompression are common, but most are not serious. In large series, up to 35.8% of consecutive patients have one or more complications, including facial palsy (7% to 18.6%), hearing deficit (1% to 7.2%), and lower cranial nerve palsies (2.8%) (65; 66). Facial palsy may be delayed up to 16 days following surgery. High-frequency hearing loss has been reported on both the ipsilateral and contralateral sides following surgery (157). Intraoperative monitoring of the brainstem auditory evoked potential may reduce the likelihood of permanent deafness by allowing surgeons to modify their technique in response to changes in the waveform (131). Other complications of these procedures include facial or scalp hypesthesia, dysequilibrium, encephalopathy, anosmia, aseptic or bacterial meningitis, phlebitis, serous otitis media, wound infection, cerebrospinal fluid rhinorrhea, tension pneumocephalus, pneumonia, pulmonary embolus, and gastrointestinal bleeding (63). Death (152) and brainstem infarction (120) are rare complications. Elderly patients have a similar response to surgery and side effect profile, according to one study (77). Endoscopic vascular decompression has shown comparable efficacy to microvascular decompression with less morbidity and shorter hospital stay (05).

In a series of three patients, radiosurgery (single dose of 3 Gy in one patient) and hyperfractionated stereotactic radiotherapy (15 Gy divided into five fractions in two patients) targeting the vestibulocochlear-facial bundle just outside the internal acoustic foramen resulted in markedly reduced symptoms over a mean follow-up duration of 24 months (46).

Hemifacial spasm related to nerve compression by large vascular structures, such as arteriovenous malformations or fusiform aneurysms of the basilar artery, may be amenable to treatment with endovascular embolization (124).

Botulinum toxin has emerged as the treatment of choice for hemifacial spasm (82; 59). For the treatment of hemifacial spasm, botulinum toxin is injected into the subcutaneous tissue overlying the orbicularis oculi muscle at 3 to 10 sites around the eye. When lower facial spasms are disabling, small amounts of toxin may be injected into the lower facial muscles as well. A systematic review of published clinical trials revealed only one small placebo-controlled study of botulinum toxin A, which supported the conclusions of multiple open-label studies (32). A review of published articles between 1991 and 2021 regarding the effectiveness and safety of botulinum toxin in hemifacial spasm as well as any reported differences among botulinum toxin showed that efficacy of BoNT for hemifacial spasm treatment ranged from 73% to 98.4% (135). A cross sectional study of 1003 patients with hemifacial spasm at 15 tertiary centers in China demonstrated a mean latency of botulinum toxin A effect was (5.0 ± 4.7) days, and the mean total duration of the effect was (19.5 ± 11.7) weeks (147). A randomized, double-blind crossover trial of 20 patients revealed no significant difference in the onset time and therapeutic efficacy between high concentration (50 U/mL) and low concentration (25 U/mL) groups; however, the duration of efficacy was longer in the high-concentration group than in the low-concentration group. Patients in the high-concentration group had more serious adverse reactions, and the reactions lasted for longer duration (94).

A study comparing patients with and without prior botulinum toxin treatment after microvascular decompression revealed no significant differences in the outcomes and complications after a minimum of nine months of follow-up (148). In another study that followed 976 patients with hemifacial spasm, 452 patients underwent microvascular decompression only and 524 patients first had botulinum toxin treatment then underwent microvascular decompression. The patients were followed for 7 to 9 years, and the long-term postoperative relief rate was 98.23% for the first group and 87.21% for the second group, suggesting that patients treated with microvascular decompression experience better long term efficacy (158). A report confirmed the high efficacy and safety profile of botulinum toxin A for long-term treatment of hemifacial spasm (28). The benefits of botulinum toxin are typically longer than the average 3 to 4 months seen in other indications, and the effects can be sustained for several decades (118). Common adverse effects are eyelid and facial weakness (up to 100%), ptosis (up to 23%), and diplopia (up to 13%). Less commonly reported adverse effects include excessive tearing, dry eyes, corneal exposure and keratitis, and fatigue (47). Excessive facial weakness is more common in patients in whom preexisting denervation can be demonstrated by EMG (04). In a retrospective analysis of patients with post-facial palsy synkinesis and hemifacial spasm treated with botulinum toxin, complications were higher in patients treated for post-facial palsy synkinesis (77%) compared to hemifacial spasm (34%) (19). Repeated injections are well tolerated, and benefit is maintained over years of treatment (82; 33; 134). Both serotypes A and B are effective (143). Botulinum toxin injections result in improved patient quality of life (138). An evidence-based medicine report found one class II and one class III study in support of botulinum toxin injections for hemifacial spasm (130). In a retrospective study of 16 patients who received 30 or more serial botulinum toxin A treatments, the mean duration of relief was 12.4 weeks (55). Injection of doxorubicin into the eyelids has also been reported effective for up to six years (153).

A controlled study of acupuncture and pressure on the otopoints in 86 patients suggested that the combination of these two traditional medicine approaches was superior to either technique alone (93). A questionnaire completed by 96 patients with hemifacial spasm indicated that complimentary therapies, most commonly acupuncture and facial massage, were tried in more than 50% of these patients (115).