Pathophysiology Ofhemimasticatory Spasm 45

Home , Deep temporal nerves, Facial muscles, Medial pterygoid muscle, Temporal muscle

J'ournal ofNeurology, Neurosurgery, and Psychiatry 1994;57:43-50 43

Pathophysiology of hemimasticatory spasm J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.1.43 on 1 January 1994. Downloaded from

G Cruccu, M Inghilleri, A Berardelli, G Pauletti, C Casali, P Coratti, G Frisardi, P D Thompson, M Manfredi

Abstract the trigeminal nerve. We also reviewed pub- Two patients aged 21 and 50 years pre- lished reports of patients literature in whom a sented with facial hemiatrophy and uni- diagnosis of HMS was supported by EMG lateral spasms of the masticatory findings, to examine the presence of a com- muscles. Masticatory muscle biopsy mon pathophysiological mechanism. showed normal findings in both patients and facial skin biospy specimens only showed atrophy, although morphoea Case reports (localised facial scleroderma) had been PATIENT 1 diagnosed nine years previously in the A young man first noticed a hollowing of his second patient. The involuntary move- left cheek at the age of 17. One year later, ments consisted of brief twitches and brief involuntary twitches appeared in his left prolonged contractions clinically and temporalis muscle. Clinical examination dis- electromyographically similar to those of closed dilatation of the left pupil and the hemifacial spasm and cramps. The jaw facial asymmetry; CT and magnetic reso- jerk and the silent periods were absent in nance imaging (MRI) of the brain were nor- the affected muscles. Direct stimulation mal. Left facial hemiatrophy was diagnosed. of the muscle nerve and transcranial At the age of 20 years, when the patient stimulation of the trigeminal root came to our observation, the involuntary demonstrated slowing of conduction and movements were still restricted to the tempo- after-activity due to autoexcitation. ralis muscle. Small twiches alternated with Observations in other reported cases and severe and painful, prolonged contractions these two patients suggest that hemimas- (lasting up to several minutes). Clinical ticatory spasm is produced by ectopic examination showed only marked hypertro- activity secondary to focal demyelination phy of the left temporalis muscle and hollow- of the trigeminal motor nerve fibres. The ing of the cheek. Facial sensation was normal proposed cause of the neuropathy is focal and the patient never complained of paraes- damage to the masticatory nerves caused thesias or neuralgia. by compression, possibly resulting from No clinical or laboratory evidence-includ- University of Rome the deep tissue changes that occur in ing antinuclear antibodies, extractable 'La Sapienza', Rome, facial hemiatrophy. nuclear antigen, native DNA, and mitochon- http://jnnp.bmj.com/ Italy drial antibody titres-was found of connec- Department of (T Neurol Neurosurg Psychiatry 1994;57:43-50) tive tissue disease. Neurosciences G Cruccu The patient has been taking 1200 mg car- M Inghilleri bamazepine daily for six months and reports A Berardelli Hemimasticatory spasm (HMS) is a rare con- moderate benefit. G Pauletti P Coratti dition, characterised by unilateral forceful

M Manfredi contractions of one or more masticatory mus- PATIENT 2 on September 29, 2021 by guest. Protected copyright. Institute of Nervous cles. In most patients it is associated with A 50-year-old woman first noticed small and Mental Diseases ipsilateral facial hemiatrophy." The cause of spots of abnormal pigmentation on the right C Casali the spasms is not understood. side of the face at the age of 41 years. On her Department of Kaufman,3 who first studied the elec- admission to a dermatological division, a skin Odontology tromyographic (EMG) characteristics of biopsy was consistent with the diagnosis of G Frisardi hemimasticatory spasm, stressed the close MRC Human localized scleroderma of the face, or Movement and similarity with hemifacial spasm, and pro- morphoea. Corticosteroids were prescribed. Balance Unit and posed that the spontaneous activity was gen- At the age of 44 years she began to have University erated in the trigeminal nerve fibres. involuntary twitches of the Department of right masseter and Clinical Neurology, Thompson and Carroll4 found similar EMG temporalis muscles; the twitches increased in Institute of Neurology, characteristics and also proposed a trigeminal duration and frequency over the years. London, UK neuropathy. Other authors, however, sug- At her first neurological assessment, the P D Thompson gested CNS, sympathetic ganglia, or muscle dyschromia was less apparent, but the face Correspondence to: Dr Giorgio Cruccu, dysfunction.56 Because none of the previously was frankly asymmetrical because of a slight Dipartimento Scienze reported patients underwent masticatory hollowing of the right cheek and marked Neurologiche, Viale Universita' 30, I-00185 nerve stimulation, conduction velocity and hypertrophy of the right masseter and tempo- Rome, Italy. after-discharges could not be studied. ralis muscles. The involuntary masticatory Received 12 January 1992 We studied two patients with HMS and muscle contractions consisted of brief and in revised form 13 April 1993. facial hemiatrophy and demonstrated ectopic twitches and prolonged spasms, occurring Accepted 19 April 1993 activity and slowing of motor conduction in many times a day either spontaneously, or 44 Cruccu, Inghilleri, Berardelli, Pauleti, Casali, Coratti, Frisardi, Thompson, Manfredi

more often, triggered by chewing, speaking or A J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.1.43 on 1 January 1994. Downloaded from other voluntary movements of the mouth and jaw. Tonic contractions could last up to a few minutes; they were most painful and pre- vented the patient from opening her mouth. ,-1i Like patient 1, she had never had sensory sysmptoms other than the pain induced by the muscle spasms. Neurological examination disclosed no abnormal signs apart from the spasms and hypertrophy of the masticatory B muscles. MRI and CT scans showed hyper- I n.l-1-ii- trophy of the right temporalis, lateral ptery- 1111117 goid, and masseter muscles, and dislocation of the right temporomandibular joint. V- Serum tests for speckled and homogeneous antinuclear antibodies, and smooth muscle r if' antibodies gave positive results; circulating immune complexes were increased. Cu Extractable nuclear antigen, native DNA, and mitochondrial antibody titres were negative. No evidence was found of kidney, intestinal, or other systemic manifestations of vasculitis. The painful spasms were attenuated by carbamazepine 600 mg daily and transiently blocked by injection of local anaesthetics into Figure 1 Involuntary EMG activity in patient 1. (A) Highfrequency discharges of a single motor unit the masseter and temporalis muscles. potential. Calibration 10 msll mV. (B) Briefbursts of Diazepam was ineffective. After completion multiple motor unit potentials (corresponding clinically to of all the twitches). Calibration 100 msll mV. (C) Large investigations, injection of botu- compoundpotential ofmultiple synchronised motor units linum A toxin (30-50 U, Oculinum) into the discharging tonically at a 60 Hzfrequency (corresponding affected muscles was repeated three times, clinically to prolonged spasms). Calibration 50 msllO m V. with clinical benefit. The following investigations were approved by the local ethical committee and both and the right masseter and temporalis mus- patients gave their informed consent. cles in patient 2. Involuntary EMG activity was similar in the two patients. Brief trains Investigations (30-100 ms) of two to seven single motor MUSCLE AND SKIN BIOPSY unit potentials reaching a discharge frequency In both patients, an elliptical biopsy speci- of 100 Hz (fig IA), or brief bursts (50-200 men, including skin and underlying tempo- ms) of multiple motor unit potentials ralis muscle, was obtained from the affected occurred spontaneously, irregularly, and side. In patient 2, a further skin biopsy speci- arrhythmically (fig 1 B). The bursts would men was taken from an area of slight dyschro- occasionally become more frequent and mia on the right forehead. intense, with the progressive recruitment and http://jnnp.bmj.com/ In both patients, the muscle was histologi- synchronisation of more and more motor cally normal and skin specimens only showed units, leading to a large (12-15 mV) com- a flattened epidermis and atrophy of dermal pound potential, discharging tonically at a appendages. In particular, no evidence of high frequency (fig 2C). The discharge fre- scleroderma was found in patient 2, in whom quency was always around 60 Hz in patient 1 morphoea had been diagnosed nine years and 70 Hz in patient 2. This activity corre-

beforehand. sponded to the painful tonic contractions. on September 29, 2021 by guest. Protected copyright. A masticatory nerve biopsy was excluded The maximum duration of the contractions because of the consequent functional dam- was 2 minutes in patient 1 and 30 s in patient age. 2. In patient 2, the prolonged spasm sometimes affected the temporalis muscle alone STANDARD EMG EXAMINATION with complete electrical silence in the mas- The masseter, temporalis, suprahyoid, and seter or vice versa. At other times the spasm facial muscles were bilaterally examined by moved from one muscle to the other (fig 2), concentric needle recordings (filters 50-5000 but never spread to involve the facial muscles Hz). No abnormalities were found in either or contralateral muscles of the jaw. In both patient in the muscles clinically unaffected by patients, these patterns of involuntary EMG the spasm, neither were denervation poten- activity usually followed a strong voluntary tials found in the affected muscles. The vol- clenching of the teeth, but could also be untary and involuntary motor unit potentials elicited by electrical stimulation of the men- examined were all normal, although difficulty talis or infraorbital nerves, though not the in providing a steady level of voluntary activ- supraorbital nerve. ity and the frequent occurrence of involuntary activity hindered an exhaustive analysis of STIMULATION OF TRIGEMINAL MOTOR NERVE single motor unit potentials. FIBRES Paroxysmal spontaneous activity was Signals were recorded through surface elec- recorded from the left temporalis in patient 1 trodes (filters 10-2000 Hz) placed bilaterally Pathophysiology ofhemimasticatory spasm 45

Figure 2 Temporalis and over the masseter and temporalis78 and by J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.1.43 on 1 January 1994. Downloaded from masseter spasm in patient concentric needles inserted into the anterior 2. Simultaneous recordings from right temporalis (T) temporalis. and masseter (M) muscles. The masseteric and deep temporal nerves Large compound potentials T in the infratemporal fossa (fig 3) were stimu- ofsynchronised motor units discharging tonically in the lated by delivering electrical shocks (0-5-5 temporalis muscle whereas mA, 0-1 ms) through two fine needle elec- the masseter is silent (1). trodes inserted below the zygomatic arch and After afew seconds, the M ------number of motor units anterior to the temperomandibular joint, and discharging synchronously direct motor responses (M waves) were in the temporalis decreases 2 recorded from the masseter and temporalis whereas similar activity muscle.78 In patient 1, the masseter M wave starts in the masseter (2). owt A q .-. W, The spasm ceases T was normal and symmetrical; the left tempo- simultaneously in the two ralis M wave was slightly delayed (right muscles (3). Calibration latency 1-8 ms, left latency 2-2 ms) and nor- 50 ms/2 m V. mal in amplitude (4.8 mV on the right, 5-1 mV on the left). In patient 2, the response in m the right masseter was clearly delayed (2.4 ms I I MI on the right, 1-5 ms on the left), although symmetrical in amplitude (6 mV on both sides) (fig 4). 3 I The trigeminal root was stimulated trans- 48'VYW V~--VYyvu,yTry V' cranially910 by means of a high tension, low impedance electrical cortical stimulator (Digitimer), with the anode placed at the ver- tex and the cathode about 10 cm laterally, slightly anterior and superior to the ear. The A H ~~~~R

1 II

B -j http://jnnp.bmj.com/

2 on September 29, 2021 by guest. Protected copyright.

Figure 3 Schematic drawing ofthe infratemporalfossa and course of the masticatory nerves. Lateral view (left) andfrontal section (right). The mandibular nerve leaves the Figure 4 Motor conduction study. Recordingfrom the cranial cavity through theforamen ovale of the greater wing of the sphenoid, and runs right (R) and left (L) temporalis muscle in patient 1 (A) down into the infratemporalfossa, behind the lateralpterygoid muscle. At its exitfrom the and the right and left masseter muscle in patient 2 (B). foramen ovale itfirst gives off the motor branches that supply thejaw closing muscles. The Electrical stimulation of the masticatory nerves (masseteric temporo masseteric nerve turns laterally, running horizontally between the greater wing of and deep temporal nerves) in the infratemporalfossa (1) the sphenoid and the upper belly of the lateral pterygoid muscle (3), and divides into the and transcranial stimulation of the trigeminal motor root posterior temporal nerve (1) and masseteric nerve (2). The temporal muscle is also (2). In (A), the direct motor responses evoked by innervated by the median and anterior temporal nerves, which originate from the stimulation of the deep temporal nerves (1) are only mandibular nerve and the buccal nerve (4). The deep temporal nerves (1) turn sharply slightly asymmetrical, whereas those evoked by stimulation around the temporal crest and run upwards between the skull and the temporal muscle. of the trigeminal motor root (2) are markedly delayed in The masseteric nerve (2) turns downwards, passing between the zygomatic arch and the the left temporalis. In (B), the responses evoked by mandibular notch to reach the masseter musclefrom its inner surface. The lateralpterygoid stimulation of both the masseteric nerve (1) and the muscle is innervated by nerve branches (not shown) originating, close to the muscle, from trigeminal motor root (2) are clearly delayed on the right the mandibular nerve and the buccal nerve; the medial pterygoid muscle is innervated by a masseter. In both patients, the amplitude of the responses to small nerve (not shown), which originates slightly below theforamen ovale and remains distal stimulations (1) is normal. Averages offour trials, posterior to the mandibular nerve. All the nerve branches that supply thejaw closing except the bottom records in (A), wherefour single trials muscles originate before the mandibular nerve divides into its terminal mixed nerves: the are superimposed to differentiate the direct motor response lingual (5) and inferior dental (6) nerves. Thejaw opening muscles are innervated by the (reproducible, time locked waves) from the late mylohyoid nerve (7), which originates from the inferior dental nerve immediately before it asynchronous activity. Horizontal calibration is 2 ms. enters the mandibular canal. The dashed line (left) represents the coronoid process of the Vertical calibration is 2 m V in (A) (except in the bottom mandible. Sl and S2 (right) indicate the approximate sites ofnerve excitation with records, where the calibration is 0 5 mV), and 5 m V in percutaneous needle stimulation (Sl) and transcranial stimulation (S2). (B). 46 Cruccu, Inghilleni, Berardelli, Pauletti, Casali, Coratti, Frisardi, Thompson, Manfredi

Figure 5 After-discharges J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.1.43 on 1 January 1994. Downloaded from evoked by stimulation of A A the masticatory nerves. (A) In the right masseter ofpatient 2, the M wave is followed by after-activity consisting of highfrequency discharges ofsynchronised multiple motor unit potentials lasting 250 ms. (B) In the left temporalis B ofpatient 1, the M wave is followed by late activity consisting ofasynchronous motor unit potentials, B I starting after a briefpause ofelectrical silence. Calibration is 50 msll mV in (A1l), 10 ms/2 mVin vrw-.-w-,v -4 vvwv v'v Y, (A2), and 5 msll mVin 'wI.rv- fv r V' (B). C1 electrical field is thought to excite the trigeminal motor nerve fibres intracranially, near their exit from the skull9 10 (fig 3). In patient R 1, the responses from the masseter were normal and symmetrical. Because a large stimu- I IbI kIEikI I VIILL Elill lus artifact and signals from facial muscles ! I l 1 L obscured surface recordings, responses from Mlfy ilmw * lfwim PI pi IIqfn the temporalis muscle could only be examined by needle recording: in this condition, the latency of the response in the left tempo- C2 ralis (4 ms) was markedly longer than that in the right temporalis (2-5 ms) and longer than R the response simultaneously recorded in the left masseter (2-2 ms). In our experience the and masseter to stimula- responses d.-- temporalis p wlll. L tion at different sites along the trigeminal nerve pathway differ minimally in latency 8 Figure 6 J7awjerk and silent periods. Thejawjerk (A) (mean 0-2 ms).7 In patient 2, the response in and the silent period (B) evoked by taps to the chin are the right masseter was markedly delayed (3'5 absent in the left temporalis muscle ofpatient 1. The early ms on the right, 2 ms on the left) and dis- and late silent periods evoked by electrical stimulation of the mentalis nerve (C) are absent in the right masseter of persed (fig 4). patient 2, after stimulation ofthe right side (Cl) and left We concluded that both patients had side nerve (C2). Calibrations are 5 ms/05 mVin (A), marked slowing of conduction of the mastica- and 20 ms/0I5 mV in (B) and (C). tory nerves, more pronounced between the two sites of stimulation (fig 3), without a http://jnnp.bmj.com/ reduction in amplitude of the M waves evoked by distal stimulation. Stimulation of the masticatory nerve fibres TRIGEMINAL REFLEXES in both patients also induced autoexcitation The jaw jerk evoked by tapping the patient's with involuntary discharges in the target mus- chin with a triggered hammer'011 was cle. In patient 2, the M wave in the right mas- recorded through surface electrodes from the

seter was often followed by an after-activity masseter and temporalis muscles bilaterally. on September 29, 2021 by guest. Protected copyright. consisting of high-frequency (110 Hz) dis- In patient 1 the jaw jerk was normal and sym- charges of synchronised multiple motor unit metrical in the masseter, whereas it was potentials lasting about 250 ms (fig 5A). In absent in the left temporalis and normal in patient 1, the M wave in the left temporalis the right (right latency 7 ms) (fig 6A). In was occasionally followed, after a 10 ms patient 2 the jaw jerk was absent in the right pause of electrical silence, by a late-activity of masseter and normal in the left (left latency 7 asynchronous motor unit potentials, lasting ms). about 30 ms (fig 5B). The silent period of jaw closing muscles The presence of ephaptic responses in the after the chin tap and the early and late silent masseter muscle after stimulation of the deep periods after electrical stimulation of the temporal nerves or in the temporalis muscle mental or infraorbital nerve (SPI and SP2 of after stimulation of the masseteric nerve the masseter inhibitory reflex, also called could not be properly evaluated, because exteroceptive suppression) were recorded stimulation of the masticatory nerves in the during maximum clenching of the teeth.'01' infratemporal fossa is not sufficiently selec- Although both patients had some difficulty in tive. Transcutaneous electrical stimulation of producing strong and steady contractions the masseter muscle belly, however, did not without interference by involuntary move- evoke ephaptic responses in the temporalis ments, the silent periods were apparently nor- and vice versa, whereas it often triggered pro- mal in the unaffected muscles and even in the longed spasms in the affected muscles. affected muscles, when these were spasm- Pathophysiology ofhemimasticatory spasm 47

Table Reports ofpatients with hemimasticatory spasm demonstrated by EMG examination J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.1.43 on 1 January 1994. Downloaded from Sex and Facial hemiatrophy age ofonset and connective tissue Proposed site oflesion Study (years) Muscles EMG Trigeminalfunction disease and mechanism Kaufman, F25 L masseter Mild chronic denervation. Clinically normal Facial hemiatrophy Peripheral nerve 19803 Bursts of single or multiple unit 200 Hz discharges. Recruitment of synchronised units leading to tonic contraction Lapresle and F15 R masseter Polyphasic potentials of Clinically normal Multiple localized Muscle and CNS Desi, 19826 variable amplitude. No scleroderma. Antinuclear clearcut myogenic or antibodies. Facial neurogenic signs. Cramp hemiatrophy recorded from the masseter Thompson F57 L temporalis 20-100 ms bursts of high Previous episodes of Absent Peripheral nerve. and Carrol, and masseter frequency discharges. Crescendo facial numbness. Inflammatory 19834 of multiple unit bursts leading to Normal clinical disorder or prolonged spasm (200 ms-i s) examination and vascular compression blink reflexes Thompson F3 1 R jaw closers 20-40 ms bursts of motor Normal facial sensation. Morphoea. Facial Peripheral nerve et al, 1986"4 facial muscles unit discharges Normal blink reflexes. hemiatrophy Delayed jaw-jerk Parisi et al, F38 R masseter Rapid discharges ofhigh Clinically normal Multiple localized Sympathetic cervical 19875 frequency motor units scleroderma. Facial ganglia (or peripheral hemiatrophy nerve, or nucleus) Yoshii et al, M44 L masseter and Synchronised spasm discharges Clinically normal ND ND 1989'5 pterygoids lasting for 1-2 s Patient 1 by F20 R masseter and 10-200 ms bursts ofhigh Normal clinical Absent Motor root or Auger et al, temporalis frequency discharges. Multiple examination, blink reflexes nucleus 1992"6 unit activity during prolonged and jaw-jerk. Efferent spasms lasting up to 1 minute block of silent periods Patient 2 by F20 R temporalis and As above As above Absent Motor root or Auger et al, masseter nucleus 199216 Patient 1 M18 L temporalis 50-100 ms bursts of high Normal clinical Facial hemiatrophy Peripheral muscle- (this report) frequency multiple unit discharges. examination and blink nerve. Focal Recruitment of synchronised reflexes. Absent jaw jerk. demyelination by motor units leading to 60 Hz tonic Efferent block of silent compression activity lasting up to 2 minutes periods Patient 2 F44 R jaw closers As above with 70 Hz tonic As above Morphoea. Antinuclear Peripheral muscle- (this report) activity lasting up to 30 s and smooth muscle nerve. Focal antibodies. Facial demyelination by hemiatrophy compression ND = No data. free. During the spasms, the affected mus- sclerosis, and unilateral dystonia of the jaw.'4 cles-that is, the left temporalis in patient 1 Clinical ambiguities can easily be resolved and the right jaw closers in patient 2- by the characteristic electromyographic find- showed an efferent block: in other words, the ings. In this review, we have therefore col- silent periods were completely absent in these lected only the reports of unilateral muscles and normal in the contralateral mus- involuntary movements of the jaw corrobo- cles, regardless of the side of stimulation (fig rated by an EMG description and corre- 6). Electrical stimulation of the mentalis and sponding to our definition of HMS. As the http://jnnp.bmj.com/ infraorbital nerve on the affected side evoked table shows, these criteria were met in 10 normal silent periods in the contralateral patients only. Two of them are described in muscle (patient 1, right temporalis latency 12 the present report, seven were also diagnosed ms, patient 2, left masseter latency l11 ms); as having unilateral masticatory or masseter this finding was important because it allowed spasm,35 14-16 and one had a different diagno- us to exclude damage to or interference on sis.6

the input from cutaneous afferents. Hemimasticatory spasm more commonly on September 29, 2021 by guest. Protected copyright. Electrical stimulation of the right and left presents in women (4:1) in the third and supraorbital nerves consistently evoked nor- fourth decade (age range 15-57 years). It is mal and symmetrical early and late blink often, although not always, associated with reflexes in the orbicularis oculi muscles11 12 in facial hemiatrophy. Localised scleroderma both patients. and clinical or laboratory signs of connective tissue disease are relatively frequent. Neurological examination is usually normal- Discussion except for the spasm-and facial sensation is DIAGNOSIS OF HEMIMASTICATORY SPASM always normal. The spasm involves one or The occurrence of 'masticatory spasms', in more jaw closing muscles on one side, most association with facial hemiatrophy, described frequently the masseter. Jaw openers are first in the last century' 13 has since been doc- never affected. Ipsilateral involvement of umented in several clinical reports. The clini- other cranial muscles has been described in cal diagnosis is difficult, however, even if the two patients.46 spasm is unilateral, because involuntary Clinically, the involuntary movement con- movements of the jaw are present in a variety sists of brief twitches (resembling those of of conditions, including mechanical or hemifacial spasm) and prolonged spasms inflammatory disorders of the mandible and (lasting a few seconds to several minutes, and temporomandibular joint, cephalic tetanus, resembling cramps) or prolonged spasms focal motor epilepsy, tonic spasms of multiple alone. The spasms are intensely painful, vio- 48 Cnuccu, Inghiaeci, Berardelli, Paulett, Casali, Coramt, Fnsardi, Thompson, Manfredi

lent, and sometimes of sudden onset: during ish the reflex response.'0 In contrast, medium J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.1.43 on 1 January 1994. Downloaded from a spasm patients may bite their tongue,'4 dis- size exteroceptive fibres (AD3) were spared, as locate their temporomandibular joint,3 or shown by the general finding of a normal tac- even break teeth. The involuntary movement tile sensation and normal latency of extero- may be evoked by yawning, speaking, closing ceptive reflexes when tested (table). the mouth, chewing, or other voluntary Temperature and pain sensation are invari- movements of the mouth and jaw, as well as ably normal in patients with HMS. One by electrical shocks delivered to the muscle explanation could be that the disease prefer- belly or the facial skin. entially affects the largest nerve fibres. EMG recordings show no denervation Alternatively, it only affects muscle nerves, in potentials and in most cases the motor unit which the function of small afferent fibres is potentials are normal. The spontaneous activ- difficult to assess. ity resembles that of hemifacial spasm (short Muscle-nerve damage would explain not bursts of 100-200 Hz discharges of one or only why patients with HMS have no sensory few synchronised motor unit potentials), and disturbance, but also why they often have that of muscle cramps (tonic 50-70 Hz dis- spasms in one or two jaw closers only (table), charges of a compound potential comprising yet never have spasms in the jaw openers, many synchronised motor unit potentials). muscles that are- innervated by a branch of The duration of these paroxysms, as well as the inferior dental nerve (fig 3). These obser- the number of motor units involved, varies vations argue against damage to the motor widely from case to case and even from time root or the intracranial portion of the to time. The hallmarks are high frequency mandibular nerve, where the motor fascicles discharges and the recruitment of synchro- are closely grouped,'8 but favour damage to nised motor unit potentials, features that are the individual muscle nerves that pass easily identified and only present (in mastica- through the infratemporal fossa. This was so tory muscles) in this rare condition. in our two patients, in whom the slowing of The finding of an absent or delayed jaw conduction was prominent in the infratempo- jerk in the affected muscles is also common, ral fossa (figs 3 and 4). yet not diagnostic, as jaw jerk asymmetries The mechanism of facial paroxysmal invol- can also be found in multiple sclerosis and untary activity has been dicussed by even in temporomandibular joint dysfunc- Kaufnan3 and by Thompson and Carroll,4 tion. Almost unique to HMS is the finding of who emphasised the close similarity between an efferent block of the masseter silent peri- hemimasticatory and hemifacial spasm. In ods: during the spasm, no inhibition of the our view, our patients and others reported affected muscles can be exerted reflexly by also had cramp-like activity. As in cramps, in any sensor input, regardless of the site of HMS the muscle may be hypertrophied, the stimulation. This unusual feature may occur contractions may also be prolonged, decid- in a few other conditions, one of which is edly painful, and sometimes intense enough cephalic tetanus. Nevertheless, a complete to produce severe damage to tongue, teeth, efferent block to the muscles on one side of and temporomandibular joint. In the EMG, the face or even in one muscle alone is excep- these prolonged spasms nicely fit the descrip- tional.'7 tion of cramps-that is, irregular motor unit To attenuate the spasm, carbamazepine, discharges that progressively increase, leading http://jnnp.bmj.com/ phenytoin, haloperidol, clonazepam, and to recruitment of a large part of the muscle diazepam have been tried with varying and to synchronous discharges at rates from results, the best being obtained with carba- 40 to 60 Hz.12 Common to hemifacial spasm mazepine. Other treatments tried include and cramps, however, is ectopic excitation. mandibular anaesthesia, 'rubbing' of the This may be responsible for the high fre- mandibular nerve,3 masseter myotomy,'4 sur- quency (100-200 Hz) discharges, synchroni-

gical lesions of the motor root,416 and local sation of the whole muscle or more muscles, on September 29, 2021 by guest. Protected copyright. injection with botulinum A toxin.'6 and after-activity. Synchronisation is tenta- tively explained by lateral spread of dis- PATHOPHYSIOLOGY OF HEMIMASTICATORY charges to adjacent nerve fibres,'9 20 leading to SPASM local circuits of re-excitation. After-activity We have demonstrated a peripheral motor consists of paroxysmal discharges that may nerve lesion, as originally proposed.' Nerve follow a voluntary orthodromic contraction or conduction studies showed slowing of con- antidromic impulses,2122 and is attributed to duction in the extracranial course of the mas- autoexcitation of the same axons after the ticatory nerve fibres, without a reduction in passage of an impulse. amplitude of the M waves or obvious EMG In our two patients, we observed the syn- signs of chronic denervation. Biopsy speci- chronisation of the whole or a large part of mens of the affected temporal muscle the muscle (fig 2C); autoexcitation was appeared histologically normal in both demonstrated by recording after-discharges patients. These findings indicate demyelina- following the direct motor response evoked tion with sparing of the axon. by stimulation of the masticatory nerves The jaw jerk was absent or delayed in sev- (fig 5). These findings lend support to the eral patients (table). This indicates damage to hypothesis that the spontaneous activity the large diameter (Aa) afferent fibres from 'arises' in a demyelinated peripheral nerve.222' muscle spindles; a demyelination of even a In our patients, unlike those with hemifacial few of the jaw jerk afferents could easily abol- spasm, we were unable to record ephaptic Pathophysiology ofhemimasticatory spasm 49

responses-that is, 'delayed' responses in one atrophy.'5 Facial hemiatrophy often involves J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.1.43 on 1 January 1994. Downloaded from muscle after stimulation of the nerve fibres not only the skin and subcutaneous tissues, directed to another muscle. Cross-talking but also deep tissues such as fat, muscles, lig- between nerve fibres for different muscles is aments, cartilage and bone.'4'5 Deep tissue likely to take place in hemifacial spasm changes might lead to stretching, angulation, because the supposed site of lesion is at a or compression followed by focal demyelina- proximal point, where nerve fascicles are tion, of the masticatory nerves. Owing to closely grouped. In our patients with HMS, their anatomical relations (fig 3), these nerves the nerve was damaged extracranially, proba- may suffer entrapment, in particular the bly between the infratemporal fossa and the nerves supplying the masseter and temporal distal nerve branches, where fascicles are muscles, which are most frequently affected more widely separated by perineurial tissue. in HMS. The temporomasseteric nerve runs in a confined space between the lateral ptery- ETIOLOGIC CONSIDERATIONS goid muscle and the unyielding surface of the Although still debatable, the primary cause of skull; the deep temporal nerves and the mas- hemifacial spasm is thought to be compres- seteric nerve turn sharply around bone crests, sion-possibly inducing demyelination- and finally run a narrow course between the of the nerve near its emergence from the bone and their own target muscles. pons.2425 Cramps occur in a number of different conditions associated with neuropathy.2023 Addendum Evidence has been found of segmental Since this paper was submitted, Auger and demyelination and axonal degeneration.2026 colleagues (Neurology 1992;42:2263-6) have The cause of the trigeminal neuropathy reported a third patient with HMS, similar to that produces masticatory spasms is the two patients described in the table. unknown. Some patients with HMS also have clinical or laboratory signs of immune connective tissue diseases; these may cause mononeu- ropathies through several mechanisms. Although connective tissue diseases are often associated with trigeminal damage, these neu- 1 Gowers WR. A manual ofdiseases of the nervous system, 2nd ropathies nearly always produce sensory ed. Philadelphia: Blakiston, 1897;2:221-4. 2 Wartenberg R. Progressive facial hemiatrophy. Arch impairment alone.27 Only three of the 103 Neurol Psychiatry 1945;54:75-96. patients described by Lecky et aP8 and Hagen 3 Kaufman MD. Masticatory spasm in facial hemiatrophy. Ann Neurol 1980;7:585-7. et aP9 also had signs of a motor trigeminal 4 Thompson PD, Carroll WM. Hemimasticatory spasm: a impairment. In addition, blink reflex as well peripheral paroxysmal cranial neuropathy? J Neurol Neurosurg Psychiatry 1983;46:274-6. as histologic studies indicated axonal loss 5 Parisi 1, Valente G, Dell'Anna C, Mariorenzi R, et al. A rather than demyelination25; in our patients, case of facial hemiatrophy associated with linear sclero- derna and homolateral masseter spasm. ItalJ Neurol Sci the absence of motor unit changes in the 1987;8:63-5. EMG and the normal muscle biopsy findings 6 Lapresle J, Desi M. Sclerodermie avec hemiatrophie faciale progressive et atrophie croisee de l'hemicorps. ruled out axonal loss. Rev Neurol (Paris) 1982;138:815-25. Motor 7 Cruccu G. Intracranial stimulation of the trigeniinal nerve trigeminal neuropathy is always an http://jnnp.bmj.com/ in man. I. Direct motor responses. J Neurol Neurosurg exceptional finding, probably arising only Psychiaty 1986;49:411-8. when focal pressure causes damage restricted 8 Cruccu G. Central and peripheral conduction in the human trigeminal motor system, In van Steenberghe D, to the muscle nerve. As studies of experimen- De Laat A, eds. Electromyography ofjaw reflexes in man. tally induced pressure neuropathies demon- Leuven: University Press, 1989:3-17. 9 Cruccu G, Berardelli A, Inghilleri M, Manfredi M. strate, compression can and does lead to Functional organization of the trigeminal motor system ectopic activity.30"3 in man. Brain 1989;112:1333-50. 10 Cruccu G, Inghilleri M, Fraioli B, Guidetti B, et al. The possibility of cross-compression by an Neurophysiologic assessment of trigeminal function arterial loop-as has been proposed for hemi- after surgery for trigeminal neuralgia. Neurology 1987; on September 29, 2021 by guest. Protected copyright. 37:631-8. facial spasm and trigeminal neuralgia-seems 11 Ongerboer de Visser BW. Anatomical and functional unlikely. Despite frequent surgical documen- organization of reflexes involving the trigeminal system in man: jaw reflex, blink reflex, comeal reflex and tation of cross-compressions between vessels exteroceptive suppression. Adv Neurol 1983;39:729-38. and cranial nerves in the posterior fossa,'2" 12 Kimura J. Electrodiagnosis in diseases of nerve and muscle: principles and practice, 2nd edn. Philadelphia: FA Davis, none of these reports mentions a concurrent 1989. motor disturbance resembling masticatory 13 Sachs B. Progressive facial hemiatrophy with some unusual symptoms. Med Rec 1890;37:292-7. spasm. Furthermore, in two patients with 14 Thompson PD, Obeso JA, Delgado G, Marsden CD. HMS, surgical exploration failed to disclose Focal dystonia of the jaw and the differential diagnosis of unilateral jaw and masticatory spasm. J Neurol anomalous vascular contacts along the Neurosurg Psychiatry 1986;49:651-6. intracranial root'6 or the distal course of the 15 Yoshii K, Seki Y, Aiba T. A case of unilateral masticatory spasm without hemifacial atrophy. No To Shinkei nerve.3 1989;41:343-6. In our opinion, when searching for the 16 Auger RG, Litchy WJ, Ahlskog JE, Cascino TL. Hemimasticatory spasm: clinical and electrophysiologic cause of HMS we should consider its fre- observations. Neurology 1992;42(suppl 3):352P. quent association with facial hemiatrophy- 17 Ongerboer de Visser BW, Cruccu G, Manfredi M, Koelman JHTM. Effects of brainstem lesions on the surely not a chance event. In the patients masseter inhibitory reflex. Functional mechanisms of reported not to have facial hemiatrophy, this reflex pathways. Brain 1990;113:781-92. 18 Pennisi E, Cruccu G, Manfredi M, Palladini G. might not be clinically apparent. The process Histometric study of myelinated fibers in the human may take 10 years to develop34 and muscle trigeminal nerve. J Neurol Sci 1991;105:22-8. 19 Nielsen VK. Pathophysiology of hemifacial spasm: II. spasms are sometimes prodromic to the Lateral spread of the supraorbital nerve reflex. Neurology development of the superficial tissue 1984;34:427-3 1. 50 Cruccu, Inghilleri, Berardelli, Pauleri, Casali, Coratti, Frisardi, Thompson, Manfredi

20 Thompson PD. Stiff people. In Fahn S, Marsden CD, 28 Lecky BR, Hughes RAC, Murray NM. Trigeminal sen- J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.1.43 on 1 January 1994. Downloaded from eds. Movement disorders 3. London: Butterworths, 1993: sory neuropathy: a study of 22 cases. Brain 1987; 367-99. 110:1463-85. 21 Auger RG. Hemnifacial spasm: clinical and electrophysio- 29 Hagen NA, Stevens JC, Michet CJ Jr. Trigeminal sensory logic observations. Neurology 1979;29: 1261-72. neuropathy associated with connective tissue diseases. 22 Nielsen VK. Pathophysiology of hemifacial spasm: I. Neurology 1990;40:891-6. Ephaptic transmission and ectopic excitation. Neurology 30 Burchiel KJ. Abnormal impulse generation in focally 1984;34:418-26. demyelinated trigeminal roots. Neurosurg 1980;53: 23 Rowland LP. Cramps, spasm and muscle stiffness. Rev 674-83. Neurol (Paris) 1985;141:261-73. 31 Rasminsky M. Ectopic impulse generation in pathologic 24 Auger RG, Piepgras DG, Laws ER, Miller RH. nerve fibers. In Dyck PJ, Thomas PK, Lambert EH, Microvascular decompression of the facial nerve for Bunge R, eds. Peripheral neuropathy, vol. 2, 2nd edn. hemifacial spasm: Clinical and electrophysiologic obser- Philadelphia: WB Saunders, 1984:911-8. vations. Neurology 1981;31:346-50. 32 Jannetta PJ. Microsurgery of cranial nerve cross-compres- 25 Nielsen VK, Jannetta PJ. Pathophysiology of hemifacial sion. Clin Neurosurg 1979;26:607-15. spasm: III. Effects of facial nerve decompression. 33 Jannetta PJ. Surgical treatment: Microcascular decom- Neurology 1984;34:891-7. pression. In Fromm GH, Sessle BJ, eds. Tigeminal neu- 26 Welch LK, Appenzeller 0, Bicknell JM. Peripheral neuralgia. Boston: Butterworth-Heinemann, 1991:145-57. ropathy with myokimia, sustained muscular contraction 34 Bruyn GW, Bruyn RPM. Hemiatrophies and hemihyper- and continuous motor unit activity. Neurology 1972;22: trophies. In de Jong JMBV, ed. Handbook ofclinical neu- 161-9. rology, vol. 15. Amsterdam: Elsevier, 1991:475-86. 27 Hughes RAC. Diseases of the fifth cranial nerve. In Dyck 35 Delaire J, Lumineau JP, Mercier J, Plenier V. Syndrome PJ, et al, eds. Peripheral Neuropathy, 3rd edn. de Romberg. Hemi-atrophie faciale progressive. Rev Philadelphia: WB Saunders, 1993:801-17. Stomatol ChirMaxilofac 1983;84:313-321. http://jnnp.bmj.com/ on September 29, 2021 by guest. Protected copyright.