Relationship Between Cochleovestibular Disorders in Hemifacial Spasm and Neurovascular Compression

Transcription

1 The Laryngoscope Lippincott Williams & Wilkins, Inc., Philadelphia The American Laryngological, Rhinological and Otological Society, Inc. Relationship Between Cochleovestibular Disorders in Hemifacial Spasm and Neurovascular Compression Hung Thai Van, MD; Olivier Deguine, PhD; Marie Jose Esteve-Fraysse, MD; Alain Bonafe, PhD; Bernard Fraysse, PhD Objective: To investigate the evolution of cochleovestibular symptoms before, during, and after microvascular decompression 0) of the facial nerve in hemifacial spasm. Study DesZgx Prospective study in patients with hemifacial spasm. Among our 13 patients who underwent MVD of the facial nerve from 1995 to 1997,6 had associated cochleovestibular disorders confirmed by neurotologic tests. Results: In four of these patients, a concomitant compression of the eighth and facial nerves was found at surgery. Preoperative magnetic resonance angiography studies had shown three cases of this double neurovascular compression. Intraoperative auditory brainstem response monitoring showed that interposition of Teflon between vessel and facial nerve was highly critical to the auditory function. Auditory brainstem response monitoring was used to guide the surgeon during this critical phase. Surgery improved at least one cochleovestibular symptom in each patient. Concbsions: The authors propose two pathophysiologic hypotheses. First, the concomitant facial and cochleovestibular symptoms may be due to a hyperactivity of both the facial and vestibular nuclei. According to theories about cryptogenic hemifacial spasm, the origin of this hyperactivity could be an ectopic excitation focus. However, the two nerves may have different sites of ectopic excitation. According to the second hypothesis, a pulsatile compression of the facial nerve may be transmitted to the eighth nerve. This could take place even if only the facial nerve is in contact with a vascular loop. Laryngoscope, 109: ,1999 Presented at GEM0 (Groupe #Etudes MBdecine et Otologie), Courchevel, France, February 6, From the Departments of Ear, Nose, and Throat Medicine (H.T.v., o.d., M.J.E.-F., B.F. 1 and Neuroradiology (A.B.), Purpan Hospital, Toulouse, France. Editor s Note: This Manuscript was accepted for publication November 18, Send Correspondence to Hung Thai Van, MD, Consultation d ORL, Service du Pr. B. Fraysse, CHU Purpan, Place du Docteur Baylac, Toulouse Cedex 31059, France. INTRODUCTION Relationships between hemifacial spasm (HFS) and cochleovestibular symptoms have been reported occasionally in the literature. In a population of 143 HFS patients, Moller and Mollerl found 33 cases (23%) of significant hearing loss on the spasm side. These patients had either sharp notches of auditory pure-tone thresholds in the midfrequency ranges or low-frequency hearing loss, as in Meniere disease, but without any fluctuations in hearing. The authors consider these audiometric abnormalities a sign of auditory nerve irritation by the vessel causing the spasm. Conversely, Mo11er2 noted that, of 41 patients who had undergone a microvascular decompression of the vestibular nerve, 8 had facial twitches. The same team3 considered that progressive sensorineural hearing loss in the high-frequency ranges, as well as asymmetry of hearing thresholds, in a patient consulting for vertigo was possibly related to a vascular compression of the eighth nerve. According to them, tinnitus may also be linked to cochlear nerve compression. Abnormalities of preoperative auditory brainstem response (ABR) have been described in HFS patient^,^ mainly, significant lengthening of peak I11 latencies and even more of peak V latencies on the spasm side. This implies that the vascular abnormalities causing HFS are not limited to the facial nerve but may extend to adjacent cranial nerves. We present our findings in six patients who underwent microvascular decompression (MVD) of the facial nerve in the Department of Ear, Nose, and Throat (ENTI Medicine at Purpan Hospital, Toulouse, France, in the last 2 years. The surgery was conducted under intraoperative monitoring of facial and auditory functions. These six patients were studied specifically because they had cochleovestibular symptoms and severe HFS simultaneously. Each of them underwent a preoperative audiovestibular evaluation including pure-tone audiogram, stapedius reflex test, ABR recording, and vestibular investigation by videonystagmography (VNG). This report investigates the evolution of the cochleovestibular func- 741

2 tion before, during, and after MVD of the facial nerve in these HFS patients. The possibility of a concomitant compression of the cochleovestibular and the facial nerves is studied using audiovestibular tests, radiologic data, and intraoperative findings. MATERIALS AND METHODS Patient Selection The period of this prospective study was from the summer of 1995 to the end of Patients were recruited fiom the population of 13 patients who underwent MVD for HF S during this period in Purpan Hospital. Of these 13 patients, 8 were men, 5 were women, and the average age was 55 years. They had had HFS for an average of 10 years. All patients who satisfied at least one of the following criteria were selected significant sensorineural hearing loss (HL) and vertigo assessed by positive vestibular test results. The level of inclusion for HL was set at 35 db HL or greater on at least one frequency between 250 and 8,000 Hz on the spasm side with an asymmetry of at least 15 db HL compared with the unaffected side. Patients with any fluctuations in hearing or a history of acoustic trauma were excluded. To assess vertigo, vestibular function was explored by VNG, including saccadic testing, light-tracking tests, optokinetic testing, and measurement of spontaneous and positional nystagmus, and study of the vestibulo-ocular reflex (VOR) by kinetic stimulation on rotatory seat and by bithermic caloric stimulation at 30 C and 44 C. Objective criteria were a caloric weakness greater than 20% on the spasm side and/or a directional preponderance higher than 2 degrees per second (or 25%). Patients with tinnitus were included only if they also had either sensorineural hearing loss or vertigo. Neuroradiologic Investigation Preoperative neuroradiologic diagnosis was performed using magnetic resonance imaging (MRI) and magnetic resonance angiography (MRA). All patients had T1-weighted images, gadolinium-enhanced T1-weighted images, and T2-weighted images to ensure that HFS was not caused by a tumor (intrapetrous cholesteatoma, epidermoid tumor, meningioma, facial nerve neurinoma). The presumptive diagnosis of a vascular compression syndrome was supported by Constructive Interference in Steady State (CISS) studies. CISS studies visualize arteries, cochlear nerve, superior and inferior vestibular nerves, and facial nerve in the internal auditory canal (IAC) and in the cerebellopontine angle up to their root outlet zone. Vascular loops were identified by three-dimensional (3-D) MRA studies using time-of-flight (TOF) techniques with multiplanar reconstructions (MPR) and maximum intensity projection (MIP) imagess All MRI and MRA investigations were performed on Magnetom Vision 1,5 Tesla (Siemens, Erlangen, Germany). Surgical Technique The retrosigmoid approach was employed for the MVD. All patients were operated on under general anesthesia in the dorsaldecubitus position. With this technique, despite a limited craniotomy, a good visualization of the anatomic and functional positions of the seventh and the eighth nerves is obtained. Moreover, the cerebellum retracts spontaneously without the aid of retractors: once the dura has been opened and the cerebrospinal fluid (CSF) has flown out. The arachnoid was incised exposing the facial nerve. A dissection between the flocculus and the facial nerve was performed up to its root outlet zone. The facial nerve root was 742 separated from the artery causing the conflict by changing the leading axis of the loop and by inserting a piece of Teflon between the loop and the nerve. The veins at the surface of the nerve s root were at least partially elevated after dissection of the surrounding arachnoid. When an arterial loop was found to be in conflict with both the seventh and the eighth nerves in their cisternal route or at the porus acusticus, the loop was elevated and Teflon was inserted between it and the facial nerve. Intraoperative Neurophysiologic Monitoring The auditory function was monitored by recording ABR, starting after craniotomy. The auditory stimulus consisted of filtered clicks delivered through ear-inserts (Nicolet, Madison, WI). The intensity of the stimulus was 100 db HL, allowing an average effective stimulation of 80 db HL. With a number of repetitions per average of 700 to 1,000 and a stimulation rate of 20 Hz, an ABR was obtained every 40 to 50 seconds. Sensitivity of the amplifier was set at 56 to 75 mv. We used a 160- to 1600-Hz pass-band filter (Centor system, Racia-Alvar, France). Peak latencies were lengthened by 2 milliseconds because of the length of silicon joints between the receivers and the earinserts. Intraoperative ABRa were compared with preoperative ones and with the responses obtained at the beginning of the surgery after anesthetic induction. Audiovestibular Follow-up All patients were followed during at least 6 to 12 months after surgery. At day 1 and day 10 after surgery, patients were assessed for their clinical status. Two, 6, and 12 months later, they underwent the following tests, depending on their clinical symptoms: pure-tone audiogram, ABR recording, and VNG. RESULTS All patients (four women and two men, aged 51 to 68 years) had severe HFS, leading to a handicap of more than 6 hours duration per day, psychological and social discomfort and, for one of them, sleeping disorders. They had been given treatment with botulinum toxin (Allergan; on average, 25 units per session) for 1 to 4 years before surgery. Compared with the total HFS population operated on in our department in the same time span, patients who presented associated cochleovestibular symptoms were slightly older (average age, 62 y), more were women, and they had had HFS for a similar time (average duration, 10 y). However, no significant difference in patient profile can be derived from such a small population. Cochleovestibular symptoms were as follows: 1) four cases of vertigo, 2) five cases of tinnitus, and 3) two cases of sensorineural HL of more than 35 db on the spasm side. Two patients had hearing loss and tinnitus, one had vertigo alone, and three had vertigo with tinnitus. The four cases of vertigo had the following different characteristics: 1) one case of persistent positional vertigo, despite liberatory maneuvers, 2) One case of continuous unsteadiness with a dramatically decreased caloric excitability on the spasm side, 3) one case of unsteadiness with a significant caloric weakness on the affected side, and 4) one case of positional vertigo spells with a major caloric weakness on the side of the vascular compression and a directional preponderance on the opposite side.

3 Five patients had on the spasm side high-pitched tinnitus described as unpleasant continuous whistle, leading to sleeping disorders in one patient. The bone conduction thresholds of one of the two patients with HL were 35 db HL at 2,000 Hz and 60 db HL at 4,000 Hz; the thresholds were, respectively, 20 db HL and 35 db HL on the nonspasm side. Preoperative ABR suggested a retrococdear origin for the HL on the spasm side: peak I pattern was abnormal, and there was a significant interaural peak V latencies shift (0.36 ms). Pure-tone audiogram of the other patient with HL showed, on the affected side, a notch of 60 db HL at 4,000 Hz without any history of acoustic trauma. Air and bone conduction thresholds were normal at all other fiequencies. His ABR was clearly retrocochlear on the spasm side with a I-V intenvave latency of 4.62 milliseconds and a intenvave latency of 2.67 milliseconds. CASE REPORTS Case 1 A 59-year-old man had a 19-year history of severe HFS on the left side. For 1 year he had been reporting symptoms of an almost continuous whistle in the left ear. Pure-tone audiogram showed an asymmetric, predominantly left-sided high-frequency sensorineural HL. MRA showed a posterior inferior cerebellar artery (PICA) loop crossing the facial nerve root, as well as an anterior inferior cerebellar artery (AICA) loop separating the facial and cochleovestibular nerves in the cerebellopontine cistern. The AICA branch clearly reached the porus acusticus on MPR in coronal oblique section (Fig. 1). Preoperative ABR revealed abnormalities in the waveforms of peaks I and 111, and only peak V was reproducible. At surgery we found a twofold neurovascular compression in the cerebellopontine angle: a large PICA in close contact with the brainstem causing cross-compression of the seventh nerve root, and an AICA between the facial and the cochleovestibular nerves. Once the dura had been opened and the cerebellum had retracted, egress of CSF led to a clear improvement in ABR waveshape allowing the neurophysiologist to distinctly identify peaks I and I11 (Fig. 2). A first decompression was performed by interposing a thick Teflon prosthesis between the PICA and the facial nerve root. Another piece of Teflon was then inserted between the branch of the AICA and the compressed part of the facial and cochleovestibular nerves. At the end of surgery ABR showed normal waveforms and normal and I-V delays. In the days following the MVD the patient was free of symptoms except his tinnitus. Twelve months after the surgery he was still relieved of HFS. Audiologic evaluation confirmed normal ABR and revealed an improvement in audiometric bone conduction thresholds of +10 db HL at 250 Hz, +20 db HL at 500 Hz and 1,000 Hz, and + 15 db HL at 4,000 Hz. Tinnitus persisted. Case 2 A 63-year-old woman presented with a 7-year history of left-sided HFS. She was also reporting a symptom of a highpitched, constant tinnitus on the spasm side, as well as continuous unsteadiness that had started at the same time as HFS. This unsteadiness was complicated by vertiginous spells described as motion intolerance precipitated by head movement. Preoperative electrophysiologic testing demonstrated an involvement of the left-sided cochleovestibular function. Pure-tone audiogram was normal, but ABR showed an interaural I-V interval shift of 0.21 milliseconds at 90 db HL and of 0.24 milliseconds at 70 db HL. Tympanograms were normal, but ipsilateral Fig. 1. Magnetic resonance images of a patient presenting with left-sided hemifacial spasm, ipsilateral tinnitus, and ipsilateral hearing loss. A. Axial time-of-flight magnetic resonance angiographic section shows a posterior inferior cerebellar artery (PICA) loop cross-compressing the facial nerve root at the brainstem and a second vessel passing through the seventh and eighth nerves in the cerebellopontine angle (maximum intensity projection techniques identify this vessel as a branch of the anterior inferior cerebellar artery [AICA]). B. Coronal time-of-flight magnetic resonance angiographic section shows the double conflict involving the acusticus facial bundle. Note that the AKA loop reaches the porus acusticus. and contralateral stapedius reflexes were absent on the spasm side. Caloric testing had been performed twice showing a left canal paresis of 94%. MRA showed a loop of the PICA crosscompressing the facial nerve in its root entry zone with a left vertebral artery looping. At surgery arachnoid adhesions were found which were covering the cochleovestibular nerve. These adhesions were removed first; this resulted in a traction of a small artery in close proximity to the cochlear nerve. ABR immediately decreased in amplitude, and the wave V latency was significantly prolonged. It was decided to preserve these adhesions rather than to injure the cochlear nerve s vascularization. The conflict was found to occur in the vicinity of the posterior and inferior portion of the seventh nerve root, which was compressed by a large vertebral artery and a loop of the PICA. Interposition of Teflon between the vessels and the facial nerve was followed again by alterations in ABR. Evoked potentials went back to normal afterward and remained stable until the end of the surgery. 743

4 Case 3 A 64-year-old woman reported a 4-year history of left-sided HFS. She also had a left-sided tinnitus and experienced positional vertigo, apparently since a head injury 6 years earlier. Three liberatory maneuvers had been ineffective. Caloric testing demonstrated a left-side caloric weakness of 30%. Pure-tone audiometry and ABR results were normal. On preoperative neuroradiologic scans a branch of the AICA was found to crosscompress at right angle the facial nerve root, while a loop of the AICA was in contact with the facial nerve and the superior vestibular nerve at the porus acusticus. Surgery confirmed a twofold neurovascular compression syndrome caused by the same branch of the AICA looping twice. This vascular loop was seen, adjacent to the porus acusticus, compressing the facial and superior vestibular nerves. Behind and under the facial nerve the loop developed a second S pressing on the facial nerve root. An interposition of Teflon prosthesis was performed between the artery and the facial nerve both at the pons acusticus and at the root. ABR recordings showed alterations at the following surgical stages: arachnoid dissection over the seventh and eighth nerves (increased latency wave V), traction of the cochlear nerve root (increased latency waves I1 and III), and liberation of the facial nerve at the root entry zone (increased latency peaks I-V, without any modifications of and I-V delays). After manipulation of the AICA loop near the pons acusticus, peaks I-V disappeared completely. They reappeared after insinuating the piece of Teflon. The patient achieved complete relief of her spasm the day following surgery. She was free of tinnitus too, but she rapidly developed vestibular troubles. Neurotologic examination revealed an obvious left-sided vestibular deficit. A videonystagmograph was performed 8 days after surgery showing a left-side canal paresis of 100% and a right-side nystagmus of degree 111 (Alexander classification). Intensive vestibular physiotherapy was undertaken. The last videonystagmograph, performed 2 months after MVD, demonstrated gradual improvement of vestibular function showing a left-side caloric weakness of 50% without any spontaneous nystagmus and a good vestibular compensation. Fig. 2. lntraoperative monitoring of the auditory function by auditory brainstem response (ABR) recording (without 50-Hz filter) during a microvascular decompression of the facial nerve (case 1). Readings are as follows: 8:45 A.M.: ABR before the beginning of surgery; 10:37 A.M.: ABR after cerebellar retraction and egress of cerebrospinal fluid; 11 :05 A.M.: Dissection of arachnoid adhesion recovering the PICA; transitory flattening of ABR peaks; 11 :14 A.M.: Dissection of a first conflict between a large PICA and the facial nerve root at the brainstem; preservation of ABR peaks; 11:24 A.M.: ABR during a spasm; A.M.: Interposition of Teflon between the cochlear nerve and the AICA; alterations in peaks I and 111 waveforms; 11:36 A.M.: Traction on the cochlear nerve root; major alteration in ABR waveshapes; A.M.: Closing the dura; 11 :54 A.M.: ABR waveshapes are normalized; 12:04 A.M.: Normal ABR at the end of surgery. The day after the surgery the patient was completely relieved of both her spasm and her tinnitus. At postoperative neurotologic examinations 2 months and 6 months later, she denied vertigo and disequilibrium. Vestibular tests confirmed that she had recovered normal and symmetric caloric excitability. 744 Case 4 A 51-year-old man before surgery had a constant ear whistle on the spasm side. Pure-tone audiometry revealed an ipsilateral sharp dip at 4,000 Hz, and ABR an abnormal interval. On MRA scans using TOF 3-D techniques, a PICA loop was identified crossing the seventh nerve root. Another vascular loop was found in close contact with the cochleovestibular nerve near the porus acusticus on CISS studies. Intraoperative findings consisted, first, of facial nerve root cross-compression by a branch of PICA. Moreover, a branch of the AICA was coursing through the cochleovestibular and facial nerves in the cerebellopontine cistern. Two implants of Teflon were inserted, one between the seventh nerve s root and the offending branch of PICA and another between the facial nerve and the branch of AICA after lowering the facial nerve. ABR was modified at these two critical stages. At the end of surgery ABRs went back to their initial status. Forty-eight hours after surgery the patient was relieved of spasm. He asserted cessation of tinnitus. This patient refused the audiologic follow-up. Case 5 A 66-year-old woman reported symptoms of a predominantly right-sided, very high-pitched, unpleasant tinnitus and unsteadiness, in addition to a right-sided HFS. She had had, a few months earlier, a paroxysmal vertigo spell. Vestibular testing revealed a remaining right-sided, significant caloric weakness. MRA revealed only a cross-compression by the PICA of the seventh nerve root. At surgery, further adding to the conflict causing

5 the spasm, an arteriole was found passing along the seventh and the eighth nerves in their cisternal route. This arteriole was mobilized away, and we interposed a piece of Teflon between it and the facial nerve to maintain the decompression. ABR was normal at the end of surgery. Ten days after surgery the patient was completely free of facial and cochleovestibular symptoms. Vestibular testing performed 2 months and 6 months later, normal and symmetric caloric excitability was shown. Case 6 A 68-year-old woman with HFS was having a symptom of motion intolerance precipitated by head movements. Vestibular investigations revealed a left-sided canal paresis of 80% on the spasm side and a right-side directional preponderance. MRI studies showed a conflict between the seventh nerve root and the PICA. A branch of the latter separated the seventh and eighth nerves within the IAC up to the fundus. The decompression consisted of mobilizing the PICA near the brainstem and interposing Teflon between the PICA and the facial nerve root. Again, this modified ABR temporarily. The patient was relieved both of spasm and of vertigo. Caloric testing results were strictly normal 2 months and 6 months later. As shown above, the surgery improved at least one of the cochleovestibular symptoms in each one of the 6 HFS patients. The following results were obtained: 1. cessation of vertigo in three of four cases of vertigo, 2. cessations of tinnitus in four of five cases of tinnitus, and 3. hearing improvement in one of two cases of HL. DISCUSSION Usefulness of Neuroradiologic Investigations Preoperative MRI and MRA data appear to correlate well with intraoperative findings. A concomitant compression of the cochleovestibular nerve and the facial nerve has been found at the operative session in four patients (cases 1,3,4, and 5). Three of the four patients (cases 1,3, and 4) had MRA images pointing to a vascular compression of the eighth nerve. In one patient (case 3) the vessel was compressing the superior vestibular nerve at the porus acusticus. In the other two patients (cases 1 and 4) the neurovascular conflict was located in the cisternal route of the acusticus facial bundle. In fact, preoperative neuroradiologic studies had shown that the conflicting loop was also adjacent to the porus acusticus. In only one patient (case 5) did we find a compression of the eighth nerve at surgery even though we had not suspected it preoperatively. This lack of radiologic diagnosis may be caused by MRA spatial resolution limits: Holley et al5 reported having difficulty obtaining a signal for vessels smaller than 1 ml, even with MPR techniques. In our series the only case of disappointing correlation between neuroradiologic and intraoperative data involved an arteriole instead of an artery. Benefits of Intraoperative Auditory Brainstem Response Monitoring Intraoperative auditory function monitoring is obviously of great interest in this type of surgery because of the close proximity of the facial and the cochleovestibular nerves in the cerebellopontine angle. Moreover, the vascular compression may involve both the facial and the cochlear nerves, making the surgical decompression risky for the auditory function: this illustrates the usefulness of monitoring. Most incidents of postoperative deafness following MVD have been reported as cases without auditory function montoring.7 In our patient population, monitored by ABR recordings, no postoperative deafness occurred. In case 2, ABR monitoring probably allowed us to preserve the internal auditory artery during dissection of arachnoidal adhesions. Case 1 shows that intraoperative ABR monitoring can immediately detect an improvement in cochlear nerve conduction. This improvement was confirmed by postoperative audiometric and electrophysiologic investigation. As a general rule, decompression of the facial nerve caused an alteration in the waveform of ABR. In all these patients, ABR monitoring provided guidance to the surgeon when interposing Teflon between the vascular loop and the facial nerve. This surgical stage appears to be the most critical stage for the auditory function. The correlation in this series between intraoperative ABR changes and postsurgical clinical results on cochlear symptoms may indicate 1) that preservation of ABR during surgery is a good indication of preservation of the auditory function after surgery and 2) that improvement of ABR during surgery is a sign of improvement of auditory function, as in case 1. This link between intraoperative ABR changes and postoperative auditory outcomes was also noted by Sindou et a1.8 They observed in their series of 17 patients who underwent MVD for HFS that, although minor and/or transitory changes did not lead to HL, postoperative HL was always associated with major intraoperative alterations of ABR that were not normalized before the end of the surgery. However, an obvious limit of ABRs is that they are not predictive of another function at risk in HFS surgery, the vestibular function. In our case 3, the dissection of the AICA cross-compressing both the facial and the superior vestibular nerves at the porus acusticus was associated with temporary abolition of ABR. This suggested that internal ear vascularization had been momentarily interrupted as a result of labyrinthine arteries stress. Although ABR went back to normal (a sign of improvement of cochlear vascularization), there was no indication that vestibular vascularization had not been permanently impaired. This would explain the postoperative vestibular troubles of the patient. Cochleovestibular Disorders in Patients With Hemifacial Spasm Analysis of the patient population and its symp tom. The main contribution of this study consists in revealing and integrating a group of cochlear and vestibular symptoms in a vascular facial nerve cross-compression syndrome. The possibility of a concomitant compression of both the eighth and the seventh nerves is still being discussed in current literature. When the cochleovestibular symptoms accompanying spasm are limited to tinnitus, doubt remains on the etiology of the tinnitus even if it improves afier MVD of facial nerve.9 Therefore we have not considered tinnitus alone, especially bilateral tinnitus, as a criterion of eighth nerve compression. On the other hand, we 745

6 considered that ABR showing a prolonged cochlear nerve conduction ipsilateral to the spasm was a contributing factor in diagnosing an eighth nerve vascular compression when no tumoral cause was found. This was the case for three patients. Abnormalities identified during vestibular examination-recurrent vertigo, unsteadiness, disequilibrium precipitated by head movement, caloric weakness, and directional preponderance-are described in the literature as elements of a cochleovestibular nerve compression syndromen (CNCS).10 Six HFS patients with associated disorders represented a large proportion of the total number of patients (13) operated on for HFS in the Department of ENT Medicine at Purpan Hospital during the same period. This high proportion may be surprising. However, it has been confirmed by audiovestibular investigation before and after facial nerve MVD, as well as by surgical findings in four cases and by preoperative neuroradiologic data in three cases. Results analysis. Another contribution of this study consists in showing an improvement in cochleovestibular troubles after MVD of the facial nerve. This demonstrates the concomitant compression of the eighth nerve in some patients with HFS. In two patients relieved of vertigo and HFS simultaneously (cases 2 and 6), the hypothesis of concomitant compression is supported primarily by clinical and vestibular postoperative data (complete normalization of an hitial canal paresis). However, in three patients (cases 1,4, and 5) who had improved cochleovestibular symptoms, the second conflict accompanying the compression of the facial nerve root was clearly evident during the surgery. The second conflict was caused by a branch of the AICA or by an arteriole passing between the cochleovestibular and facial nerves in their cisternal crossing. MVD of the facial nerve simultaneously freed the eighth nerve, allowing improvement of some or all symptoms and even normalization of the audiovestibular tests. The surgical findings are consistent with the data in the literature about MVD of the eighth nerve. Regarding the type of the conflicting vessel, Schwaber and Hall10 studied a population of 13 patients operated on for a CNCS. They found 11 cases of vascular compression of the eighth nerve caused by the AICA and 3 cases caused by an arteriole, possibly associated with a vein. According to Leclerq and Hil1,ll contrary to the classic findings in HFS, the conflicting vessel does not necessarily compress the cochleovestibular nerve at the root entry zone. The compression may occur anywhere along the eighth nerve between the brainstem and the porus acusticus. This can be explained by the location of the glial- Schwann sheath junction, which sometimes may be very peripheral. In some cases it may even be located as far away as within the IAC.12 At this transition zone the myelin is very thin, and a pulsatile compression may alter the eighth nerve conduction.13 Pathophysiologic hypotheses. Two pathophysiologic hypotheses must be raised. First, the compression by a branch of the AICA or an arteriole of both the facial 746 nerve and a part of the cochleovestibular nerve may result in a focal demyelination involving at least the eighth nerve at the compression site. The ephaptic transmission theory described by Nielsenl4 in HFS is based on the demyelination of the facial nerve at the root entry zone. Mechanical stress caused by a vascular loop in this demyelinated area would slow the conduction velocity in nervous fibers. This would trigger formation of false synapse or ephapse between axons of facial nerve fibers at the site of compression. The theory of ephaptic transmission described for HFS and trigeminal neuralgia may then be applied to the cochleovestibular nerve, as proposed by Wiet and a1.12 This hypothesis, based on a possible local demyelination of the cochleovestibular nerve, is not consistent with the histopathologic findings of Schwaber and Whetzell.15 These authors, when examining vestibular nerves of six parents operated on for CNCS, did not observe any segmental demyelinations. he major histopathologic feature in their six observations is an endoneurial fibrosis, as is often seen in response to chronic degeneration of axons following vestibular neuritis. According to these authors, this fibrosis may lead to a chronic autostimulation of the nerve, spurred by a pulsatile compression. The chronic autostimulation of the nerve results in reorganization of the vestibular nucleus, which leads to the imbalance and vestibular abnormalities observed in CNCS. According to the theory of Schwaber and Whetzell, the association of HFS and CNCS may be due to a hyperactivity of the facial and the vestibular nuclei. Hyperactivity of the facial nucleus would be linked to a continuous antidromic stimulation, while the vestibular nucleus disorders would be related to a modification of afferent vestibular inputs. The focus of ectopic excitation may differ for each of the two nerves concerned: while probably located at the root entry zone for the facial nerve, the cochleovestibular nerve would be affected between the brainstem and the porus acusticus (as found in cases 1,4, and 51, or even within the LAC. Preoperative MRI findings in case 6 suggest an IAC location of the cochleovestibular nerve compression. Whatever the pathophysiologic mechanism causing CNCS, not all vascular loops in contact with the eighth nerve lead to symptoms. In a patient operated on for HFS alone we have, indeed, found a PICA loop in contact with both the seventh and eighth nerves. The second hypothesis is that the presence of cochleovestibular symptoms in HFS syndrome may not be related to a direct compression of the eighth nerve by a vascular loop, but to repercussions on the eighth nerve adjacent to the conflict of the facial nerve s compression. Mechanical and pulsatile irritation of the facial nerve may be transmitted to the cochleovestibular nerve, which would cause the reported and demonstrated disorders. This hypothesis is supported by the ABR waveform and latency alterations we observed as Teflon was inserted at the root entry zone between the facial nerve and the conflicting vesselwhether AICA as in case 3 or PICA as in case 2. CONCLUSION In this study a cochleovestibular injury is demonstrated with audiovestibular, radiologic, and surgical data Thai Van et at.: Cochleovestibular Disorders

7 in six patients operated on for HFS. However, the pathophysiology of this association remains to be clarified, and it may be complex. Our hypotheses support the existing pathophysiologic theories on cryptogenic HFS. The fact that cochleovestibular disorders can be improved by a microdecompression of facial nerve pleads for a systematic search for a possible vascular compression of the eighth nerve, especially in patients presenting a slow cochlear conduction or any unclear vestibular symptoms. For this type of patient, our results may open some interesting therapeutic horizons. ACKNOWLEDGMENT The authors gratefully acknowledge Michelle Delore and Gisele Bouissou for their assistance in preparation of the manuscript. BJBLIOGRAPHY 1. Moller MB, Moller AR. Audiometric abnormalities in hemifacia1 spasm. Audiology 1985;24: Moller MB. Results of microvascular decompression of the eighth nerve as treatment for disabling positional vertigo. Ann Otol Rhino1 Layngol 1990;99: Moller MB, Moller AR, Jannetta PJ, Jho HD, Sekhar LN. Microvascular decompression of the eighth nerve in patients with disabling positional vertigo: selection criteria and operative results in 207 patients. Acta Neurochir Wen) 1993;125: Moller MB, Moller AR, Jannetta PJ. Brain stem auditory evoked potentials in patients with hemifacial spasm. Laryngoscope 1982;92: Holley P, Bonafe A, Brunet E, Manelfe C. Apport de I angioirm temps de vol dans l exploration des conflits neurovasculaires. J Neuroradiol 1996;21: Caces F, Chays A, Magnan J. Resultats du traitement du spasme de l hkmiface par dbcompression neuro-vasculaire ehimgicale et endoscopique: analyse de 60 dossiers. Ann Otolayngol Chir Cervicofac 1996;113: Moller MB, Moller AR. Loss of auditory function in microvascular decompression for hemifacial spasm. J Neurosurg 1985;63: Sindou M, Fobe JL, Ciriano D, Fischer C. Hearing prognosis and intraoperative guidance of brainstem auditory evoked potentials in microvascular decompression. Laryngoscope 1992; 102: Caces F. Spasme essentiel de l hkmiface: la rkalite du conflit vasculo-nerveux, ti propos de 60 observations. [Doctoral thesis]. University of Marseilles, Faculty of Medicine, Schwaber MK, Hall JW. Cochleovestibular nerve compression syndrome, I clinical features and audiovestibular findings. Laryngoscope 1992; 102: Leclercq TA, Hill CL. Retromastoid microsurgical approach to vascular compression of the eighth cranial nerve. Laryngoscope 1980;6: Wiet RJ, Schramm DR, Kazan RP. The retrolabyrinthine approach and vascular loop. Laryngoscope 1989; 10: McCabe BF, Gantz BJ. Vascular loop as a cause of incapacitating dizziness. Am J Otol 1989;10(2): Nielsen VK. Pathophysiology of hemifacial spasm, I: ephaptic transmission and ectopic excitation. Neurology 1984; Schwaber MK, Whetzell WO. Cochleovestibular nerve compression syndrome, 11: vestibular nerve histopathology and theory of pathophysiology. Laryngoscope 1992;102: Thai Van et at.: Cochleovestibular Disorders 747