MARSHALL F. WILKINSON, PH.D., AND ANTHONY M. KAUFMANN, M.D., F.R.C.S.(C)

Transcription

1 J Neurosurg 103:64 69, 2005 Monitoring of facial muscle motor evoked potentials during microvascular decompression for hemifacial spasm: evidence of changes in motor neuron excitability MARSHALL F. WILKINSON, PH.D., AND ANTHONY M. KAUFMANN, M.D., F.R.C.S.(C) Centre for Cranial Nerve Disorders, Division of Neurosurgery, University of Manitoba, Winnipeg, Manitoba, Canada Object. Hemifacial spasm (HFS) is thought to be due to a hyperactive facial motor nucleus consequent to chronic neurovascular contact. The lateral spread (LS) response is presumed to reflect changes in facial motor neuron excitability. Facial muscle motor evoked potentials (MEPs) use the same efferent pathway as LS, therefore the authors speculated that these potentials should reflect differences consistent with changes at the facial motor nucleus level. Methods. Monitoring of LS and bilateral facial MEP was performed in 10 consecutive patients undergoing MVD for HFS. Ipsilateral facial MEPs were monitored in 17 patients undergoing MVD for trigeminal neuralgia (TN). Latency, amplitude, and duration of the MEPs were compared before and after MVD. Following MVD the duration of ipsilateral MEPs decreased from to msec and their amplitude decreased from to V (p 0.01). These changes were consequent to the abolition of LS in eight of 10 patients and an approximately 50% reduction in two patients. The relationship between the reduction in MEPs and changes in LS was significant (p 0.01). Control facial muscle MEPs (nonspastic side in patients with HFS and in those with TN) did not change significantly during the MVD procedure. Spasms were alleviated in nine of 10 patients, and there was no indication of facial nerve damage intraoperatively or postoperatively. Conclusions. Facial muscle MEPs represent a novel tool for studying the neurophysiological mechanisms of HFS in particular and monitoring the facial nerve in general. Data in this study support the hypothesis that the development of HFS and its alleviation with MVD are related to changes in facial motor nucleus activity. KEY WORDS hemifacial spasm intraoperative neurophysiology facial motor evoked potential microvascular decompression H EMIFACIAL spasm is a relatively rare condition involving progressive twitching of the facial muscles. The origin of the spasms is generally agreed to be the result of vascular compression of the facial nerve REZ. Sufferers of HFS typically have two treatment options: serial botulinus toxin injections or surgical intervention using MVD to eliminate the neurovascular contact. Moller and Jannetta pioneered the use of intraoperative evoked EMG recordings during MVD procedures and eloquently described the LS response. Lateral spread is unusual because it is produced by electrical stimulation of one branch of the facial nerve, whereas the EMG response is seen in the musculature innervated by a different branch of the facial nerve. Results of studies on the neurophysiological mechanisms producing LS have led to two competing hypotheses: a peripheral mechanism via ephaptic transmission or a central phenomenon involving hyperexcitability of the facial Abbreviations used in this paper: ANOVA = analysis of variance; BAEP = brainstem auditory evoked potential; EMG = electromyography; HFS = hemifacial spasm; LS = lateral spread; MEP = motor evoked potential; MVD = microvascular decompression; REZ = root entry zone; SEM = standard error of the mean; TES = transcranial electrical stimulation; TMS = transcranial magnetic stimulation; TN = trigeminal neuralgia. 64 motor nucleus. Arguments for both explanations have created an interesting debate. 9,14,23,26,28 Transcranial electrical stimulation has been used almost exclusively for intraoperative monitoring of the spinal cord motor pathways. 4,5,13,20,27 To our knowledge there has been no published report on the application of this method for the intraoperative study of HFS or monitoring of the facial nerve. We speculated that the application of TES over the facial motor cortex would reveal differences in facial muscle activation between symptomatic and asymptomatic sides in patients with HFS. In the present study we report on the first use of TES in the intraoperative monitoring of the facial nerve in patients undergoing MVD for HFS. Moreover, we show novel facial muscle MEP dynamics that mimic changes in LS during the course of MVD surgery. Clinical Material and Methods We recently incorporated the study of facial muscle MEPs into our intraoperative monitoring protocols for all posterior fossa surgeries including MVD operations. Results from a consecutive series of patients with HFS and TN who had undergone such monitoring procedures were analyzed. Following the induction of anesthesia using a short-

2 Facial muscle motor evoked potential monitoring during MVD TABLE 1 Characteristics of facial MEPs obtained on the nonspasm side of patients with HFS and in patients with TN before and after MVD* Latency Amplitude Duration Group Value (msec) p Value Value ( V) p Value Value (msec) p Value nonspasm side of patients w/ HFS pre-mvd post-mvd patients w/ TN pre-mvd post-mvd * Values represent the means SEM. Probability values were obtained using the paired t-test. acting muscle relaxant, patients were placed in a Mayfield headrest and positioned laterally. Electrodes were situated for routine intraoperative monitoring of the seventh and eighth cranial nerves via facial muscle needle EMG and BAEPs. To stimulate MEPs we positioned corkscrew electrodes over the facial motor cortex at C-3 and C-4 for bilateral MEP monitoring through the use of the polarity-switching feature of the CV-2 stimulator (Cadwell Laboratories, Inc., Kennewick, WA). Stimuli were applied as single shocks with a pulse width of 150 to 200 sec and a voltage range of 90 to 305 V. Facial nerve function and the LS response were monitored using needle EMG recordings from the orbicularis oculi, orbicularis oris, and mentalis muscles through paired stainless-steel needle electrodes inserted subdermally. Lateral spread activity was evoked by subdermal stimulation of the temporal branch of the facial nerve, as described previously ,26 Stimuli were delivered using paired needle or surface electrodes with the cathode positioned proximally. Stimulus parameters consisted of the following: pulse width 0.1 to 0.2 msec, amplitude less than or equal to 30 ma. We observed the LS response throughout the procedure using the same stimulus amplitude. Recorded signals were continuously displayed and fed through an audio amplifier allowing EMG potentials to be heard by the surgical team. Conventional amplification and band-pass filtering ( Hz) methods were used. Several anesthetic regimens were compatible with monitoring facial MEPs, but a balanced anesthetic mixture of propofol (50 g/kg/hr), remifentanil (0.1 g/kg/ hr), and desflurane was used in most cases. In addition to obtaining bilateral facial muscle MEPs in patients with HFS, we also monitored the facial motor pathway in patients undergoing MVD for TN and compared the data. To do so was of interest for three reasons. We could observe facial MEPs in patients with a normal facial motor neuron pool, determine whether there was an effect of surgery per se on the facial MEP responses, and assess the use of the facial MEP monitoring technique in surgical scenarios other than those for HFS. Each EMG record was obtained within 1 minute of each other to maintain, as closely as possible, the temporal relationship between the evoked EMG signal and the surgical event. Individual latency, peak-to-peak amplitude, and duration of the evoked potentials were determined from 3 to 10 consecutive traces and averaged. These values were then pooled and reported as the means SEM. In a statistical analysis of the data we performed the paired t-test, Wilcoxon signed-rank test, and one-way ANOVA, with significance set at a probability value less than Results The study protocol was used in 10 consecutive patients with HFS and 17 patients with TN while undergoing MVD surgery. There was no postoperative complication such as hearing loss or new facial nerve weakness in any of these patients. Among those with HFS, nine were spasm-free after 3 months, whereas one patient experienced the recurrence of facial spasm. Figure 1 illustrates the LS and facial muscle MEP records obtained in a patient undergoing a left-sided MVD for HFS. During decompression the LS response was abolished in eight of 10 patients, and was reduced between 40 and 50% in the other two. Similarly, MEPs on the spasm side decreased in amplitude and duration consequent to the decompression procedure. Figure 2 summarizes the mean changes in facial MEP characteristics (latency, amplitude, and duration) on the spasm side in patients with HFS. In these patients, decompression of the facial nerve on the symptomatic side led to a reduction in MEP amplitude, from to V (p = 0.01, paired t-test). The morphological characteristic also changed from a complex of potentials to usually a single compound muscle action potential. This feature was reflected in the changes in the mean duration of the potentials, which decreased from msec before MVD to msec after MVD (p 0.01). Thus, as LS changed so too did the facial MEP recorded from the same muscle. These changes were significantly related (p 0.01, Wilcoxon signed-rank test). Whereas the MEPs were significantly reduced following MVD, the onset latencies remained stable ( msec before MVD and msec after MVD; p = 0.23). The facial MEP responses from the contralateral mentalis muscle (asymptomatic) were not significantly changed after MVD. As summarized in Table 1, contralateral MEP amplitudes were V before and V after MVD (p = 0.83). The duration of contralateral MEPs was msec before and msec after MVD (p = 0.25). Latencies from the nonspasm side were also unchanged following MVD ( msec before and msec after; p = 0.58). The stability of the contralateral MEPs indicates that the changes observed on 65

3 M. F. Wilkinson and A. M. Kaufmann FIG. 1. Representative EMG recordings obtained from the mentalis muscle at indicated intervals during MVD for leftsided HFS. Responses were obtained sequentially. Note the reduction in the ipsilateral (spasm side) facial MEP response corresponding to changes in LS, whereas the contralateral (nonspasm side) MEP response remained stable. Vertical scale bar denotes 200 V for MEP records and 50 V for LS records; horizontal scale bar denotes 10 msec. the spasm side were specific to the HFS syndrome and the surgical decompression rather than to fluctuations in anesthesia, damage to the facial nerve, or other nonspecific factors. Facial muscle MEP responses were also used for facial nerve monitoring during MVD for TN (17 patients) and the data are shown in Table 1. Potentials were elicited in 17 cases and data were obtained throughout the procedure in 13 cases. In one case we were unable to elicit MEPs, and three other cases required progressively higher stimulus intensities to obtain responses, probably due to inhalant anesthetic agents being used at 1 minimum alveolar concentration. These cases were excluded from analysis because the baseline stimulus voltages were lower than those used at the end of the procedure. As indicated in Table 1, there was no significant differences, pre- and post-mvd, in the latency, amplitude, or duration of facial muscle MEPs. A comparison of baseline MEP characteristics of spastic and nonspastic sides in patients with HFS and patients with TN revealed no significant difference among the patient groups. The only difference was the applied voltages used to elicit the MEP responses. The pre-mvd latencies in patients with HFS (spastic and nonspastic sides) and those with TN were not significantly different (p = 0.29, one-way ANOVA). Discussion The present analysis represents the first use of TES to elicit facial muscle MEPs for the study of HFS. We applied the stimulation technique on the spasm and nonspasm sides in patients undergoing MVD for HFS and in a separate group of patients who underwent surgery for TN. During the MVD procedure for HFS, changes in the facial MEPs were temporally related to changes in the LS response. We interpreted these novel observations as supporting the hypothesis that chronic neurovascular contact produces changes in the facial motor neuron pool, a state reversed by MVD. Thus, significant alterations in facial MEPs follow the decompression procedure and these changes are significantly related to the reduction in, or the disappearance of, the LS response. Motor evoked potentials have been used mainly for monitoring spinal cord function, 3,5 although Neuloh and Schramm 22 have recently described its use during cerebral aneurysm surgery. Because intraoperative MEPs have not been used during MVD surgery for HFS, we speculated that facial motor pathways would be demonstrably different between spasm and nonspasm sides. In addition, we planned to compare these responses in patients with TN undergoing a similar procedure during surgery when we routinely monitor facial nerve function. Note that following MVD for HFS, facial MEP amplitudes and durations significantly decreased on the affected side, whereas responses on the unaffected side or those in patients with TN remained unchanged after MVD. Although the mechanism for LS and HFS remains speculative, there is good evidence that patients with HFS have an elevated level of facial motor neuron excitability. 8,9,15,19,26 Our results clearly showed a dramatic reduction in facial MEP amplitude and duration on removal of the offending neurovascular contact. Interestingly, however, pre-mvd amplitude, duration, and latency were similar across all groups (spasm and nonspasm sides in patients with HFS and patients with TN). A priori, we may have expected some differences in patients with HFS between the spasm and nonspasm sides before MVD. Instead, we saw no difference in any parameter prior to MVD. This result may be explained by our method of determining the applied stimulus voltage (determined empirically as a suprathreshold) or may represent an as yet unexplained physiological mechanism revealed by our technique. In subsequent studies it would be prudent to control the stimulus voltage parameter more rigorously. Examination of the applied voltages between the groups revealed a significant but unexpected difference in the voltages used in patients with HFS and in those with TN (TN, V; HFS/spasm side, V; HFS/nonspasm side, V; p 0.05, ANOVA). This finding is consistent with differences in the excitability of the facial motor neuron pool between patients with HFS and those with TN. Interestingly, MEPs on the nonspasm side of the face were obtained as readily as the those on the spasm side, based on the mean voltages applied. (It is tempting to speculate that changes induced by 66

4 Facial muscle motor evoked potential monitoring during MVD FIG. 2. Bar graphs demonstrating a summary of facial MEP amplitude, duration, and latency before and after MVD (mean SEM) in patients (spasm side) with HFS. *p 0.01, paired t-test. chronic neurovascular association may be translated bilaterally.) We also found that stimulation of the motor cortex resulted in ipsilateral responses in patients with HFS or TN (data not shown), thus indicating bilateral activation of facial motor pathways. Evidence of abnormal blink-reflex responses from the asymptomatic sides in patients with HFS has been previously reported. 12 The application of TMS has been used in the study of the facial nerve in healthy volunteers 1,2 as well as in patients with HFS. 10,11,24 Overall, the activation latencies of the mentalis muscle by using TMS tend to be longer than those evoked with TES in the present study. In our patient groups, facial MEP latencies were between 5 and 8 msec, whereas TMS created latencies in excess of 10 msec. 10,11 Similar to the experience of Kojima, et al., 10 we noticed a tendency for longer latencies on the symptomatic side compared with those from the asymptomatic side, but the difference was not significant. This lack of significance was similar to that in other investigations involving TMS. 11,24 The absolute latency values were no doubt influenced by the use of different cortical stimulation methods, which are known to activate different cortical components selectively. 6 Examination of the facial muscle MEP latency data revealed no change before or after MVD in any of our three groups. This outcome indicates that the MVD procedure was atraumatic and that the changes we observed in MEPs and LS are not adequately explained by facial nerve injury. This point is crucial as it could be argued that demyelinated fibers (due to chronic neurovascular contact) would be susceptible to surgical trauma and rendered inactive through conduction block. Thus, a peripherally generated mechanism would be sufficient to explain not only changes in the LS response, but also our observations with regard to the MEPs. Nevertheless, we reject this position based on both stable recorded facial MEP latencies and our consistent observations that the fast (early) components of the MEP reflect changes in amplitude (Fig. 1). If demyelinated fibers are rendered inactive by surgical trauma, then changes in the fast (that is, myelinated) components of the MEP should not have occurred. In addition, we as well as others 19 have observed that the process of opening the dura mater can cause a reduction in, or disappearance of, the LS response. If lateral spreading is caused by a demyelinating phenomenon and its disappearance caused by surgical trauma, then this result should not occur. This reasoning is further supported by the work of Murakami, et al., 21 who showed that the HFS and LS caused by a vertebral artery aneurysm impinging the facial nerve REZ could be abolished following endovascular coil insertion in the aneurysm. This information would rule out surgical trauma as an explanation for the loss of the LS response. We also observed no neurogenic (injury) EMG potentials intraoperatively or any instance of new postoperative weakness in any patient who underwent MVD. If fiber dropout was occurring because of surgical trauma, then we should have seen evidence on EMG monitoring as the damaged facial nerve fibers depolarized and entered conduction block. Finally, it is noteworthy that the degree of neurovascular compression (by visual inspection) was in no way related to the dynamic changes in either the LS or MEPs. If a demyelinating mechanism of action was a sufficient explanation, then we would have expected to see the greatest electrophysiological effects in patients with the most extensive REZ compression. This result was not observed. Although we acknowledge the relatively small sample size in this study, at this juncture we would rule out peripheral nerve injury as a sufficient explanation for HFS in general or our data specifically. In their study using TMS, Kojima, et al., 10 reported a reversal of the threshold of activation in the one patient with HFS who had been evaluated before and after MVD. This finding is comparable to our observation of reduced spasmside MEP amplitude and duration following MVD. In all patients with HFS, amplitude and duration significantly decreased following MVD on the spasm side. These changes were not observed from the nonspasm side or in the TN group, indicating that surgical decompression and not other variables factor into the mechanism of facial MEP changes. The median reduction in MEP amplitude was 66.2% (range %). Concurrently, the majority of patients (eight of 10) had complete abolition of LS on completion of MVD. Thus, there was a significant statistical relationship between the change in LS and that in facial MEP (p 0.01, Wilcoxon signed-rank test). We believe that the changes in LS and MEP indicate a common mechanism of action. Our observations are consistent with the hypothesis that chronic neurovascular contact results in a labile facial motor 67

5 M. F. Wilkinson and A. M. Kaufmann neuron pool and that removal of that stimulus reduces or abolishes the abnormal facial motor responses. Indeed, our data demonstrate the reduction in facial motor neuron excitability to levels below those in the two control groups. Previous observations in the blink reflex 15,19 and facial muscle F-waves 8,9 measured before and after MVD are consistent with our results. Collectively, changes in the blink reflex, F-waves, and LS and now our observations of changes in MEP indicate that pulsatile irritation of the facial nerve is necessary to cause and sustain abnormal central excitability changes. As discussed previously, an alternative explanation of cross-transmission (ephaptic) at the site of neurovascular contact has been proposed as a sufficient explanation for the features of HFS. 23 Longer blink reflex R1 latencies establish a necessary precondition for ephaptic transmission and such data have been reported. 1,2,25 Nevertheless, data from additional studies on the blink reflex R1 component have been equivocal, given that increased R1 latencies have not been observed. 7,11,23 In addition, blink reflexes measured intraoperatively have been shown to be present on the spasm side before MVD and absent following the decompression procedure; simultaneously monitored R1 latencies on the nonspastic side were not measurable. 19 Like the blink reflex, MEPs provide a means of examining the complete efferent pathway of the facial nerve. The changes in spasm-side MEP duration following MVD indicate that more facial motor units were activated (or were available for activation) prior to decompression, even though the stimulus parameters remained the same. These observations were concurrent with stable MEPs on the nonspasm side. We believe that this result is best explained by motor neuron excitability rather than ephaptic transmission. Moreover, we observed no change in facial MEP latency during the MVD procedure, indicating that peripheral nerve compression had no marked effect on the orthodromic conduction of the facial nerve. Conclusions We have shown that facial muscle MEPs are compatible with routine surgical monitoring of the facial nerve and can proceed uninterrupted during surgery. We applied this technique for the first time in patients with HFS and TN who were undergoing MVD surgery. The results in patients with HFS are consistent with vascular compression induced plasticity changes within the facial motor nucleus and provide an interesting line along which to continue this study. Moreover, our experience with facial MEPs during surgery for TN indicates that the method has broader neurosurgical applications for facial nerve monitoring and we now routinely use these potentials during surgery for vestibular schwannoma and cerebral aneurysm. References 1. Benecke R, Meyer BU: Magnetic stimulation of corticonuclear systems and of cranial nerves in man: physiological basis and clinical application. Electroencephalogr Clin Neurophysiol Suppl 43: , Benecke R, Meyer BU, Schonle P, Conrad B: Transcranial magnetic stimulation of the human brain: responses in muscles supplied by cranial nerves. Exp Brain Res 71: , Burke D, Hicks RG: Surgical monitoring of motor pathways. J Clin Neurophysiol 15: , de Haan P, Kalkman CJ, Jacobs MJ: Spinal cord monitoring with myogenic motor evoked potentials: early detection of spinal cord ischemia as an integral part of spinal cord protective strategies during thoracoabdominal aneurysm surgery. Semin Thorac Cardiovasc Surg 10:19 24, Deletis V, Kothbauer K: Intraoperative neurophysiology of the corticospinal tract, in Stolberg E, Sharma HS, Olsson Y (eds): Spinal Cord Monitoring. New York: Springer, 1998, pp Di Lazzaro V, Oliviero A, Pilato F, Saturno E, Dileone M, Mazzone P, et al: The physiological basis of transcranial motor cortex stimulation in conscious humans. Clin Neurophysiol 115: , Eekhof JL, Aramideh M, Speelman JD, Devriese PP, Ongerboer De Visser BW: Blink reflexes and lateral spreading in patients with synkinesia after Bell s palsy and in hemifacial spasm. Eur Neurol 43: , Ishikawa M, Namiki J, Takase M, Ohira T, Nakamura A, Toya S: Effect of repetitive stimulation on lateral spreads and F-waves in hemifacial spasm. J Neurol Sci 142:99 106, Ishikawa M, Ohira T, Namiki J, Ajimi Y, Takase M, Toya S: Abnormal muscle response (lateral spread) and F-wave in patients with hemifacial spasm. J Neurol Sci 137: , Kojima A, Ohira T, Takase M, Kawase T: Long-latency response to transcranial magnetic stimulation in patients with hemifacial spasm. Electroencephalogr Clin Neurophysiol 109: , Kotterba S, Tegenthoff M, Malin JP: Hemifacial spasm or somatoform disorder postexcitatory inhibition after transcranial magnetic cortical stimulation as a diagnostic tool. Acta Neurol Scand 101: , Lamas G, Poignonec S, Fligny I, Soudant J, Willer JC: Recovery of normal excitability of the facial motor nucleus following facial nerve decompression in hemifacial spasm. Eur Arch Otorhinolaryngol (Suppl): S555 S556, Lang EW, Beutler AS, Chesnut RM, Patel PM, Kennelly NA, Kalkman CJ, et al: Myogenic motor-evoked potential monitoring using partial neuromuscular blockade in surgery of the spine. Spine 21: , Moller AR: Vascular compression of cranial nerves. I. History of the microvascular decompression operation. Neurol Res 20: , Moller AR, Jannetta PJ: Blink reflex in patients with hemifacial spasm. Observations during microvascular decompression operations. J Neurol Sci 72: , Moller AR, Jannetta PJ: Hemifacial spasm: results of electrophysiologic recording during microvascular decompression operations. Neurology 35: , Moller AR, Jannetta PJ: Microvascular decompression in hemifacial spasm: intraoperative electrophysiological observations. Neurosurgery 16: , Moller AR, Jannetta PJ: Monitoring facial EMG responses during microvascular decompression operations for hemifacial spasm. J Neurosurg 66: , Moller AR, Jannetta PJ: Physiological abnormalities in hemifacial spasm studies during microvascular decompression operations. Exp Neurol 93: , Morota N, Deletis V, Constantini S, Kofler M, Cohen H, Epstein FJ: The role of motor evoked potentials during surgery for intramedullary spinal cord tumors. Neurosurgery 41: , Murakami H, Kawaguchi T, Fukuda M, Ito Y, Hasegawa H, Tanaka R: Monitoring of the lateral spread response in the endovascular treatment of a hemifacial spasm caused by an unruptured vertebral artery aneurysm. Case report. J Neurosurg 101: , Neuloh G, Schramm J: Monitoring of motor evoked potentials compared with somatosensory evoked potentials and microvascu- 68

6 Facial muscle motor evoked potential monitoring during MVD lar Doppler ultrasonography in cerebral aneurysm surgery. J Neurosurg 100: , Nielsen VK: Pathophysiology of hemifacial spasm: II. Lateral spread of the supraorbital nerve reflex. Neurology 34: , Oge AE, Yazici J, Boyaciyan A, Tanyeri S, Celik M, Kenyalioglu R, et al: Magnetic stimulation in hemifacial spasm and post-facial palsy synkinesis. Muscle Nerve 16: , Sood S, Vyas L, Taori GM: Hemifacial spasm: early postoperative normalization of blink reflex latency. Br J Neurosurg 7: , Tankere F, Maisonobe T, Lamas G, Soudant J, Bouche P, Fournier E, et al: Electrophysiological determination of the site involved in generating abnormal muscle responses in hemifacial spasm. Muscle Nerve 21: , Taylor BA, Fennelly ME, Taylor A, Farrell J: Temporal summation the key to motor evoked potential spinal cord monitoring in humans. J Neurol Neurosurg Psychiatry 56: , Yamashita S, Kawaguchi T, Fukuda M, Suzuki K, Watanabe M, Tanaka R, et al: Lateral spread response elicited by double stimulation in patients with hemifacial spasm. Muscle Nerve 25: , 2002 Manuscript received August 25, Accepted in final form March 15, Address reprint requests to: Marshall F. Wilkinson, Ph.D., Intraoperative Monitoring Program and Centre for Cranial Nerve Disorders, Section of Neurosurgery, Health Sciences Centre, Winnipeg, Manitoba, Canada R3A 1R9. mwilkinson@exchange. hsc.mb.ca. 69