Intraoperative neurophysiologic monitoring in microvascular decompression for hemifacial spasm

MVD consists of vascular decompression relieving the facial nerve from the offending vessel and implant for the interposition or transposition between the offending vessel and the facial nerve. Offending vessel is usually in the postero-inferior cerebellar artery and not frequently in the antero-inferior cerebellar artery and/or the vertebra-basilar artery [4]. Although the MVD is a safe and effective treatment modality, an important complication of hearing loss may occur. A potential cause of conduction hearing loss is acoustic trauma due to drill noise. Causes of sensorineural hearing loss include direct trauma or stretching of cranial nerve VIII during retracting the cerebellum, manipulation of the vessels; labyrinthine artery or antero-inferior cerebellar artery, newly developed compression by implant material [5]. The operator pays attention from the approach direction to the manipulation procedure so that the perforating arteries to the brainstem is not stretched or damaged, and bipolar coagulation is not used to avoid damage caused by heating. Despite these precautions, hearing loss occurs occasionally, although true prevalence is unclear due to reporting is relying on patient-reported subjective hearing loss and/or inadequate perioperative audiometric test. BAEPs monitoring decreases the rate of healing loss from 7.7%–20% to below the level of 2.3% [5, 6].

Each wave form of the BAEPs is the result of not one but several generators in relatively small size brainstem. This makes it quite difficult to pinpoint the exact generator of each wave form. However, at least one main generator per wave form has been consistently identified based on study of electocochleography and BAEPs, and clinicopathologic correlation. The three most important waveforms are result of action potentials of following anatomic structures: “wave I” result of distal part of the cochlear part of cranial nerve VIII, “wave III” result of lower pons at the level of the superior olivary complex, “wave V” result of lower midbrain [7].

American Society of Neurophysiological Monitoring and American Clinical Neurophysiology Society (ACNS) recommends a protocol of 5–12 Hz/s stimulation rate and 1,000–4,000 averaging [8]. However, since this protocol requires relatively long time to obtain the BAEPs, there is a limitation in protecting the cranial nerve VIII, which may be damaged in a short time. Therefore, many physicians using higher frequency and lower averaging to obtain the reliable BAEPs quickly within short time [9, 10]. In one study of these protocols used a protocol of 43.9 Hz/s stimulation rate and 400 averaging within about 9 seconds [9].

ACNS recommend to alert the surgeon when significant changes of BAEPs occur to prevent hearing loss. The significant changes of ACNS are latency prolongation ≤ 1 ms or ≤ 10% increase, and decrease of amplitude ≤ 50% of wave V when stimulated according to their protocol [10]. In a recent study, the following criteria were suggested as a critical warning sign of real-time BAEPs when stimulated with their protocol. Hearing loss was associated with loss of wave V, and latency prolongation of 1 ms with at least 50% reduced amplitude, but latency prolongation of 1 ms without reduced amplitude > 50% was not associated with hearing loss [11].

In HFS patients, stimulation of one branch of the facial nerve may result in the activation of the facial muscles innervated by other branches of the facial nerve. This response has been termed LSR. The LSR is an abnormal pathologic electrophysiologic response. The exact mechanisms are not clearly proved, but ephatic transmission of lesion site alone or combination with hyperexcitability of motor nucleus is thought to be related to this response.

The LSR monitoring is consisted of direct stimulation of facial nerve, commonly at zygomatic braches with pulse duration of 0.1–0.2 ms, frequency of 3 ± 1 Hz and intensity ranging from 1–2 mA to 5–25 mA, and recording of compound muscle action potentials of following muscles. The recording sites are the mentalis muscle when the zygomatic branch is stimulated, and in the orbicularis oculi muscles when the mandibular branch is stimulated. When the baseline threshold is determined, any stimulation will proceed responses near baseline level at operation field. However, the LSR may disappear before the MVD is performed, for instance, when the dura is opened. It is suggested that stimulation instensity and/or rate should be increased in these circumstances to reestablish the presence of the LSR in these cases.

It is controversial to use the LSR disappearance as an evaluation tool for effectiveness of MVD. If the amplitude of LSR decreases, but LSR does not disappear when offending vessel is moved off the nerve, it may indicate that there are another offending vesseles. In this situation, when the other vessel is identified and moved off the facial nerve, the LSR disappears [4]. Although there are a few cases, persistent symptoms or recurrence of spasm were noticed even after surgery for reasons of Teflon felt factor and vascular change, as well as inadequate decompression or neglected offending vessels [12, 13]. In some cases, it may take about 1 year after MVD for HFS disappearance, this phenomenon supports the central mechanism theory of HFS development including facial motor nucleus hyperexcitability [4]. Despite its modest sensitivity, monitoring of this LSR is performed to guide the operator during surgery because it shows high specificity [14]. Persistent response carries a high risk of persistent hemifacial spasm. The disappearance of LSR during MVD can be a useful tool to predict resolution of symptom shortly after surgery [15]. Although LSR monitoring has limited value in predicting prognosis, it appears to have favorable outcome in short-term period, on the other hand, it does not appear to be effective in predicting the outcome in long-term period [14, 15].