ORIGINAL ARTICLE

Microvascular Decompression for Hemifacial Spasm: Evaluating

Outcome Prognosticators Including the Value of Intraoperative
Lateral Spread Response Monitoring and Clinical Characteristics in
293 Patients
Parthasarathy D. Thirumala,*† Aalap C. Shah,* Tara N. Nikonow,* Miguel E. Habeych,* Jeffrey R. Balzer,*‡
Donald J. Crammond,* Lois Burkhart,* Yue-Fang Chang,* Paul Gardner,* Amin B. Kassam,*
and Michael B. Horowitz*

Abstract: Hemifacial spasm is a socially disabling condition that mani-

fests as intermittent involuntary twitching of the eyelid and progresses to
H emifacial spasm (HFS) is a condition involving involuntary,
repetitive, unilateral contraction of the muscles innervated
by the facial nerve (cranial nerve [CN] VII). Typical HFS is
muscle contractions of the entire hemiface. Patients receiving microvas-
caused by facial nerve irritation secondary to vascular compres-
cular decompression of the facial nerve demonstrate an abnormal lateral
sion at the root exit zone (RExZ), leading to involuntary, inter-
spread response (LSR) in peripheral branches during facial electromyo-
mittent spasms beginning at the orbicularis oculi muscle and
graphy. The authors retrospectively evaluate the prognostic value of
progressing down to the mentalis muscle.
preoperative clinical characteristics and the efficacy of intraoperative
Neurophysiologic investigations have provided insight into
monitoring in predicting short- and long-term relief after microvascular
the underlying mechanisms responsible for the abnormal muscle
decompression for hemifacial spasm. Microvascular decompression was
response, which appears as the lateral spread response (LSR)
performed in 293 patients with hemifacial spasm, and LSR was recorded
during routine intraoperative monitoring. Previous neurophysio-
during intraoperative facial electromyography monitoring. In 259
logic studies (Nielsen, 1985) have demonstrated demyelination/
(87.7%) of the 293 patients, the LSR was attainable. Patient outcome was
axonal injury and hyperexcitability of the facial motonucleus, as
evaluated on the basis of whether the LSR disappeared or persisted after
being responsible for the residual LSR. In a rat model, Kuroki
decompression. The mean follow-up period was 54.5 months (range,
and Moller (1994) showed that the facial motonucleus was
9 –102 months). A total of 88.0% of patients experienced immediate
involved in HFS, but previous injury causing demyelination (eg,
postoperative relief of spasm; 90.8% had relief at discharge, and 92.3%
pulsatile compression near the RExZ) was also required (Ruby
had relief at follow-up. Preoperative facial weakness and platysmal
and Jannetta, 1975). It is feasible that HFS is a total of the
spasm correlated with persistent postoperative spasm, with similar trends
electrophysiologic phenomenon between the facial motonucleus,
at follow-up. In 207 patients, the LSR disappeared intraoperatively after
given the facilitated orthodromic activity in peripheral branches
decompression (group I), and in the remaining 52 patients, the LSR
of CN VII, and demyelination.
persisted intraoperatively despite decompression (group II). There was a
Nonsurgical treatments, such as medications and local
significant difference in spasm relief between both groups within 24
intramuscular botulinum toxin (BT) injections, have been inef-
hours of surgery (94.7% vs. 67.3%) (P ⬍ 0.0001) and at discharge (94.2%
fective as long-term solutions for HFS. The only method for
vs. 76.9%) (P ⫽ 0.001), but not at follow-up (93.3% vs. 94.4%) (P ⫽
providing a long-term cure has been retromastoid craniotomy and
1.000). Multivariate logistic regression analysis demonstrated indepen-
facial nerve microvascular decompression (MVD), which has
dent predictability of residual LSR for present spasm within 24 hours of
proved effective in curing ⬎90% of patients (Moller and Jan-
surgery and at discharge but not at follow-up. Facial electromyography
netta, 1984). During surgery, concurrent monitoring of brainstem
monitoring of the LSR during microvascular decompression is an effec-
auditory evoked responses is routinely used to detect eighth nerve
tive tool in ensuring a complete decompression with long-lasting effects.
dysfunction (Haines and Torres, 1991; Yamashita et al., 2002).
Although LSR results predict short-term outcomes, long-term outcomes
The eighth cranial nerve (vestibulocochlear nerve) enables the
are not as reliant on LSR activity.
patient to hear and is pertinent to maintaining balance and body
Key Words: HFS, Spasm, LSR, MVD, Decompression, BT, Facial nerve, position. Another important monitoring tool is intraoperative
CN VII, Monitoring. electromyography and recording of LSR, which can help sur-
geons to determine whether adequate decompression has been
(J Clin Neurophysiol 2011;28: 56 –66)
achieved. LSRs elicited by stimulation of the facial nerve
branches denote the electrophysiologic perturbations consistent
From the *Departments of Neurological Surgery, †Neurology, and ‡Neuro- with HFS. When the offending vasculature is moved off the facial
science, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania,
15213, U.S.A.
nerve, the LSR is known to disappear or become markedly
Address correspondence and reprint requests to Parthasarathy D. Thirumala, attenuated. However, the practical value of LSR, as a predictor of
M.D., Center for Clinical Neurophysiology, UPMC Department of Neurosur- surgical outcome and long-term prognosis, remains controversial.
gery, UPMC Presbyterian, Suite B-400, 200 Lothrop Street, Pittsburgh, PA A retrospective study focusing on three time points (post-
15213, U.S.A.; e-mail: thirumalapd@upmc.edu.
Copyright © 2011 by the American Clinical Neurophysiology Society operative, discharge, and follow-up) was conducted on 293
ISSN: 0736-0258/11/2801-0056 patients who underwent MVD as a treatment for HFS. The study

56 Journal of Clinical Neurophysiology • Volume 28, Number 1, February 2011

Journal of Clinical Neurophysiology • Volume 28, Number 1, February 2011
investigated the efficacy of intraoperative monitoring in predict-
ing spasm persistence or resolution and identified clinical char-
acteristics that can potentially predict surgical outcome.
METHODS
Microvascular Decompression
Between January 2000 and December 2007, our center
performed 326 retromastoid MVD procedures for HFS. Preoper-
atively, all patients received a cranial MRI, audiometry, and
facial EMG testing. Decompression was achieved by placing
Teflon pledgets between the facial nerve as it exited the brain-
stem and the offending vessels, and/or by elevating and cauter-
izing compressive veins that could not be safely decompressed
with Teflon. During surgery, facial EMG monitoring was per-
formed from the initiation of general anesthesia until the time of
dural closure. Stimulating needle electrodes were inserted intra-
dermally over the zygomatic branches of the facial nerve at the
middle point of a line between the ipsilateral tragus and external
canthus of the eye. A 0.2- to 0.3-millisecond pulse wave with an
intensity of 5 to 25 mV was applied. On stimulation of the
zygomatic branch, which primarily innervated the orbicularis
oculi muscle, the team recorded and reviewed the evoked LSR
that appeared in the other facial muscles via peripheral branches,
including the frontalis (temporal), orbicularis oris (buccal) and
mentalis (marginal mandibular) muscles (Fig. 1). The eighth
cranial nerve (CN VIII) function was concurrently monitored by
looking for waveform shifts during the recording of brainstem
auditory evoked responses. In some cases, use of neuromuscular
blockade during anesthesia led to alterations in the LSR phenom-
enon due to muscular paralysis. Therefore, the anesthesiologist
used a technique that maintained the train-of-four ratio at a level
of at least 0.75. When complete decompression was achieved, the
LSR was found to disappear in most patients. When the LSR
persisted or simply decreased in amplitude, the surgeon looked
again for persistent arterial or venous compression. Residual LSR
was characterized as either an LSR that returned after initially
disappearing during the procedure (after disappearing [AD]) or
an LSR that persisted relatively unchanged throughout the de-
compression (never disappearing [ND]) (Fig. 2). After confirm-
ing that there were no further offending vessels, the surgeon
terminated the procedure and closed the craniotomy in a routine
fashion.
Data Collection and Analysis
A retrospective study was conducted with Institutional
Review Board approval from the University of Pittsburgh (IRB #:
PR008120394). Of 326 MVD procedures, data were collected
Copyright © 2011 by the American Clinical Neurophysiology Society
Microvascular Decompression for Hemifacial Spasm
FIGURE 1. A, Branches of the facial nerve (CN
VII); cervical branch not depicted. B, Monitoring
for the paradoxical lateral spread response (LSR).
The zygomatic branch is stimulated (STIM), and
an evoked EMG response can be recorded in the
orbicularis oculi (REC 1). In patients with HFS, an
abnormal evoked response, the LSR, can be seen
in the mentalis muscle (REC 2). T, temporal; Z,
zygomatic; B, buccal; M, marginal mandibular.
from 322 operations performed on 293 patients with HFS; 29
operations were reexploration surgeries due to persistent or
recurrent spasm. The patient population consisted of 103 men and
190 women ranging in age from 17 to 82 years (mean: 52.25
years). Clinical outcome data were obtained immediately after
the operation, at discharge (mean: 3.91 ⫾ 1.98 days), and at a
follow-up phone call during June 2008. Follow-up data were
collected from 208 patients who had a minimum follow-up period
of 9 months (mean: 54.5 ⫾ 27.8 months). We attempted to
contact every patient identified during the record screening to
obtain information regarding the patient’s present spasm status
and operative complications and to confirm the accuracy of data
collected from clinical notes and statements. Outcomes were
divided into two categories: success (spasm relief) and failure
(persistent spasm). Postoperative success was defined as com-
plete spasm resolution with no residual twitching within 24 hours
of operation and no more than two episodes of residual eye
twitching before discharge. At follow-up, complete relief was
defined as the reported absence of HFS, allowing for residual eye
twitching at a frequency no more than one episode per month.
Patients who experienced waxing and waning symptoms were
asked to rate the frequency and severity of their current symp-
toms on a 1 to 10 scale. Patients were instructed to consider their
preoperative symptoms as a 10 and rate their current symptoms in
comparison to that value. Patients who reported spasm with
frequency and severity ⬎3 on a 10 scale when compared with
their preoperative spasm were considered to have persistent
spasm. To minimize bias, an investigator other than the operating
surgeons and neurophysiologists conducted all of the telephone
interviews. Also, investigators responsible for collecting patient
data at all three time points were blinded to LSR results.
Statistical analyses were performed using SAS version 9.1.3
(SAS Institute, Cary, NC). Continuous variables were presented as
mean ⫾ standard deviation and categorical variables as frequency
(%). Group differences in demographic, clinical characteristics, and
outcomes were assessed using t tests, ␹2 tests, and Fisher exact tests
when appropriate. Logistic regression models were conducted to
evaluate the association of LSR and outcome at each time point
while adjusting for age, gender, prior BT use, platysmal spasm,
preoperative facial weakness, and side of spasm. P ⬍ 0.05 was
considered as statistically significant.
RESULTS
Demographics
Mean patient age was 52.25 ⫾ 12.05 years (range, 17– 82
years), with women to men ratio being 1.8:1. No patients exhibited
57
P. D. Thirumala et al. Journal of Clinical Neurophysiology • Volume 28, Number 1, February 2011

FIGURE 2. Recording of the stimulus-evoked EMG responses from two different patients at the orbicularis oculi (A) and men-
talis (B) muscles in response to stimulation of the zygomatic branch of the facial nerve. On decompression (arrow; A), the
paradoxical LSR (mentalis) becomes variable in amplitude and morphology (after disappearing) and on complete decompres-
sion (B), the LSR eventually disappears with no significant change observed from the oculi muscle group.

bilateral HFS (see Appendix, Table A2). The mean follow-up period
was 54.5 ⫾ 27.8 months (range: 9 –102 months) (Table 1).
Medical Hx
Medical and surgical histories were obtained from each
patient undergoing MVD. Twenty-two patients underwent a prior
MVD at an outside institution. Two hundred seventeen patients
(74.6%) received prior BT treatment, with the average time
between the last injection and the most recent operation being 12
months (range: 1– 49 months). Patients who underwent BT treat-
ment did so over a variable period (average: 4.23 years, range: 2
months–14 years). Common medications used to control HFS
symptoms, including anticonvulsants and antipsychotics, are
listed separately (see Appendix, Fig. A2). There was no gender
bias regarding patients undergoing prior BT treatment (P ⫽
0.561).
Preoperative Characteristics
Seven patients had evidence of a Chiari I malformation on
radiographic imaging. Nine patients had symptoms of atypical HFS
(spasms began in the buccal-oral muscles and then progressed to
involve the orbicularis oculi muscle). Sixty-seven (22.8%) pa-
tients exhibited moderate-to-severe preoperative facial weakness,
as determined by the House-Brackmann score (House and Brack-
mann, 1985) (grade III and higher). Of note, patients with prior
BT treatment were significantly more likely to demonstrate
moderate-to-severe preoperative facial weakness (P ⫽ 0.01). One
hundred thirty-seven (49.6%) patients had platysmal spasm, while
187 (85.8%) exhibited tonus (frequent eyelid locking), signs associated
with an extended spasm history. One hundred twenty-six (43.0%)
patients reported specific triggers that would initiate spasms. Thirty-six
(12.3%) patients demonstrated functional hearing loss, as determined
by the pure tone average during preoperative audiograms.

58 Copyright © 2011 by the American Clinical Neurophysiology Society

Journal of Clinical Neurophysiology • Volume 28, Number 1, February 2011 Microvascular Decompression for Hemifacial Spasm

TABLE 1. Summary of Demographic and Clinical TABLE 3. Hemifacial Spasm Resolution

Characteristics Rates—Microvascular Decompression of the Facial Nerve
(CN VII)
Variable n (%)
No. cases 293 (100.0) Postoperative Discharge Follow-Up
Discharge data available 292 (99.7) All subjects 88.0 (257) 90.8 (265) 92.3 (192)
Follow-up available 208 (70.9) Gender
Mean discharge time (days) 3.91 ⫾ 1.98 Male 90.3 (93) 95.2 (98) 98.7 (74)
Mean follow-up time (years) 4.54 ⫾ 2.32 Female 86.8 (164) 88.4 (167) 88.7 (118)
Mean age at operation (years) 52.25 ⫾ 12.25 P 0.376 0.056 0.010
Gender Age
Female 190 (64.8) ⬍50 90.4 (103) 91.2 (104) 93.2 (69)
Male 103 (35.2) ⱖ50 86.5 (154) 90.5 (161) 91.8 (123)
Spasm presentation P 0.325 0.823 0.707
Left:right 164:129 Values are presented as % (n). Bold indicates statistical significance.
Preoperative Botox usage 217 (74.6)
Tonus 187 (85.8)
Platysmal spasm 137 (49.6)
Specific triggers 126 (43.0)

TABLE 2. Compressing Vasculature Seen Near Facial Nerve

Root Exit Zone During Operation
Compressing Vessel
AICA
PICA
VA
Unnamed artery
Vein
Perforator
AICA, anterior inferior cerebellar artery; PICA, posterior inferior cerebellar artery;
VA, vertebral artery.
Operative Findings
Intraoperative Remarks
The vessels compressing the RExZ, as identified by the
surgeon, are summarized in Table 2. A majority (70.7%) of patients
had multiple compressing vessels. Of 21 patients (7.2%) requiring a
two-stage operation, 13 exhibited marked (ⱖ2 milliseconds) shifted
brainstem auditory evoked responses during the first operation,
which prompted the surgeon to terminate the procedure. Four
patients had a planned staged procedure because of a preexisting
Chiari I malformation (Chiari I decompressed during the first stage),
while three other patients demonstrated dangerous cerebellar swell-
ing during the first operation, which necessitated procedure termi-
nation and repeat surgery. During one operation, equipment mal-
function necessitated procedure termination and a second-stage
surgery.
Operative Outcomes
Facial spasm resolved in 257 patients (88.0%) within 24
hours of MVD (Table 3). Success rate increased to 90.8% at the time
of discharge (3.91 ⫾ 1.98 days). Twenty-nine patients had persistent
or recurrent spasm and underwent reexplorative surgery; results
were similar to those of first-time patients undergoing MVD at each
time point (see Appendix, Tables A4 and A5). It is noteworthy that
18 patients with immediate postoperative spasm experienced com-
plete relief at discharge, and 10 patients with postoperative relief had
n (%)
147 (50.2)
132 (45.1)
82 (28.0)
58 (19.8)
124 (42.3) FIGURE 3. Outcome as a factor of demographics (age and
58 (19.8) gender). *P ⬍ 0.05.
spasm recurrence by discharge. Among patients with follow-up data,
only 16 patients (7.7%) had not experienced marked relief despite
MVD. Of these 16 patients, 12 reported that their spasm had
recurred during the follow-up time period, despite having complete
postoperative relief. An additional 16 patients who were discharged
with persistent spasm reported complete relief at follow-up (54.5 ⫾
27.8 months). Although no significant gender differences were
present when comparing outcomes within 24 hours of surgery (P ⫽
0.376), a greater proportion of men had spasm resolution at dis-
charge (95.2% vs. 88.4%) (P ⫽ 0.056). However, there was a
significant gender difference in spasm relief at follow-up (98.7% vs.
88.7%) (P ⫽ 0.01) (Fig. 3). When dividing patients by age ⬍50
years and ⱖ50 years, younger patients had a slightly higher yet
statistically insignificant relief rate (see Appendix, Table A1).
Among patients with preoperative platysmal spasm, 22
(16.1%) experienced HFS 24 hours after the surgery. This rate was
significantly greater than those without platysmal spasm (P ⫽
0.022). There was a similar but statistically insignificant trend at
discharge and at follow-up. Preoperative facial weakness also influ-
enced operative outcome; patients with a H-B grade III or IV
(moderate-to-severe) facial palsy were more likely to demonstrate
persistent spasm than those with a grade I or II (mild) facial palsy
within 24 hours of the surgery (P ⫽ 0.012) and at discharge (P ⫽
0.016). Prior BT treatment did not predict a poor surgical outcome
within 24 hours after the operation or at discharge. However, at the
time of follow-up, 15 patients with a history of BT treatment
reported persistent or recurrent spasm, whereas only one patient
without prior BT continued to experience spasm (P ⫽ 0.076). There
was no significant correlation between spasm laterality, preoperative

Copyright © 2011 by the American Clinical Neurophysiology Society 59

P. D. Thirumala et al. Journal of Clinical Neurophysiology • Volume 28, Number 1, February 2011

tonus, or type of decompressed vasculature and outcome at any of history of BT injections, or tonus/platysma involvement (Fig. 3).
the time points. Although no statistical difference existed with regard to preoperative
facial weakness, we observed an increasing proportion of patients
Intraoperative Lateral Spread Monitoring with greater degrees of preoperative paresis having postoperative
Categorizing LSR at Termination of MVD residual LSR (grade I, II: 15.5%; grade III: 17.0%; grade IV:
Data regarding intraoperative monitoring of the LSR during 40.0%). More patients requiring venous decompression had residual
MVD were available for 259 (87.4%) of the 293 patients. LSR on postdecompression LSR (51.9%) than did those without venous
other 12.3% of patients was collected but could not be located for involvement (42.0%) (P ⫽ 0.199). Reexploration patients (including
review at the time of this study. We divided these 259 patients into operations at our institution) were more likely to have LSR resolu-
two groups, according to the disappearance (group I: LSR ⫽ 0) or tion at MVD termination (85.1%) when compared with first-time
persistence (group II: LSR ⬎0) of facial EMG activity immediately operations (78.8%) (P ⫽ 0.327).
after decompression (see Appendix, Fig. A1). Demographic data
were not significantly different between groups (Table 4). LSR Outcomes: Short-Term Prognostic Value
disappearance (group I) was observed after facial nerve decompres- Within group I, (patients in whom the LSR disappeared after
sion in 207 of 259 patients (79.9%). Fifty-two patients (20.1%) had decompression), spasms completely disappeared postoperatively in
residual LSR postoperatively (group II). Five of these patients had 195 of 207 patients; LSR monitoring therefore had a negative
an LSR amplitude increase when compared with baseline. predictive value (NPV; proportion of patients without residual LSR
There was no statistically significant difference between that are spasm-free) of 94.7%. In contrast, only 67.3% of group II
group I and II with regard to laterality, compressing vasculature, patients (patients in whom residual LSR was present after decom-
pression) experienced immediate postoperative relief; the positive
predictive value (proportion of patients with residual LSR that have
TABLE 4. Patients With Measured LSR: Demographics persistent spasm) was 32.7% (Fig. 4). The specificity and sensitivity
of intraoperative LSR monitoring for predicting postoperative sur-
Group I: Group II: gical outcome were 60.7% and 84.8%, respectively (see Appendix,
Parameter LSR ⴝ 0 LSR >0 P
Table A3).
No. patients 207 52 — There was a statistically significant difference in the postop-
Gender (M:F) 72:135 19:33 0.813 erative spasm relief outcomes between the two groups (P ⬍ 0.0001),
Age 0.638 a trend which was also evident at discharge (94.2% vs. 76.9%) (P ⫽
Mean age (years) 51.9 52.8 0.001). Five group II patients with postoperative spasm had com-
No. patients aged ⬍50 years 85 19 plete relief by discharge. When 32 patients in group II were
No. patients aged ⱖ50 years 122 33
subdivided by whether the residual LSR was present throughout the
entire operation (ND), or returned after initially disappearing during
Operation 0.327
the operation (AD), no statistically significant difference in out-
First-time 167 45 comes was identified. All group II patients with follow-up data
Reexploration 40 7 whose LSR reappeared after initially resolving during the decom-
Mean time to FUP (months) 57.1 51.1 — pression (AD) described complete spasm relief at follow-up. Of six
Residual LSR classification — group II patients whose residual LSR was persistent during the
After disappearing — 23 entirety of the decompression (ND), one patient declared persistent
Never disappearing — 9 spasm at follow-up.
When separating patients in groups I and II by BT history or
LSR, lateral spread response; FUP, follow-up.
gender, we found no differences from the outcomes seen when

FIGURE 4. Clinical characteristics of

patients without residual lateral
spreads (group I: lateral spread re-
sponse [LSR] ⫽ 0) and with residual
lateral spreads (group II: LSR ⬎0) dur-
ing intraoperative electromyography.
AICA, anterior inferior cerebellar ar-
tery; PICA, posterior inferior cerebellar
artery; VA, vertebral artery.

60 Copyright © 2011 by the American Clinical Neurophysiology Society

Journal of Clinical Neurophysiology • Volume 28, Number 1, February 2011 Microvascular Decompression for Hemifacial Spasm

FIGURE 5. MVD Outcomes at three major time points. A, Outcomes in patients without prior Botox treatment. B, Out-
comes in patients with prior Botox treatment. C, Operative outcomes with respect to residual LSR status. ***P ⬍ 0.0005;
**P ⬍ 0.005, when compared with patients without residual LSR.

TABLE 5. Logistic Regression Analysis of the Association Between Perioperative Risk Factors and Spasm Persistence
Postoperative Discharge Follow-Up
OR (95% CI) P OR (95% CI) P OR (95% CI) P
Residual LSR 9.59 (3.73–24.65) ⬍0.0001 5.50 (2.21–13.73) ⬍0.001 1.03 (0.19–5.43) 0.977
Gender 0.75 (0.28–2.03) 0.572 0.43 (0.14–1.27) 0.127 0.13 (0.02–1.07) 0.058
Age ⱖ50 years 1.93 (0.71–5.21) 0.197 1.04 (0.42–2.61) 0.931 1.37 (0.36–5.15) 0.643
Prior BT use 0.42 (0.15–1.18) 0.098 0.41 (0.15–1.12) 0.081 3.03 (0.34–26.92) 0.321
Platysmal spasm 5.46 (1.88–15.90) 0.002 2.74 (1.03–7.27) 0.043 6.29 (1.25–31.55) 0.025
Weakness (H-B III, IV) 1.86 (0.67–5.19) 0.238 1.65 (0.58–4.65) 0.347 0.21 (0.02–1.77) 0.150
Left-sided spasm 0.74 (0.29–1.91) 0.537 0.95 (0.38–2.38) 0.907 1.35 (0.38–4.79) 0.641
Reference group: gender (female), age ⬍50 years, no prior BT use, no platysmal spasm, H-B score 0/I/II (none or mild weakness), left-sided spasm.
LSR, lateral spread response; BT, botulinum toxin; OR, odds ratio; CI, confidence interval. Bold indicates statistical significance.

comparing groups I and II as a whole. In patients without previous
BT use (n ⫽ 68), group I subjects (no postoperative LSR) were
significantly more likely than those in group II (residual postoper-
ative LSR) to have complete relief within 24 hours of the operation
(96.3% vs. 50.0%) (P ⫽ 0.0001) and at discharge (94.4% vs. 57.1%)
(P ⫽ 0.0002). This association was also true for patients with
previous BT injections (n ⫽ 189) postoperatively, although with less
statistical significance (94.1% vs. 75.7%) (P ⫽ 0.002). However, the
outcomes of group I and II patients did not significantly differ at
discharge (94.1% vs. 83.8%) (P ⫽ 0.081) when considering prior
BT use (Fig. 5). Within the BT use subgroup, postoperative spasm
disappeared in three patients with residual LSR by the time of
discharge. When separating patients by gender, the association
between residual LSR and persistent spasm was strongest within 24
hours of the operation (men: P ⫽ 0.0017; women: P ⬍ 0.0001) and
remained significant for women at the time of discharge (P ⬍
0.001).
Outcomes: Long-Term Prognostic Value
At the time of follow-up phone call with 208 patients, the
outcomes between group I and II patients did not significantly differ
(93.3% vs. 94.4%) (P ⫽ 1.000), although the NPV of LSR moni-
toring for long-term outcomes was 93.3%. There was no significant
statistical difference even when specifically considering patients
without previous BT injections (P ⫽ 0.688), female patients (P ⫽
1.000), and patients with recent follow-up (⬍2 years) (P ⫽ 0.402).
Multivariate Logistic Regression Model
The association between predisposing factors and spasm
status at each time point was examined with a multivariate logistic
regression model (Table 5). Predicting variables controlled for
included residual LSR, gender, age (ⱖ50 years or ⬍50 years), prior
BT treatment, platysmal spasm, preoperative facial weakness (H-B
grade 0/I/II [absent/mild], or III/IV [moderate/severe]), and lateral-
ity (left or right). Postoperatively, persistent spasm was associated
with residual LSR (odds ratio [OR]: 9.59; 95% confidence interval
[CI]: 3.73–24.65; P ⬍ 0.0001) and preoperative platysmal spasm
(OR: 5.46; 95% CI: 1.88 –15.90; P ⫽ 0.002). At discharge, residual
LSR (OR: 5.50; 95% CI: 2.21–13.73; P ⬍ 0.001) and platysmal
involvement (OR: 2.74; 95% CI: 1.03–7.27; P ⫽ 0.043) were also
predictive of present spasm. However, at follow-up, only preoper-
ative platysmal spasm was associated with persistent spasm (OR:
6.29; 95% CI: 1.25–31.55; P ⫽ 0.025). Men tended to report spasm

Copyright © 2011 by the American Clinical Neurophysiology Society 61

P. D. Thirumala et al. Journal of Clinical Neurophysiology • Volume 28, Number 1, February 2011
less frequently than women (OR: 0.13; 95% CI: 0.02–1.07; P ⫽
0.058).
DISCUSSION
CN VII MVD is an effective treatment for HFS. In our
experience with 326 operations, postoperative outcomes, with re-
spect to demographic variables, concurred with the results of pre-
vious studies (Barker et al., 1995; Ishikawa et al., 2001; Lovely et
al., 1998) Men demonstrate a greater relief rate compared with
women, a trend that was especially apparent at the time of follow-
up, and patients aged 50 years and older at the time of operation
have a similar resolution rate to that of younger patients (Shin et al.,
1997). Although a majority of patients presented with left-sided
spasm, laterality did not affect outcome. Outcome was not contin-
gent on vessel type, although multiple vessels were frequently seen
compressing the nerve. Excellent outcomes were achieved in the
majority of patients within 24 hours of surgery (88.0% spasm-free).
Eight additional patients achieved complete relief during their inpa-
tient stay, increasing the resolution rate to 90.8% at discharge. Only
16 patients with follow-up (7.7%) had symptoms, demonstrating the
long-lasting effects of MVD. The delayed spasm resolution may be
attributed to the time required for remyelination of the damaged
area, as well as the return of normal excitability of the facial
motonucleus (Ishikawa, et al., 1997; Moller and Jannetta, 1985,a,b;
Yamashita et al., 2001).
Facial EMG monitoring aids in the perioperative diagnosis of
HFS and has been considered by us to be a valuable intraoperative
tool in ensuring a meticulous CN VII decompression. Pulsatile
compression at the CN VII RExZ leads to the LSR (Ishikawa et al.,
1996b; Jannetta et al., 1970; Moller and Jannetta, 1986). In line with
recent studies (Isu et al., 1996; Kong et al., 2007; Sekula et al., 2009;
Shin et al., 1997) demonstrating the usefulness of intraoperative
LSR monitoring, we found that the disappearance of LSR at the end
of MVD was predictive of spasm relief both immediately after the
surgery (NPV ⫽ 94.7%) and at discharge. This finding was also true
when focusing on individual genders and BT history, although the
short-term correlation was weaker for those with prior BT injections.
In light of the excellent surgical outcomes in ⬎90% of patients at
discharge, we believe that intraoperative monitoring is an effective
tool in identifying culprit vessels. Given the significant negative
predictive value of LSR monitoring, the surgeon can be reassured
that an adequate decompression has been achieved and avoid un-
necessary operation time and resultant complications, especially
when multiple vessels are involved. However, we did find that
several patients with residual LSR were spasm-free after the surgery,
reducing the positive predictive value of LSR monitoring for short-
term outcomes. This appeared to be more common when the nerve
was found to be compressed by veins.
In five patients with residual LSR and immediate postopera-
tive spasm, symptoms disappeared by the time of discharge. Several
more reported relief at follow-up, which detracted from the positive
predictive value of LSR for long-term outcomes. All patients with
follow-up data whose residual LSR reappeared after initially disap-
pearing (AD) were spasm-free at follow-up, whereas only one
patient with persistent intraoperative LSR (ND) had refractory
spasm. A returning LSR (AD) may be the factor of additional minor
vessel or dural involvement that resolves after hospital stay. Of note,
when the LSR is found to reappear after initially disappearing, the
surgeon will make the decision whether to continue surgery and look
for additional compressing vasculature. Therefore, the proportion of
patients with postoperative residual LSR at our center is compara-
tively less because several potential AD cases were resolved after
further investigation into the compression. A persistent intraopera-
tive LSR (ND), although not significantly predictive of spasm at
62
follow-up, could also be indicative of a different pathology for the
patient’s spasm or an incomplete decompression. It is hypothesized
that delayed LSR resolution is due to a variable duration for
restoration of CN VII firing thresholds or remyelination in different
patients (Goto et al., 2002; Huang et al., 1992; Ishikawa 2001; Li,
2005). In addition, BT-induced and postoperative facial weakness
can make it difficult to ascertain spasm status in patients with subtle
but persistent symptoms, detracting from the predictive value of the
LSR in these subgroups.
Some authors have described the significant predictive value
of the LSR for outcomes at 1-year follow-up (Kong et al., 2007;
Moller and Jannetta, 1987), whereas others question its value (Ha-
tem et al., 2001; Joo et al., 2008; Kiya et al., 2001). We, too,
investigated the predictive value of the LSR on long-term outcomes
in our series. Our follow-up data were collected from patients with
a mean follow-up period of 54.5 months, whereas prior studies
document relatively short-term follow-up, with a mean period ⬍1
year (Sekula et al., 2009). Although we did not find a significant
statistical correlation, LSR resolution was predictive of spasm relief
at follow-up (NPV ⫽ 93.3%). However, residual LSR does not
always correlate with a poor outcome because it may take several
months for nerve excitability to normalize.
Therefore, we recognize the importance of intraoperative
LSR monitoring and agree with prior studies that recommend
postoperative facial EMG testing. In addition to detecting when the
LSR is fully normalized, postoperative testing can confirm that
HFS-related complexes, synkinesis, and cross talk are eliminated
and can thus be helpful in determining prognosis of persistent
spasm, ascertaining recurrence, and planning for reexploration. Kim
and Fukushima (1984) showed that synkinesis remains in the orbic-
ularis oris and mentalis muscles on postoperative facial EMG
monitoring 10 days after surgery. In evaluating outcomes after MVD
for HFS, it is therefore important to continue observing patients with
persistent spasm and discuss the likelihood of delayed resolution
with the patient before considering reoperation. This can be difficult
in practice because the refractory symptoms can be debilitating to
the already anxious patient, and there is no guarantee of resolution
even after 1 year. Also, early reoperation has been found to correlate
with better outcomes compared with patients receiving late reopera-
tions in one study (Engh et al., 2005), but complication rates for
additional MVD operations need to be considered. Other institutions
have advocated waiting 1 to 2 years before considering reoperation
(Goto et al., 2002; Ishikawa et al., 2001; Li et al., 2005). Repeat
EMGs may provide more insight into the neurophysiologic status of
patients with persistent spasm and can be implemented before
reoperation.
The use of BT injections has long been advocated as a quick,
noninvasive treatment for HFS, as well as blepharospasm (Laskawi,
2008), dystonias (Benecke and Dressler, 2007), and various other
disorders involving muscle overactivity. However, BT frequently
results in facial paresis (Yamashita et al., 2002). Patients also tend
to become refractory to treatment, requiring larger and more fre-
quent toxin doses over time, resulting in recurrent spasms. It is
significant that almost three quarters of patients who underwent
MVD have tried and been unsuccessful with BT in the past. This
represents a larger proportion of patients who tried BT compared
with that of previous series (Yamashita et al., 2002), pointing to
increasing popularity. Although a statistically insignificant trend, the
observation that 15 of 16 patients expressing persistent spasm at
follow-up had received preoperative BT injections raises the issue
that this drug may predispose a patient to recurrent spasm. We also
found that severe preoperative facial weakness, which can occur
after repeated BT treatment (House-Brackmann grade III or greater),
is predictive of poor surgical outcome. To explain these findings, it
Copyright © 2011 by the American Clinical Neurophysiology Society
Journal of Clinical Neurophysiology • Volume 28, Number 1, February 2011
is possible that paretic muscles (including those with weakness
secondary to BT injections) have lower thresholds for firing during
a period of spasmodic activity. A recent study suggests the HFS
leads to mild facial nerve injury, which leads to greater vulnerability
to BT injections (Misawa et al., 2008).
In addition to preoperative paresis, involvement of the
platysma muscle is another indicator of long-standing and severe
spasm. Preoperative platysma muscle spasm was also found to
correlate with poor outcome within 24 hours of surgery. In the
univariate analyses, platysmal involvement did not correlate with
spasm at discharge or follow-up, suggesting that the normal firing
activity in different branches of CN VII is not immediately restored
after decompression. However, we found a significant association
between platysmal spasm and postsurgical spasm status when con-
trolling for other perioperative characteristics in the multivariate
analyses. Therefore, patients with these progressive symptoms are
strongly encouraged to seek neurologic consultation and consider
MVD to improve their prognosis after surgery.
Study limitations include the inherent bias in evaluating LSR
monitoring, which was actively being used to make intraoperative
decisions, along with the absence of a control group. In addition, the
“all-or-nothing” nature of the LSR categorization (LSR ⫽ 0 or LSR
⬎0) may account for some of the patients with persistent LSR but
resolved spasm; even 95% disappearance of LSR was considered
LSR persistence in this study. Finally, follow-up data may be
problematic because some patients may not be objective about their
clinical condition. Further investigations, which decrease the reli-
ance on self-interpretation of surgical results, would be beneficial.
CONCLUSIONS
MVD for HFS is an effective treatment that can offer perma-
nent symptom resolution. We identified preoperative clinical char-
acteristics that affect outcome after MVD. Pertinent findings of this
study include the following:
Copyright © 2011 by the American Clinical Neurophysiology Society
Microvascular Decompression for Hemifacial Spasm
1. Men expressed a greater relief rate compared with women at
follow-up (P ⫽ 0.01).
2. A BT treatment history, which was positive in three quarters of
patients, was associated with preoperative facial weakness
(P ⫽ 0.01) and was also common in patients reporting spasm
at follow-up (P ⫽ 0.076).
3. Preoperative facial weakness correlated with persistent spasm
within 24 hours of operation (P ⫽ 0.012) and at discharge
(P ⫽ 0.016), with a similar trend at the time of follow-up (P ⫽
0.056).
4. Preoperative platysmal spasm correlated with persistent spasm
within 24 hours of operation (P ⫽ 0.022) and exhibited a
similar trend at the time of follow-up (P ⫽ 0.069). When
controlling for other perioperative characteristics, we found a
significant association of preoperative platysmal involvement
with poor postoperative surgical outcome (OR: 5.46; 95% CI:
1.88 –15.90; P ⫽ 0.002), positive spasm status at discharge
(OR: 2.74; 95% CI: 1.03–7.27; P ⫽ 0.043), and persistent or
recurrent spasm at follow-up (OR: 6.29; 95% CI: 1.25–31.55;
P ⫽ 0.025). It is recommended that patients with progressive
symptoms (affecting multiple facial muscle groups) consider
MVD and undergo surgery before muscles of the lower face
and neck become involved.
5. LSR monitoring was predictive of surgical outcomes within 24
hours of operation (P ⬍ 0.0001) and at discharge (P ⫽ 0.001).
When we adjusted for clinical and demographic variables, we
found residual LSR to be significantly predictive of persistent
postoperative spasm (OR: 9.59; 95% CI: 3.73–24.65; P ⬍
0.0001) and spasm at discharge (OR: 5.50; 95% CI: 2.21–
13.73; P ⬍ 0.001).
6. Although LSR monitoring was not statistically associated with
long-term reported outcomes (P ⫽ 1.00), postoperative LSR
resolution was predictive of long-term spasm relief (NPV:
93.3%).
63
P. D. Thirumala et al. Journal of Clinical Neurophysiology • Volume 28, Number 1, February 2011

APPENDIX

FIGURE A1. LSR, lateral spread response; ND, never disap-

pearing; AD, after disappearing.

FIGURE A2. Common medications

of patients referred for microvascu-
lar decompression.

64 Copyright © 2011 by the American Clinical Neurophysiology Society

Journal of Clinical Neurophysiology • Volume 28, Number 1, February 2011 Microvascular Decompression for Hemifacial Spasm

TABLE A1. Microvascular Decompression Success Rate by TABLE A4. Reexploration Surgeries (University of
Gender-Age Subgroups Pittsburgh Medical Center)
Age (Years) Male (%) Female (%) No. Patients
n With Spasm
Postoperative
⬍50 86.4 95.3 Time since previous MVD (n ⫽ 29)
ⱖ50 91.1 88.7 ⬍1 month 16 2
P 0.676 0.163 ⬍1 year 8 0
Discharge ⱖ1 year 5 0
⬍50 90.9 98.4 Time since recurrence (n ⫽ 17)
ⱖ50 91.1 90.1 ⱕ1 week 9 1
P 0.999 0.065 ⬎1 week 8 0
Follow-up MVD, microvascular decompression.
⬍50 100.0 97.5
ⱖ50 96.9 90.4
P 0.999 0.228 TABLE A5. Prognosis of First-Time vs. Reexploration
Patients
First- Redo: Redo:
TABLE A2. Microvascular Decompression—Demographics Time Reexploration Other Within
Outcome Patient Patient P Institution UPMC P
Parameter Male Female
Postoperative 87.6 90.2 0.597 86.4 (19) 93.1 (27) 0.641
Mean age (years) 53.2 50.9
Discharge 91.6 86.3 0.283 90.9 (20) 82.8 (24) 0.684
No. patients postoperative 75 148
Follow-up 93.6 85.7 0.155 78.6 (11) 90.5 (19) 0.369
% patients with follow-up data (n) 68.0 (51) 66.2 (98)
Mean time to follow-up (months) 66.1 70.7 Values are presented as % or % (n).
UPMC, University of Pittsburgh Medical Center.
% with recorded (n) 84.0 (63) 89.2 (132)
% with repeat operations (n) 12.0 (9) 18.2 (27)

TABLE A3. Predictive Value of Intraoperative Lateral Spread

Monitoring on Spasm Relief During Microvascular
Decompression, by Time Point
Parameter Postoperative Discharge Follow-Up
␹ test
2
P ⬍ 0.0001 P ⫽ 0.001 P ⫽ 1.00
Sensitivity* 60.7 50.0 16.7
Specificity† 84.8 82.9 80.5
Positive predictive value‡ 32.7 23.1 5.6
Negative predictive value§ 94.7 94.2 93.3
Values are presented as %.
*Proportion of patients with postoperative spasm that exhibit residual LSR at end
of MVD.
†Proportion of patients without postoperative spasm that exhibit disappearance of
LSR at end of MVD.
‡Proportion of patients with residual LSR at end of MVD that exhibit postoperative
persistent spasm.
§Proportion of patients with disappearance of LSR at end of MVD that are without
postoperative spasm.
LSR, lateral spread response; MVD, microvascular decompression.

Copyright © 2011 by the American Clinical Neurophysiology Society 65

P. D. Thirumala et al. Journal of Clinical Neurophysiology • Volume 28, Number 1, February 2011
REFERENCES
Barker FG II, Jannetta PJ, Bissonette DJ, et al. Microvascular decompression for
hemifacial spasm. J Neurosurg. 1995;82:201–210.
Benecke R, Dressler D. Botulinum toxin treatment of axial and cervical dystonia.
Disabil Rehabil. 2007;29:1769 –1777.
Eekhof JL, Aramideh M, Speelman JD, et al. Blink reflexes and lateral spreading
in patients with synkinesia after Bell’s Palsy and in hemifacial spasm. Eur
Neurol. 2000;43:141–146.
Engh JA, Horowitz M, Burkhart L, et al. Repeat microvascular decompression for
hemifacial spasm. J Neurol Neurosurg Psychiatry. 2005;76:1574 –1580.
Esteban A, Molina-Negro P. Primary hemifacial spasm: a neurophysiological
study. J Neurol Neurosurg Psychiatry. 1989;49:58 – 63.
Frueh BR, Preston RA, Musch DC. Facial nerve injury and hemifacial spasm.
Am J Ophthalmol. 1990;110:421– 423.
Goto Y, Matsushima T, Natori Y, et al. Delayed effects of the microvascular
decompression on hemifacial spasm: a retrospective study of 131 consecu-
tive operated cases. Neurol Res. 2002;24:296 –300.
Haines SJ, Torres F. Intra-operative monitoring of the facial nerve during decompres-
sive surgery for hemifacial spasm. J Neurosurg. 1991;74:254–257.
Hatem J, Sindou M, Vial C. Intra-operative monitoring of facial EMG responses
during microvascular decompression for hemifacial spasm. Prognostic value
for long-term outcome: a study in a 33-patient series. Br J Neurosurg.
2001;15:496 – 499.
Hopf HC, Lowitzsch K. Hemifacial spasm: location of the lesion by electrophys-
iological means. Muscle Nerve. 1982;5:S84 –S88.
House JW, Brackmann DE. Facial nerve grading system. Otolaryngol Head Neck
Surg. 1985;93:146 –147.
Huang CI, Chen IH, Lee LS. Microvascular decompression for hemifacial spasm:
analyses of operative findings and results in 310 patients. Neurosurgery.
1992;30:53–57.
Ishikawa M, Nakanishi T, Takamiya Y, Namiki J. Delayed resolution of residual
hemifacial spasm after microvascular decompression operations. Neurosurgery.
2001;49:847–856.
Ishikawa M, Ohira T, Namiki J, et al. Electrophysiological investigation of
hemifacial spasm: F-wave of the facial muscles. Acta Neurochir (Wien).
1996a;138:24 –32.
Ishikawa M, Ohira T, Namiki J, et al. Abnormal muscle response (lateral spread)
and F-wave in patients with hemifacial spasm. J Neurol Sci. 1996b;137:109 –
116.
Ishikawa M, Ohira T, Namiki J, et al. Electrophysiological investigations of
hemifacial spasm after microvascular decompression: F waves of the facial
muscles, blink reflexes, and abnormal muscle responses. J Neurosurg.
1997;86:654 – 661.
Isu T, Kamada K, Mabuchi S, et al. Intra-operative monitoring by facial electro-
myographic responses during microvascular decompressive surgery for
hemifacial spasm. Acta Neurochir (Wien). 1996;138:19 –23.
Jannetta PJ, Hackett E, Ruby JR. Electromyographic and electron microscopic
correlates in hemifacial spasm treated by microsurgical relief of neurovas-
cular compression. Surg Forum. 1970;21:449 – 451.
Joo WI, Lee KJ, Park HK, et al. Prognostic value of intra-operative lateral spread
response monitoring during microvascular decompression in patients with
hemifacial spasm. J Clin Neurosci. 2008;15:1335–1339.
66
Kim P, Fukushima T. Observations on synkinesis in patients with hemifacial
spasm: effect of microvascular decompression and etiological considerations.
J Neurosurg. 1984;60:821– 827.
Kiya N, Bannur U, Yamauchi A, et al. Monitoring of facial evoked EMG for
hemifacial spasm: a critical analysis of its prognostic value. Acta Neurochir
(Wien). 2001;143:365–368.
Kong DS, Park K, Shin BG, Lee JA, Eum DO. Prognostic value of the lateral
spread response for intraoperative electromyography monitoring of the facial
musculature during microvascular decompression for hemifacial spasm.
J Neurosurg. 2007;106:384 –387.
Kuroki A, Itagaki S, Nagai O. Delayed facial palsy after microvascular decom-
pression for hemifacial spasm. Facial Nerve Res. 1991;11:147–150.
Laskawi R. The use of botulinum toxin in head and face medicine: an interdis-
ciplinary field. Head Face Med. 2008;4:1– 8.
Li CS. Varied patterns of postoperative course of disappearance of hemifacial
spasm after microvascular decompression. Acta Neurochir (Wien). 2005;147:
617– 620.
Lovely TJ, Getch CC, Jannetta PJ. Delayed facial weakness after microvascular
decompression of cranial nerve VII. Surg Neurol. 1998;50:449 – 452.
Moller AR, Jannetta PJ. On the origin of synkinesis in hemifacial spasm: results
of intracranial recordings. J Neurosurg. 1984;61:569 –576.
Moller AR, Jannetta PJ. Hemifacial spasm: results of electrophysiologic recording
during microvascular decompression operations. Neurology. 1985a;35:969 –
974.
Moller AR, Jannetta PJ. Microvascular decompression in hemifacial spasm:
intraoperative electrophysiological observations. Neurosurgery. 1985b;16:
612– 618.
Moller AR, Jannetta PJ. Synkinesis in hemifacial spasm: results of recording
intracranially from the facial nerve. Experentia. 1985c;41:415– 417.
Moller AR, Jannetta PJ. Physiological abnormalities in hemifacial spasm studied
during microvascular decompression operations. Exp Neurol. 1986;93:584 –
600.
Moller AR, Jannetta PJ. Monitoring facial EMG responses during microvascular
decompression operations for hemifacial spasm. J Neurosurg. 1987;66:681–
685.
Nielsen VK. Pathophysiology of hemifacial spasm: I. Ephaptic transmission and
ectopic excitation. Neurology. 1984;34:418 – 426.
Nielsen VK. Electrophysiology of the facial nerve in hemifacial spasm: ectopic/
ephaptic excitation. Muscle Nerve. 1985;8:545–555.
Poignonec S, Widaihet M, Lamas G, et al. Electrophysiological evidence for
central hyperexcitability of facial motoneurons in hemifacial spasm. Eur
Arch Otorhinolaryngol. 1994;S216 –S217.
Ruby JR, Jannetta PJ. Hemifacial spasm: ultrastructural changes in the facial
nerve induced by neurovascular compression. Surg Neurol. 1975;4:369 –
370.
Sekula RF Jr, Bhatia S, Frederickson AM, et al. Utility of intraoperative electro-
myography in Microvascular decompression for Hemifacial spasm: a meta-
analysis. Neurosurg Focus. 2009;27:E10.
Shin JC, Chung UH, Kim YC, Park CI. Prospective study of microvascular
decompression in hemifacial spasm. Neurosurgery. 1997;40:730 –735.
Yamashita S, Kawaguchi T, Fukuda M, et al. Lateral spread response elicited by
double stimulation in patients with hemifacial spasm. Muscle Nerve. 2002;
25:845– 849.
Copyright © 2011 by the American Clinical Neurophysiology Society