Microvascular decompression surgery:surgical principles and technical nuancesbased on 4000 cases

Jun Zhong, Jin Zhu, Hui Sun, Ning-Ning Dou, Yong-Nan Wang,Ting-Ting Ying, Lei Xia, Ming-Xin Liu, Bang-Bao Tao, Shi-Ting Li

Department of Neurosurgery, Xinhua Hospital (The Cranial Nerve Disease Center of Shanghai), Shanghai JiaoTong University School of Medicine, China

Background: As an etiological treatment of trigeminal neuralgia (TN) and hemifacial spasm (HFS),microvascular decompression (MVD) has been popularized around the world. However, as a functionaloperation in the cerebellopontine angle (CPA), this process can be risky and the postoperative outcomesmight not be good enough sometimes.Objective: In order to obtain a better result with less complication, this surgery should be further addressed.Methods: With experience of more than 4000 MVDs, we have gained knowledge about the operativetechnique. Through abundant intraoperative photos, each step of the procedure was demonstrated indetail and the surgical strategy was focused.Results: The principle of MVD is to separate the nerve-vessel confliction rather than isolate it withprostheses. A prompt identification of the conflict site is important, which hinges on a good exposure. Asatisfactory working space can be established by an appropriate positioning of the patient’s head and aproper craniectomy as well as a rational approach. A sharp dissection of arachnoids leads to a maximalvisualization of the entire intracranial course of the nerve root. All the vessels contacting the trigeminal orfacial nerve should be treated. Intraoperative electrophysiological mentoring is helpful to distinguish theoffending artery for hemifacial cases.Conclusion: MVD is an effective treatment for the patient with TN or HFS. Immediate relief can be achievedby an experienced neurosurgeon with good knowledge of regional anatomy. A safe surgery is the tenet ofMVD, and accordingly, no single step of the procedure should be ignored.

Keywords: Microvascular decompression, Surgical technique, Hemifacial spasm, Trigeminal neuralgia

IntroductionTrigeminal neuralgia (TN) and hemifacial spasm

(HFS) account for the majority of cranial nerve

diseases.1 The compression of the V or VII cranial

nerve root by vessel(s) has been believed to be the

etiology. It was hypothesized that an ephaptic

transmission occurred between those individual nerve

fibers or the excitability threshold of the neuron

dropped following the vascular compression of

the nerve root.2–7 Our preliminary study implied that

the autonomic nervous system in the adventitia of the

offending vessel may contribute to the disorder.7,8

Carbamazepine may relieve the symptoms for the

patients with TN or HFS in the early stage of the

disease.9–12 When patients are weaned off medications,

they may try balloon techniques,13 glycerol injections,14

stereotactic radiosurgery with Gamma Knife,15 or

Botox injections.2 Nevertheless, in spite of being less

invasive, all these above modalities are destructive at

the cost of partial nerve function loss and may leave

facial numbness or palsy.9,16–22 However, as a non-

destructive surgery, microvascular decompression

(MVD), has nowadays become the most effective

remedy for TN or HFS since it was first introduced

by Dandy in 193423 and then popularized by

Janetta.24 Although the efficacy of MVD has been

reported to be more than 90%,18 not all institutes

reported perfect and consistent results.25–27 To date,

we have performed 4000 MVDs in China.18 Most of

the surgeries were performed in recent years, and the

relief rate is increasing continuously while the

complication is decreasing. Therefore, we believe

that, theoretically, as soon as the offending vessel(s) is

(are) removed away from the nerve, the symptoms of

Correspondence to: Shi-ting Li, Department of Neurosurgery, XinhuaHospital (The Cranial Nerve Disease Center of Shanghai), Shanghai JiaoTong University School of Medicine, Shanghai 200092, China. Email:zhongmdphd@126.com

882� W. S. Maney & Son Ltd 2014DOI 10.1179/1743132814Y.0000000344 Neurological Research 2014 VOL. 36 NO. 10

the patient should be gone immediately. The

improvement mainly depends on the surgeons’

understanding, experience, and technical nuances.

An unsuccessful MVD may be largely attributable to

misidentification of the culprit vessel. Although it was

claimed that endoscopy offers enhanced visualization

to identify offending vessels,28 it remains to be a sort

of assistance to MVDs. However, a good working

space may finally be obtained anyway with a fine

dissection. Apart from a meticulous manipulation

under microscope, a successful MVD depends on

many other factors, including approach, craniect-

omy, positioning, and so on. Therefore, every single

step of the process needs to be emphasized in order to

achieve an excellent postoperative outcome. In this

paper, we would like to present our experience with

emphasis on surgical principles and technical nuances

rather than postoperative outcomes, which have been

mentioned much in our previous papers.27

ImagingPreoperative imaging study is used for exclusionof mass lesions rather than identification of theoffending vesselThe disease is diagnosed in the light of the character-

istic symptom, while iconography is a supplementary

entity that is valuable in finding secondary TN or

HFS. Magnetic resonance imaging (MRI) with three-

dimensional time-of-flight (3D-TOF) sequence is recom-

mended, which may provide accurate information on

the relative locations of the nerve and the vessel

(Fig. 1A). Our previous study has observed that a

tortured vertebrobasilar artery deviated to the sympto-

matic side in 86.4% of HFS patients (Fig. 1B).29

However, this phenomenon is not so common in TN

patients. Actually, the main function of MRI is to

exclude the neoplasmal pathologies in the cerebellopon-

tine angle (CPA), e.g., meningiomas, acoustic neuromas,

cholesteatomas, etc. The occurrence of these neoplasms,

which in all intracranial tumors mentioned above, were

Figure 1 Preoperative magnetic resonance imagings (MRI). (A) MRI with three-dimensional time-of-flight (3D-TOF) sequence

showed an artery contacting the left facial nerve root. (B) A tortured vertebrobasilar artery deviated to the symptomatic side in a

hemifacial case. (C) It was a case of secondary hemifacial spasm (HFS). The space of left cerebellopontine angle (CPA) was

bigger with different signal intensity compared to the right side, which delineated a cholesteatoma.

Zhong et al. Microvascular decompression surgery

Neurological Research 2014 VOL. 36 NO. 10 883

about 13–26%,30 8–10%,31 and 3–10%,32 respectively,

according to the literature. It should be noted that

cholesteatoma is apt to be missed in the imaging before

posterior fossa exploratory surgery (Fig. 1C).33 Com-

puted tomography (CT) scan is an alternative when

MRI is not applicable because of metal implants.

IndicationsMicrovascular decompression is appropriate formost TN or HFS patientsGenerally, MVD is indicated for all the patients

suffering from drug-resistant TN or HFS as long as

their general conditions do not contraindicate general

anesthesia. Drug-resistant TN or HFS is defined as

uncontrolled pains or spasms when more than

1200 mg per day of carbamazepine is needed or

unacceptable complications occur. We would like to

point out that old age is not a contra-indication for

MVD surgery. Instead, it is much easier to operate on

old patients because of wide subarachnoid space as a

result of brain atrophy; only those with decompen-

sated dysfunction of vital organs (heart, lungs,

kidneys, or liver) should be evaluated cautiously. In

our cohort, the patient age at the time of surgery

ranged from 8 to 93 years without significant

difference in gender and side; the length of sympto-

matic illness ranged from 3 months to 22 years. We

suggest performing the surgery in the early stage

before the patient’s quality of life is awfully

influenced. Especially for those undernourished

because of reduced eating to avoid an attack of

severe pain induced by oral movement, a prompt

surgery is encouraged.

PositionA proper positioning contributes to a satisfactoryexposureGeneral anesthesia is the best option for MVD

patients. We place the patient in a park bench

position (3/4 lateral prone decubitus). This position

is superior to supine or full prone position because it

obviates the need to turn the patient’s head into an

uncomfortable position and therefore decreases the

risk of postoperative neck pain. This is especially

important for obese patients who usually have

generous supraclavicular fat pads. It is necessary to

point out that the contralateral shoulder should be

close to the edge of the bed so that the surgeon can

easily reach the surgical site. Meanwhile, the ipsilat-

eral shoulder should be slanted forward and pulled

away from the head by a shoulder belt so as to create

a satisfactory working space, which facilitates the

instruments getting in and out of the surgical field. A

fixation frame is used to hold the patient’s head stable

and make it possible to apply the retractor system

when necessary during the surgery, though we have

not used it even in cases of secondary TN or HFS

resulting from CPA masses since 2010, when we

began to choose an oblique position for the patients’

head (Fig. 2A). This inclined position with the

patient’s head turning back 15–20u from the level

surface facilitates the cerebellum to fall away under

its own gravity from the petrosal bone and obviates

the need of retractors (Fig. 2B). Actually, a good

exposure is achieved by rational positioning rather

than retracting. Brain relaxation methods such as

mannitol/hypertonic saline/hyperventilation are unne-

cessary in MVD surgeries, especially in old patients,

since cerebrospinal fluid (CSF) removal from the

cisterns works well. For TN cases, if a very steep

tentorium was showed after dural opening, the rostral

end of the table could be lowered a bit.

As the approach trajectory is somewhat different

between TN and HFS, the patient’s position varies

accordingly. As the cranial nerve VII is behind the

VIII in the surgical field, the axis of the head should

be placed 15–20u downward to expose the proximal

aspect of the facial nerve in HFS cases (Fig. 2C). In

TN cases, the axis is level.

IncisionThe incision is not ‘the smaller, the better’, buttoo large is not actually good for exposureDespite transversal, arc, or reverse ‘U’ shape incisions

having been reported,34 we choose a vertical linear

incision. It is laterally parallel to the hairline and

crosses the inion-zygomatic line with 1/3 above and 2/

3 below for TN cases and a bit lower for HFS cases

(Fig. 3). Nowadays, we have adopted a mastoid

retractor to hold the incision and no scalp clip is

needed. Owing to the limitation of the retractor’s

open angle, a longer than 7 cm incision is not really

necessary. To save a good blood supply, undue

coagulation should be avoided and a quick retraction

of the incision is recommended.

CraniectomyThe craniectomy should be as lateral as possibleA craniectomy of 3 cm in diameter is enough for

most cases. The edge of the sigmoidal sinus should be

exposed for both the TN and HFS cases to ensure an

ideal surgical corridor. Particularly, the craniectomy

should be rostral enough to the transverse sinus for

TN cases while low enough to expose the caudal

nerves for HFS cases. To avoid dural sinus injury, we

prefer craniectomy with pneumatic drill and Kerrison

rongeur to craniotomy with milling cutters. Bone

dust and chips were preserved for the later cranio-

plasty. The bone over the sigmoid sinus should be

removed in small pieces. Bone wax is effective for

homeostasis at the edge of dural sinuses. To obtain a

good working angle, the mastoid antrum could be

opened if necessary, but it should be immediately

waxed to prevent infection and CSF leak.

Zhong et al. Microvascular decompression surgery

884 Neurological Research 2014 VOL. 36 NO. 10

Figure 2 Positioning. (A) The patient is placed in a 3/4 lateral prone decubitus position with the ipsilateral shoulder slanted

forward (%a) and pulled away from the head. The patient’s head is turned back 15–20u from the level surface (%b). (B) The

sketch exhibits the benefits of an inclined head position. It offers a bigger visualizable area (B) than a flat position (A) does.

Zhong et al. Microvascular decompression surgery

Neurological Research 2014 VOL. 36 NO. 10 885

DurotomyThe dural opening is tailored to a maximalexposure of the conflict siteThe dural opening is tailored to the anatomy of each

case to allow a maximal exposure, which differs

between TN and HFS cases. We prefer to make a

wide ‘Y’ shape cutting with its tip pointing to the

junction of the transverse and sigmoid sinus for TN

cases, while to the lower outer quadrant in the

surgical area for HFS cases. The end of the durotomy

should be 1–2 mm away from the bone edge, so that a

gutter is built along the edge of craniotomy after the

dura flap is reflexed, which maintains the surgical

filed clear from potential outside bleeding. After the

dural mater is sutured back with double knotting, the

suture thread remains in place (without cutting off),

which facilitates tightening when necessary during the

procedure. This pattern of dural opening leaves the

majority of the dura on the cerebellum, which avails

the protection of the cerebellar hemisphere during the

process. To avoid shrinkage of the dura due to the

heat generated by the operating microscope’s lamp

aimed at the surgical field during the intradural

portion of the operation, pieces of wet gelforms are

placed over the dura.

ExposureA good exposure is obtained by sharpmicrodissection as well as proper positioningand craniectomy rather than harsh retractionWith analysis of our data, we found that approxi-

mately 4.5% of the patients had no symptomatic

relief postoperatively,18 and the main reason of

a failed surgery was that the exact offending vessel(s)

was(were) not recognized or the neurovascular con-

tact site was inaccessible intraoperatively. So we

believe a full exposure of the trigeminal or facial

nerve root is the key to obtaining a good result.

Usually, an unhurried suction drainage of CSF and

an ample adhesiolysis are effective enough to achieve

brain relaxation and no mannitol or lumbar puncture

is needed for most of the cases. We do not use

retracting blades because a narrow suction tube allows

more mobility and can actually afford more working

space than a wider spatula does during the operation at

a moment when a specific area is dissected. As a matter

of fact, with a good knowledge of the regional

anatomy, one does not have to visualize the whole

area of CPA while operating at a particular site (but

those surrounding structures should always be in

mind). Basically, an alternate use of a pair of the

microinstruments, i.e., a Fukushima teardrop suction

tube, a microscissors, a microdissector, and a bipolar

coagulation forceps, are enough to complete all the

intracranial manipulation. Instead of the ordinary

gun-shape forceps, a self-irrigating bipolar forceps

can keep the cottonoid over the cerebellum moist all

the time and avoid cerebellar contusion. The action

of clamping an artery should be avoided, which may

cause vasospasms. In that way, one can achieve more

effective exposure with less manipulation of the

cerebellum (Fig. 4).

Figure 3 Incision. A 7-cm vertical linear incision paralleling

laterally to the hairline is made. It crosses the inion-

zygomatic line with 1/3 above and 2/3 below for trigeminal

neuralgia (TN) cases and a bit lower for hemifacial spasm

(HFS) cases. This picture exhibited a trigeminal case.

Owing to its own gravity, the cerebellum falls away from the petrosal bone and provides more working space (D) than a flat

position (C) does. (C) For hemifacial cases, the axis of the head should be tilted down slightly, so that the facial nerve (VII) can

be easily seen from behind the acoustic nerve (VIII) (c1). Otherwise, it would be mostly covered by the VIII (c2).

Figure 4 Exposure. With a proper positioning and an

appropriate craniectomy, a satisfactory exposure is estab-

lished using a sharp microdissection of arachnoids. This was

a right trigeminal case. The complete intracranial course of

the trigeminal nerve root (V) from pons to Meckel’s cave was

visualized.

Zhong et al. Microvascular decompression surgery

886 Neurological Research 2014 VOL. 36 NO. 10

ApproachA caudorostral approach is suggestedFor TN cases, the protection and management of

petrosal veins (PV) poses the main challenges and

risks during the operation.35,36 In early cases, we

approached the trigeminal nerve from the super-

olateral aspect of the cerebellum, and PV were

obstructing the access path. We used to sacrifice the

petrosal vein or its branches to prevent unintention-

ally tearing at its entry to the superior petrosal sinus,

which is disastrous and difficult to manage. In that

case, compression with gelfoam is the only way for

homeostasis, while coagulation only makes things

worse. However, when a gelfoam is used to cover the

bleeding point, working space becomes narrower and

the following process can be very difficult. Our

previous study showed that sacrifice of the petrosal

vein may lead to a risk of delayed intracerebellar

hemorrhage, which was fatal unless timely identified

and treated.37 To detour off PV, we suggest a new

entry corridor. Recently, we started to select a

caudorostral approach (via cerebellar fissures),

through which the root entry/exit zone (REZ) of

trigeminal nerve could be accessed directly without

sacrifice of PV (Fig. 5a).

For HFS cases, the dissection should also start

from the lower cranial nerves with sharp opening of

the surrounding arachnoid, while the cerebellum is

superomedially retracted by a suction tube. The

retraction is most effective when it is parallel to the

VII/VIII cranial nerves. Arachnoid membrane over

the VII/VIII nerve complex is then cut with micro-

scissors, while the cerebellum is further retracted

medially until the pontomedullary sulcus is visualized

(Fig. 5B). The REZ of nerve VII is anterior and

slightly inferior to nerve VIII and may be directly

visualized upon gentle elevation of nerve VIII using a

fine dissector.

Identification of the Neurovascular ConflictionA full-way inspection of the nerve root increasesthe chance to find the culpritApart from I and II, the other cranial nerves are

composed of a central nervous system segment and a

peripheral nervous system segment. The central

segment has a structure similar to that of white

matter of the brain, consisting of parallel traveling

nerve fibers. It lacks funicular structure and is less

vascularized. Conversely, the winding density in the

peripheral segment is greater than in the central

segment, consisting of undulating nerve fibers, which

creates a more elastic and firmer structure. As a

result, the peripheral segment of the nerve is more

resistant to compression. The central segments of the

12 cranial nerves differ in length. Basically, motor

nerves have a shorter central segment than sensory

nerves.38 As the central segment of the trigeminal

nerve is longer than that of the facial nerve, logically,

the TN patients have more chances of lateral conflic-

tion than HFS patients.39–42 Hence, TN patients have

more chances of lateral confliction (Fig. 6A) than HFS

patients (Fig. 6B), which is in accordance with our

data (Table 1) as well as others.43 The neurovascular

confliction of TN patients is relatively complicated,

and a superior cerebellar artery (SCA) along the

shoulder of the REZ compressing the caudal side of

the trigeminal nerve ventromedially is common.9,19,44

The possibility of multiple offending vessels (arterial

and/or venous loops) should be excluded with careful

inspection (Fig. 6C). In our series, venous compres-

sion is not rare in TN patients.45 For HFS cases, the

very medial location of the contact site increases the

risk/chances of being missed by a poor exposure and

may thus lead to unsuccessful decompression and

failure of MVD. In accordance with others,46–50 we

found that the most frequent offending vessel was the

anterior inferior cerebellar artery (AICA) followed by

posterior inferior cerebellar artery (PICA) or vertebral

artery.41 Sometimes, the culprit could be those

Figure 5 Approach. A caudorostral approach is appropriate

for both hemifacial and trigeminal cases. (A) To detour off

petrosal veins (PV), dissection begins from petrosal fissure

and then extends rostrally for trigeminal cases. (B) For

hemifacial cases, dissection begins from caudal cranial

nerves and then extends rostrally.

Zhong et al. Microvascular decompression surgery

Neurological Research 2014 VOL. 36 NO. 10 887

arterioles (Table 2) (Fig. 6E). Mobilization of the

arterial loop often discloses a site of discoloration (or

even an indent) along the nerve, which confirms that

the intended pathology is found and predicts a good

outcome postoperatively (Fig. 6D).36 Accordingly, for

TN patients, we would check the entire root from pons

distally, while focusing on the medial portion of the

facial nerve for HFS patients. Nevertheless, to achieve

Figure 6 Identification of the offending vessel. (A) For trigeminal cases, the neurovascular conflict could be anyway along the

nerve root. In this case, the trigeminal nerve (V) was compressed rostrally by superior cerebellar artery (SCA) in cisternal

portion, which was separated by Teflon (T). (B) For hemifacial cases, the offending artery (*) is often discovered in the axilla

(caudal root exit zone) of the hemifacial nerve (VII). (C) Sometimes the nerve is compressed by multiple vessels. In this case, the

trigeminal nerve (V) was compressed by the SCA rostrally and by petrosal veins (PV) dorsally. (D) Sometimes an indent or

discoloration could be found in the nerve when an arterial loop is removed, which confirms that the intended pathology is found

and is predictive of a good outcome postoperatively. (E) Arterioles could also be the culprit. Note that one (a) of the arterioles

had been separated by Teflon in this left hemifacial case. These arterioles (a) looked redder in color compared to those veins

around under microscope. REZ: the root entry zone of trigeminal nerve; AICA: anterior infracerebellar artery; VIII: the

vestibulocochlear nerve; IX-X-XI: caudal cranial nerves.

Zhong et al. Microvascular decompression surgery

888 Neurological Research 2014 VOL. 36 NO. 10

a high cure rate, it is recommended that the whole

intracranial nerve course be a thorough exposed.

DecompressionThe principle of MVD is separation of the nerve–vessel confliction rather than isolation withprosthesis between themAs a matter of fact, a skillful driving of a micro-

dissector in assistance with a fine suction tube by the

operator’s two hands are enough to complete the

decompression process, including removal of the off-

ending vessel(s) and placement of Teflon felts piece by

piece, and no other instruments controlled by a third

hand are needed in such a small room. Moving the

vessel with forceps is not recommended. The sub-

stance of decompression should be separation of the

neurovascular confliction rather than isolation with

prosthesis between the nerve and the vessel (Fig. 7).

Actually, the role of the Teflon is to keep the

offending vessel from rebounding, and therefore it

is unnecessary to be always put at conflict site.

Intraoperative MonitoringElectrophysiological monitoring could be used topredict a satisfactory decompression for HFScasesElectrophysiological monitoring has been routinely used

in our department for years, and it has been proved to be

a good way to predict the results for HFS cases. Short

acting neuromuscular junction blocking medications

were used for intubation. No additional paralytic agent

was administered during electromyography (EMG)

monitoring. Our published data demonstrated that the

abnormal muscle response (AMR) vanished after the

decompression in 93.9% of the relieved HFS patients,51

which was in accordance with others.52 However, AMR

is not always reliable.53,54 Recently, we also stimulated

the offending artery directly during the operation for

HFS cases and achieved a wave analogous to the AMR

(Fig. 8). We named this new wave as ‘Z–L response’,which could only be recorded at the offending artery

and disappeared immediately after the decompression.7

To record the Z–L response, the same needle

electrodes as used in AMR recording is used. The

reference electrodes are inserted into the frontal

muscle, and the recording electrodes are inserted into

the orbicularis oculi, orbicularis oris, and mentalis

muscles. A non-invasive concentric electrode is

placed on the offending artery wall near the

compression site before it is detached from the facial

Table 1 The conflict site in our series

Cases REZ (%) non-REZ (%)

TN 52 48HFS 93 7

REZ: root entry/exit zone; non-REZ: the lateral (cisternal)segment of trigeminal or facial nerve root; TN: trigeminalneuralgia cases; HFS: hemifacial spasm cases.

Figure 7 Decompression. The substance of microvascular

decompression (MVD) process should be separation of the

neurovascular conflict rather than isolation with prosthesis

between them. This was a left trigeminal case. (A) The

superior cerebellar artery (SCA) was identified as the culprit,

which contacted with trigeminal nerve (V) rostrally. (B) The

offending artery (SCA) was separated from the nerve (V). In

case of rebound, a Teflon wadding (T) was placed between

them. Notice that the Teflon was not placed at the original

conflict site (*). It may work anyway as long as it could keep

the vessel from rebounding. VIII: the vestibulocochlear

nerve; PV: petrosal vein.

Table 2 The offending vessel(s) in our series

SCA (%) AICA (%) PICA (%) VA (%) Arterioles (%) Vein(s) (%)

TN 41 33 9 7 8 35HFS 0 36 52 9 13 1

TN: trigeminal neuralgia; HFS: hemifacial spasm; SCA: superior cerebellar artery; AICA: anterior inferior cerebellar artery; PICA:posterior inferior cerebellar artery; VA: vertebral artery; REZ: root entry/exit zone. Sometimes, more than a single vessel accounted forthe confliction.

Zhong et al. Microvascular decompression surgery

Neurological Research 2014 VOL. 36 NO. 10 889

nerve, and a square impulse (2 mA 6 0.2 ms) is

delivered. Z–L response monitoring is useful when an

AMR is not recorded before decompression or

persists after all vascular compressions are properly

treated. It could help neurosurgeons to determine the

real culprit when multiple offending vessels exist.

Accordingly, we suggest monitoring both AMR and

‘Z–L response’ during the MVD for HFS cases.

ClosureThe dura should be closed in a watertightpatternAt the end of the operation, the surgical field should

be well irrigated with 37uC normal saline to assure

that there is no bleeding as well as to vent air.

Meanwhile, attention should be paid to make sure the

implanted Teflon is stable under the flow of CSF.

Before dural closure, it is suggested to place a whole

piece of wet gelfoam over the cerebellum, which

prevents epidural hemorrhage from assessing to the

subdural space as well as protects accidental injury to

the brain tissue while suturing. In order to prevent

CSF leakage, the dura should be closed with stitch as

tight as possible. The operator should begin the stitch

process from the lowest level while the assistant is

pulling the suture thread, which will protect blood

inflow from outside. Sometimes, it is difficult to

complete a watertight closure due to dura defect. In

our series, this situation used to occur in 25%

proximately, but it reduced to 10% after we adopted

the following measurements: (1) no dural coagulation

was used; (2) dura mater was quickly sutured back as

soon as it was opened; (3) wet gelforms were inserted

between the dura and the craniotomy rims. Besides,

we would use a muscle patch to repair in case of a

defect. Then, the dura is sealed with glue and covered

by an artificial dura plus bone chips to reconstruct

the normal anatomy. Finally, the muscles, subcuta-

neous tissues and the scalp are sutured layer by layer.

No drainage is required (Fig. 3).

ComplicationsGenerally, those complications caused by serious organic

lesions, such as cerebellar hemorrhage, abrupt hydro-

cephalus, facial palsy and numbness16,17,19–22,55–58

(Table 3) can be avoided by an experienced operator.

However, those mild complications, e.g., dizziness,

disequilibrium, nausea, and vomit, may still occur in

some cases (Table 4).9,16 It might be attributable to

cerebellar laceration or pneumocrania. Therefore, a good

approach with gentle sharp dissection is very important,

which avoids undue retraction of the cerebellum.

Meanwhile, a thorough irrigation with warm normal

saline before tying the last suture of the dura stitch (which

should be in the highest position) should not be ignored.

Occasionally, an ipsilateral facial palsy or lip herpes may

occur a couple of days postoperatively for HFS or TN

cases, respectively.

Management of Failed MVDsWe have analyzed failed MVDs and found the main

reason for the failure was misidentification of the

offending vessel followed by insufficient decompres-

sion (Table 5).35,41 It was generally caused because a

Figure 8 Electrophysiological monitoring. For an abnormal

muscle response (AMR) wave to be monitored exclusively in

patients with hemifacial spasm (HFS), electrophysiology has

been widely employed to predict a satisfactory decompres-

sion. Recently, we also stimulated the offending artery

during the operation for HFS cases and achieved a wave

analogous to the AMR. It could only be recorded at

the offending artery and disappeared immediately after the

decompression. In this case, the offending artery and the

conflict site was finally proved to be the anterior infracer-

ebellar artery (AICA) in the cisternal segment of the facial

nerve. Notice that the facial nerve was not visualized in this

picture due to the positional block of the acoustic nerve (VIII).

Table 3 Postoperative malaises

Items Literatures (%) Our data (%)

Face paresthesia 1–359,16,18,56,63 6.7Nausea/vomit 1–28.29 15.2Hearing change 0.87–169,16,18,56 2Headache 28.29 22.3Dizziness/vertigo 0.89–28.29,16,17 18.2Herpes labialis (TN) 1.99 2.1Balance disorder 118 1

TN: trigeminal neuralgia.

Table 4 Complications of microvascular decompressions(MVDs)

Items Literatures (%) Our data (%)

Cerebrospinalfluid (CSF) leak

0.89–1316–20,55,56,64,65 1.3

Incisional infection 3–755 0.4Meningitis 0.6–9.29,16,17,55, 0Brain infarctions 0.69 0.2Pulmonary embolism 0.69 0Facial palsy 0.6–8.79,17–22,57,58,66 0.7Deafness 0.89–8.216,17 0.9Cerebellar edema 555 3.2Hydrocephalus 4.655 2.1Cerebellar mutism 1.255 0.1Mortality 0–2.616,18,19 0.13

Zhong et al. Microvascular decompression surgery

890 Neurological Research 2014 VOL. 36 NO. 10

good anatomical angle had not been obtained during

the operation. It may happen when PV blocked the

surgical way or even when arteries were attached the

petrosal bone and made it very difficult to remove

the cerebellum (Fig. 9). As TN or HFS is a functional

rather than life-threatening disorder, we regard

‘safety’ as the priority of MVD. In case of difficult

exposure, we would end the operation if an apparent

offending vessel has been identified at the most

haunted site, especially when a dent has been

observed at the contact point (Fig. 7D). For those

patients who do not improve at all after the surgery,

we would review the operative video records care-

fully. As long as we find that not all the entire

intracranial root of the nerve has been exposed, we

would communicate with the relatives as well as the

patient. If the patient is ready to take the risk that

arises from additional exposure, we would redo

MVD. We are not against the possibility of ‘delayed

relief’ mentioned by many authors, especially in HFS

cases.59 However, we do not think it may happen in

patients whose symptom were completely unim-

proved or even deteriorated. It might happen when

multiple vessels are involved. Once the larger

involved artery is isolated, the symptoms may

marginally improve as the main problem has been

solved. For the smaller vessel, a little movement may

allow the lesions at the interfaces to repair over time.

With the restoring of both the epineurium and

adventitia, the nerve may finally be isolated from

the vessel. Therefore, if the patient hopes for an

immediate cure and there is a chance for it, why do

we still ask him or her to suffer while anxiously

expecting a possible relief? Our philosophy looks like

a sort of Eastern view–‘Middle Way’, but it does give

the patient more choice. With this surgical strategy,

we have achieved a good postoperative outcome with

very low incidence of complications.36,45,60–62

Disclaimer Statements

Contributors Dr Zhong and Dr RLi performed all the

MVD surgeries. The others assisted the operations

and collected data; Dr Zhong finished the manu-

script; and Dr Li supervised all the work.

Funding This study was supported by Science & Tech-

nology Committee of ShanghaiMunicipal (124119a0800)

and Education Commission of Shanghai Municipal

(10YZ42).

Table 5 Comparison of surgical findings between primary and reoperative MVDs in HFS patients (adapted from Ref. 41)

Patients

1st MVD 2nd MVD

The failure reasonof the 1st MVD

Offendingartery(ies)

Conflictsite (zone)

Severity ofcompression

Offendingartery(ies)

Conflictsite (zone)

Severity ofcompression

1 VA z PICA 2 and 3 Indentation VA z PICA 2 and 3 Indentation Teflon was not enough2 VA z AICA 3 Indentation AICA 4 Contact Zone 4 missed3 AICA 3 Indentation AICA 4 Contact Zone 4 missed4 AICA 2 and 3 Indentation AICA 4 Deviation Zone 4 missed*5 AICA 2 and 3 Indentation AICA 4 Deviation Zone 4 missed*6 AICA 2 and 3 Contact AICA 4 Contact Zone 4 missed#7 AICA 2 and 3 Contact AICA 4 Deviation Zone 4 missed*8 VA z AICA 2 and 3 Indentation AICA 2 Indentation Incomplete decompression‘9 AICA 2 and 3 Contact AICA 4 Deviation Zone 4 missed*

MVD: microvascular decompression; VA: vertebral artery; PICA: posterior inferior cerebellar artery; AICA: anterior inferior cerebellarartery.*AICA went through between facial and acoustic nerves.#A vein went through between facial and acoustic nerves.‘decompression between AICA and the nerve was not enough.Zone 1: where the nerve emerges to the brainstem surface from the parenchymal and goes through the pontomedullary sulcus; Zone2: where the root attaches to the surface of the pons; Zone 3: where it is gradually transiting to be narrower, which corresponds to theObersteiner–Redlich zone (the traditional REZ); Zone 4: where the nerve fibrins separates from the brainstem and extends to theinternal meatus.

Figure 9 An unsatisfactory exposure due to individual

anatomical difference. This was a left HFS case. The medial

portion of the facial nerve could not be exposed because the

anterior infracerebellar artery (AICA) was attached to the

petrosal bone (*) and made the retraction of cerebellar

hemisphere impossible. V: trigeminal nerve; VIII: the vesti-

bulocochlear nerve; IX: glossopharyngeal nerve.

Zhong et al. Microvascular decompression surgery

Neurological Research 2014 VOL. 36 NO. 10 891

Conflicts of interest None.

Ethics approval This study was approved by the

Ethics Committee of XinHua Hospital, Shanghai

JiaoTong University School of Medicine.

Reference1 Yang K-H, Na J-H, Kong D-S, Park K. Combined hyperactive

dysfunction syndrome of the cranial nerves. J KoreanNeurosurg Soc. 2009;46:351–4.

2 Rosenstengel C, Matthes M, Baldauf J, Fleck S, Schroeder H.Hemifacial spasm: conservative and surgical treatment options.Dtsch Arztebl Int. 2012;109:667.

3 Kuroki A, Møller A. Facial nerve demyelination and vascularcompression are both needed to induce facial hyperactivity: astudy in rats. Acta Neurochir (Wien). 1994;126:149–57.

4 Wu Y, Davidson AL, Pan T, Jankovic J. Asian over-representation among patients with hemifacial spasm comparedto patients with cranial-cervical dystonia. J Neurol Sci.2010;298:61–3.

5 Love S, Coakham HB. Trigeminal neuralgia pathology andpathogenesis. Brain. 2001;124:2347–60.

6 Devor M, Seltzer Z. Pathophysiology of damaged nerves inrelation to chronic pain. In: Wall PD, Melzack R, editors.Textbook of pain, Vol. 4. London: Churchill Livingstone; 1999.p. 129–64.

7 Zheng X, Hong W, Tang Y, Wu Z, Shang M, Zhang W, et al.Sympathetic nerves bridge the cross-transmission in hemifacialspasm. Neurosci Lett. 2012;517(1):52–5.

8 Zhou QM, Zhong J, Jiao W, Zhu J, Yang XS, Ying TT, et al.The role of autonomic nervous system in the pathophysiologyof hemifacial spasm. Neurol Res. 2012;34:643–8.

9 Oesman C, Mooij JJ. Long-term follow-up of microvasculardecompression for trigeminal neuralgia. Skull Base. 2011;21:313.

10 Espay AJ, Schmithorst VJ, Szaflarski JP. Chronic isolatedhemifacial spasm as a manifestation of epilepsia partialiscontinua. Epilepsy Behav. 2008;12:332–6.

11 Perkin G, Illingworth R. The association of hemifacial spasmand facial pain. J Neurol Neurosurg Psychiatry. 1989;52:663–5.

12 Lemos L, Alegria C, Oliveira J, Machado A, Oliveira P,Almeida A. Pharmacological versus microvascular decompres-sion approaches for the treatment of trigeminal neuralgia:clinical outcomes and direct costs. J Pain Res. 2011;4:233.

13 Campos WK, Linhares MN. A prospective study of 39 patientswith trigeminal neuralgia treated with percutaneous ballooncompression. Arq Neuropsiquiatr. 2011;69:221–6.

14 Wang X, Zhou C, Guang-Jian S, Min-Hui X, Guang-Xin C,Yong-Wen Z, et al. Long-term outcomes of percutaneousretrogasserian glycerol rhizotomy in 3370 patients withtrigeminal neuralgia. Turk Neurosurg. 2011;21(1):48–52.

15 Lee JK, Choi HJ, Ko HC, Choi SK, Lim YJ. Long termoutcomes of gamma knife radiosurgery for typical trigeminalneuralgia-minimum 5-year follow-up. J Korean Neurosurg Soc.2012;51:276–80.

16 Gunther T, Gerganov VM, Stieglitz L, Ludemann W, Samii A,Samii M. Microvascular decompression for trigeminal neur-algia in the elderly: long-term treatment outcome andcomparison with younger patients. Neurosurgery. 2009;65:477–82.

17 Illingworth R, Porter D, Jakubowski J. Hemifacial spasm: aprospective long-term follow up of 83 cases treated bymicrovascular decompression at two neurosurgical centres inthe United Kingdom. J Neurol Neurosurg Psychiatry.1996;60:72–7.

18 Zhong J, Li S-T, Zhu J, Guan HX, Zhou QM, Jiao W, et al. Aclinical analysis on microvascular decompression surgery in aseries of 3000 cases. Clin Neurol Neurosurg. 2012;114:846–51.

19 Jagannath PM, Venkataramana NK, Bansal A, RavichandraM. Outcome of microvascular decompression for trigeminalneuralgia using autologous muscle graft: a five-year prospectivestudy. Asian J Neurosurg. 2012;7:125.

20 Jeon C-J, Kong D-S, Lee J-A, Park K. The efficacy and safetyof microvascular decompression for hemifacial spasm in elderlypatients. J Korean Neurosurg Soc. 2010;47:442–5.

21 Han J-S, Lee J-A, Kong D-S, Park K. Delayed cranial nervepalsy after microvascular decompression for hemifacial spasm.J Korean Neurosurg Soc. 2012;52:288–92.

22 Kim HJ, Park YS, Ryu JS, Huh R, Han I, Shin DA, et al.Intraoperative facial electromyography and brainstem auditoryevoked potential findings in microvascular decompression forhemifacial spasm: correlation with postoperative delayed facialpalsy. Stereotact Funct Neurosurg. 2012;90:260–5.

23 Dandy WE. Concerning cause of trigeminal neuralgia. Am JSurg. 1934;24:447–55.

24 Jannetta PJ, McLaughlin MR, Casey KF. Technique ofmicrovascular decompression: technical note. NeurosurgFocus. 2005;18:1–5.

25 Kondo A. Microvascular decompression surgery for trigeminalneuralgia. Stereotact Funct Neurosurg. 2001;77:187–9.

26 Park JS, Kong DS, Lee JA, Park K. Hemifacial spasm:neurovascular compressive patterns and surgical significance.Acta Neurochir (Wien). 2008;150:235–41.

27 Zhong J. An ideal microvascular decompression techniqueshould be simple and safe. Neurosurg Rev. 2012;35:137–40.

28 Ishimori T, Nakano S, Kagawa M, Yokoe K, Togami T,Asakura H. Virtual endoscopic images by 3D FASE cisterno-graphy for neurovascular compression. Magn Reson Med Sci.2003;2:145–9.

29 Guan HX, Zhu J, Zhong J. Correlation between idiopathichemifacial spasm and the MRI characteristics of the vertebralartery. J Clin Neurosci. 2011;18:528–30.

30 Zimny A, Sasiadek M. Contribution of perfusion-weightedmagnetic resonance imaging in the differentiation of meningio-mas and other extra-axial tumors: case reports and literaturereview. J Neurooncol. 2011;103:777–83.

31 Palmisano S, Schwartzbaum J, Prochazka M, Pettersson D,Bergenheim T, Florentzson R, et al. Role of tobacco use in theetiology of acoustic neuroma. Am J Epidemiol. 2012;175:1243–51.

32 Albera R, Canale A, Piumetto E, Lacilla M, Dagna F.Ossicular chain lesions in cholesteatoma. Acta Otorhino-laryngol Ital. 2012;32:309.

33 Ilıca AT, Hıdır Y, Bulakbası N, Satar B, Guvenc I, Arslan HH,et al. HASTE diffusion-weighted MRI for the reliable detectionof cholesteatoma. Diagn Interv Radiol. 2012;18:153.

34 Cohen-Gadol AA. Microvascular decompression surgery fortrigeminal neuralgia and hemifacial spasm: nuances of thetechnique based on experiences with 100 patients and review ofthe literature. Clin Neurol Neurosurg. 2011;113:844–53.

35 Li S, Hong W, Tang Y, Ying T, Zhang W, Li X, et al. Re-operation for persistent hemifacial spasm after microvasculardecompression with the aid of intraoperative monitoring ofabnormal muscle response. Acta Neurochir (Wien).2010;152:2113–8.

36 Zhu J, Li ST, Zhong J, Guan HX, Ying TT, Yang M, et al.Role of arterioles in management of microvascular decompres-sion in patients with hemifacial spasm. J Clin Neurosci.2012;19:375–9.

37 Zhong J, Li S, Xu S, Wan L, Wang X. Management of petrosalveins during microvascular decompression for trigeminalneuralgia. Neurol Res. 2008;30:697–700.

38 De Ridder D, Moller A, Verlooy J, Cornelissen M, De RidderL. Is the root entry/exit zone important in microvascularcompression syndromes? Neurosurgery. 2002;51:427–33; dis-cussion 433–4.

39 Katusic S, Beard CM, Bergstralh E, Kurland LT. Incidence andclinical features of trigeminal neuralgia: Rochester, Minnesota,1945-1984. Ann Neurol. 1990;27:89–95.

40 Auger RG, Whisnant JP. Hemifacial spasm in Rochester andOlmsted County, Minnesota, 1960 to 1984. Arch Neurol.1990;47:1233–4.

41 Zhong J, Zhu J, Li ST, Li XY, Wang XH, Yang M, et al. Ananalysis of failed microvascular decompression in patients withhemifacial spasm: focused on the early reoperative findings.Acta Neurochir (Wien). 2010;152:2119–23.

42 Zhong J, Li ST, Zhu J, Guan HX. Is entire nerve rootdecompression necessary for hemifacial spasm? Int J Surg.2011;9:254–7.

43 Katusic S, Beard CM, Bergstralth E, Kurland LT. Incidenceand clinical features of trigeminal neuralgia, Rochester,Minnesota, 1945-1984. Ann Neurol. 1990;27:89–95.

44 Prieto R, Pascual JM, Yus M, Jorquera M. Trigeminalneuralgia: assessment of neurovascular decompression by 3Dfast imaging employing steady-state acquisition and 3D time offlight multiple overlapping thin slab acquisition magneticresonance imaging. Surg Neurol Int. 2012;3:50.

45 Zheng X, Feng B, Hong W, Zhang W, Yang M, Tang Y, et al.Management of intraneural vessels during microvascular

Zhong et al. Microvascular decompression surgery

892 Neurological Research 2014 VOL. 36 NO. 10

decompression surgery for trigeminal neuralgia. WorldNeurosurg. 2012;77:771–4.

46 Kawashima M, Yamada M, Sato S, Oka H, Fujii K,Matsushima T. Hemifacial spasm caused by vascular compres-sion of the distal portion of the facial nerve associated withconfiguration variation of the facial and vestibulocochlearnerve complex. Turk Neurosurg. 2009;19:269–75.

47 Campos-Benitez M, Kaufmann AM. Neurovascular compres-sion findings in hemifacial spasm. J Neurosurg. 2008;109:416–20.

48 Sindou M. Microvascular decompression for primary hemi-facial spasm. Importance of intraoperative neurophysiologicalmonitoring. Acta Neurochir (Wien). 2005;147:1019–26.

49 Yamashita S, Kawaguchi T, Fukuda M, Watanabe M, TanakaR, Kameyama S. Abnormal muscle response monitoring duringmicrovascular decompression for hemifacial spasm. ActaNeurochir (Wien). 2005;147:933–8.

50 Polo G, Fischer C, Sindou MP, Marneffe V. Brainstemauditory evoked potential monitoring during microvasculardecompression for hemifacial spasm: intraoperative brainstemauditory evoked potential changes and warning values toprevent hearing loss-prospective study in a consecutive series of84 patients. Neurosurgery. 2004;54:97.

51 Ying T-T, Li S-T, Zhong J, Li XY, Wang XH, Zhu J. The valueof abnormal muscle response monitoring during microvasculardecompression surgery for hemifacial spasm. Int J Surg. 2011;9:347–51.

52 Fukuda M, Oishi M, Takao T, Hiraishi T, Sato Y, Fujii Y.Monitoring of abnormal muscle response and facial motorevoked potential during microvascular decompression forhemifacial spasm. Surg Neurol Int. 2012;3:118.

53 Joo W-I, Lee K-J, Park H-K, Chough CK, Rha HK.Prognostic value of intra-operative lateral spread responsemonitoring during microvascular decompression in patientswith hemifacial spasm. J Clin Neurosci. 2008;15:1335.

54 Hatem J, Sindou M, Vial C. Intraoperative monitoring of facialEMG responses during microvascular decompression forhemifacial spasm. Prognostic value for long-term outcome: astudy in a 33-patient series. Br J Neurosurg. 2001;15:496–9.

55 Dubey A, Sung W-S, Shaya M, Patwardhan R, Willis B, SmithD, et al. Complications of posterior cranial fossa surgery – aninstitutional experience of 500 patients. Surg Neurol. 2009;72:369–75.

56 Jellish WS, Benedict W, Owen K, Anderson D, Fluder E, SheaJF. Perioperative and long-term operative outcomes after

surgery for trigeminal neuralgia: microvascular decompressionvs percutaneous balloon ablation. Head Face Med. 2008;4:11.

57 Dannenbaum D, Kuzmina E, Lejeune P, Torrie J, Gangbe M.Prevalence of diabetes and diabetes-related complications inFirst Nations communities in Northern Quebec (EeyouIstchee), Canada. Can J Diabetes. 2008;32:46–52.

58 Kalkanis SN, Eskandar EN, Carter BS, Barker FG II.Microvascular decompression surgery in the United States,1996 to 2000: mortality rates, morbidity rates, and the effects ofhospital and surgeon volumes. Neurosurgery. 2003;52:1251–62.

59 Goto Y, Matsushima T, Natori Y, Inamura T, Tobimatsu S.Delayed effects of the microvascular decompression on hemi-facial spasm: a retrospective study of 131 consecutive operatedcases. Neurol Res. 2002;24:296–300.

60 Zheng X, Hong W, Tang Y, Ying T, Wu Z, Shang M, et al.Discovery of a new waveform for intraoperative monitoring ofhemifacial spasms. Acta Neurochir (Wien). 2012;154:799–805.

61 Yang XS, Li ST, Zhong J, Zhu J, Du Q, Zhou QM, et al.Microvascular decompression on patients with trigeminalneuralgia caused by ectatic vertebrobasilar artery complex:technique notes. Acta Neurochir (Wien). 2012;154:793–7;discussion 797.

62 Zhong J, Zhu J, Li ST, Guan HX. Microvascular decompres-sions in patients with coexistent hemifacial spasm andtrigeminal neuralgia. Neurosurgery. 2011;68:916–20; discussion920.

63 Rhee DJ, Kong DS, Park K, Lee JA. Frequency and prognosisof delayed facial palsy after microvascular decompression forhemifacial spasm. Acta Neurochir (Wien). 2006;148(8):839–843.

64 Hyun SJ, Kong DS, Park K. Microvascular decompression fortreating hemifacial spasm: lessons learned from a prospectivestudy of 1,174 operations. Neurosurg Rev. 2010 Jul;33(3):325–34; discussion 334. doi: 10.1007/s10143-010-0254-9. Epub 2010Mar 27.

65 Park JS, Kong DS, Lee JA, Park K. Intraoperative manage-ment to prevent cerebrospinal fluid leakage after microvasculardecompression: dural closure with a ‘‘plugging muscle’’method. Neurosurg Rev. 2007 Apr;30(2):139–42; discussion142. Epub 2007 Jan 13.

66 Barker FG, Jannetta PJ, Bissonette DJ, Larkins MV, Jho HD.The long-term outcome of microvascular decompression fortrigeminal neuralgia. N Engl J Med. 1996 Apr 25;334(17):1077–83; discussion 142. Epub 2007 Jan 13.

Zhong et al. Microvascular decompression surgery

Neurological Research 2014 VOL. 36 NO. 10 893