Introduction

Hemifacial spasm (HFS) is a condition involving involuntary, repetitive, unilateral contraction of the muscles innervated by the facial nerve (cranial nerve CN VII) [1]. Typical HFS is caused by facial nerve irritation secondary to vascular compression at the root exit zone (REZ) [2, 3], leading to involuntary, intermittent spasms beginning at the orbicularis oculi muscle and progressing to the mentalis muscle. Retromastoid craniotomy and facial nerve microvascular decompression (MVD) have been proven to be effective cure for patients [4, 5]. However, wide variations in the cure rate for HFS have been reported ranging from 86 to 92 % [2, 6]. There is no agreement concerning what is the suitable time for the treatment of persistent or recurrent patients with HFS after the original MVD [7, 8]. So far, little has been published regarding repeat MVD for HFS in those patients who failed their initial operation [9–11], and less has been reported about patients who had late repeat MVD [9]. In present study, the main objective was to investigate characteristics, outcomes, and complications of the late repeat MVD (beyond 1 month after the original MVD) for persistent or recurrent HFS after initial failure of MVD.

Methods

Study design

A retrospective study was conducted with institutional review board approval from the University of Pittsburgh (IRB #: PR008120394). Of 276 in 293 patients with typical HFS, 33 patients with HFS underwent repeat MVD due to persistent or recurrent spasm (group I), and 243 patients underwent their first MVD (group II). In group I, 28 patients underwent reoperation for clinical failure or recurrence after original MVD, and five patients had two prior MVDs. The interval to repeat MVD ranged from 1.6 months to 16 years (Fig. 1). Clinical outcome data were obtained after the immediate operation, at discharge (mean, 4.00 ± 2.00 days), and at a follow-up phone call during June 2008. Follow-up data were collected from 195 patients who had a minimum follow-up period of 9 months (mean, 54.48 ± 27.84 months).

Fig. 1
figure 1

Number of patients and interval time of latest two MVDs

Full size image

Microvascular decompression

Between January 2000 and December 2007, we reviewed patients who had HFS and underwent MVD procedures at the University of Pittsburgh Medical Center (UPMC). The operation was performed using a routine retrosigmoid approach [12, 13]. Short-acting neuromuscular junction-blocking medications were used for intubation. No additional paralytic agent was administered during electromyography (EMG) monitoring. The dissection was started from the caudal cranial nerves, using withdrawal of CSF and gentle exposure of pertinent anatomy with handheld suction. Following careful exposure of the CN VII–CN VIII complex, any suspect arteries or veins compressing CN VII anywhere from the brainstem to beyond the REZ were treated. For redo MVDs, the previous conflict site was reexamined and checked to confirm whether a satisfactory decompression was completed. The operation was complete when the nerve no longer demonstrated any visible evidence of vascular compression. After confirming that there were no further offending vessels, the surgeon terminated the procedure and closed the craniotomy in a routine fashion.

Intraoperative neurophysiological monitoring and alarm criteria

During surgery, free run EMG (f-EMG) monitoring of the facial, glossopharyngeal, and vagus nerves was performed, in addition to lateral spread response (LSR). Lateral spread response is a delayed abnormal muscle response recorded in the mentalis muscle following the stimulation of zygomatic branch of the facial nerve [5]. All instances of f-EMG activity, regardless of type (spikes, bursts, neurotonic discharges), were made audible to and also immediately reported to the surgeon and recorded in the patient’s record. Auditory nerve function was monitored using brainstem auditory evoked potentials performed with our institution’s alarm criteria [14]. Physician oversight and interpretation were performed using a combined on-site and remote model utilized by UPMC [15].

Data analysis

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 (in percent). Group differences in demographic, clinical characteristics, and outcomes were assessed using χ 2 tests, and the Fisher exact test correction was used when needed. P < 0.05 was considered as statistically significant.

Results

Demographics

Two hundred seventy-six patients with HFS had a mean age of 52.17 ± 12.13 years (range, 17–82 years), with a female-to-male ratio being 1.9:1 and left-to-right ratio being 1.2:1. Medical and surgical histories were obtained from each patient undergoing MVD. Two hundred two patients (73.19 %) received prior botox treatment. Of the 276 patients, 33 underwent repeat MVDs, among which, 29 patients were referred to UPMC after prior operations elsewhere, and four patients had their first operation at UPMC. These patients underwent reoperation at UPMC beyond 1 month of their original decompression. No patients exhibited bilateral HFS.

Preoperative characteristics

There is no significant statistical difference in preoperative clinical characteristics (gender, age, side of MVD, botox usage, facial weakness, etc.) between the two groups (Table 1).

Table 1 Comparison of preoperative clinical characteristics between group I and group II
Full size table

Intraoperative findings

Compressing vasculature seen near facial nerve REZ

The vessels compressing the REZ, as identified by the surgeon, are summarized in Table 2. A majority (70.7 %) of patients had multiple compressing vessels. The compression was commonly caused by the anterior inferior cerebellar artery (AICA), posterior inferior cerebellar artery, vertebral artery, and some veins in both groups.

Table 2 Comparison of intraoperative findings including LSR, offending vessels, and outcomes between group I and group II
Full size table

Significantly higher number of patients had a vein as offending vessel intraoperatively in group I as compared to group II (P = 0.02; Table 2).

Intraoperative lateral spread monitoring

Data regarding intraoperative monitoring of the LSR during MVD were available for 255 (92.39 %) of the 276 patients. LSR disappeared in 66.67 % (20/30) of the patients in group I as compared to 65.33 % (147/225) in group II (P = 0.89; Table 2).

Operative outcomes

In 33 patients, 84.85 % experienced immediate postoperative relief of spasm, 90.91 % had relief at discharge, and 85.00 % had relief at follow-up period. No significant difference in spasm relief between two groups was observed at the immediate postoperative, discharge, or follow-up stages.

Postoperative complications

Postoperative complications were observed in both patient groups (Table 3). Complications included hearing loss, balance problem, CSF leakage, diplopia, headache, and dizziness/vertigo. There was no significant difference in complications between the two groups. Complication rates in repeat MVD were as follows: hearing loss 15.15 %, facial weakness 12.12 % (two cases of grade II, one grade III, and one grade IV, respectively), balance disorder 3.03 %, cerebrospinal fluid leak 3.03 %, diplopia 3.03 %, headache 6.06 %, and wound infection 6.06 %.

Table 3 Comparison of post-operative complications between Group I and Group II
Full size table

Discussion

HFS is caused by vascular compression of the REZ of the facial nerve [16, 17]. Microvascular decompression is a highly accepted and effective method for treatment of patients with HFS [1, 2, 18]. Much has been published regarding the high efficacy of MVD for HFS, with cure rates ranging anywhere from approximately 85 to 92 %, and excellent long-term results 5 years after MVD in 85 % of patients [7, 19]. Some authors have suggested the possibility of symptom relief within 1 year; the treatment of persistent or recurrent patients with HFS should be more than 1 year after the original MVD [6, 16, 20]. Sindou et al. [7, 8] recommend a waiting period of about 1 year after initial surgery before reoperation. Hyun [21] believed that any judgment and decision regarding re-treatments should be done after 12 months after the original surgery, no matter how much the residual spasm existed after the first MVD. In contrast, Kassam [9] and Zhong et al. [22] reported that patients undergoing early reoperation were significantly more likely to be cured or improved than patients undergoing late reoperation. In the present study, we analyzed patients who underwent late repeat MVD. As a matter of fact, the relief of HFS did happen after the late redo MVD in 28 of the 33 patients postoperatively. Therefore, it was noted that late repeat MVD for HFS is an effective and safe procedure.

In previous reports, MVD had immediate spasm relief rates varying from 76.5 to 88.3 % [22–24]. The rest exhibited delayed gradual resolution of spasm after MVD during follow-up observations. After follow-up of 1–3 years after surgery, the successful spasm relief rate increased to 79–94.6 % [10, 22, 24], which is consistent with the pathophysiological mechanisms of primary HFS. Some think 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 [4, 25, 26]. However, because it cannot explain the immediate relief after a successful surgery, it is still unclear about the delayed relief. According to Moller [27], primary HFS may be due to hyperactivity of the facial nucleus, progressively induced by the chronic compression pulsation of the neurovascular conflict, especially at REZ. Thus, it would be physiologically logical that the effect of surgical decompression takes time to decrease and normalize the clinical spasm. However, as a matter of fact, most of the patients rather than a minority of the patient relieved immediately after a successful MVD. Zheng et al. [28] developed a hypothesis about the mechanism of HFS, which is the cross-transmission is bridged by sympathetic nerve fibers between the facial nerve fibers and offending vessels. Hence, this might explain most of patients who were relieved immediately after facial nerve and offending vessels wall were separated by Teflon. Some of the reasons for failure of first or repeat MVD may include not identifying the real offending vessels, Teflon pledget movement after closure, or incomplete decompression [9, 10]. Therefore, the primary reason to perform a repeat MVD would be to re-explore the entire root and explore missed compressive vessel [22]. The second reason would be to reexamine the relationship between REZ of CN VII and pledget to make ensure complete decompression.

Based on the spasm relief during the immediate postoperative period, the results indicate that intraoperative monitoring is an effective tool in helping identify offending vessels during repeat MVD [5, 29]. Our LSR disappearance rate in both groups was comparable; in addition, it was similar to previously published data [4, 30]. Sekula et al. [31] reported that the chance of resolved HFS if the LSR was abolished during surgery was 4.2 times greater than if the LSR persisted. Given the significant negative predictive value of LSR monitoring, the surgeon can be reassured that an adequate decompression has been achieved, thus avoiding unnecessary operative time and resultant complications especially when multiple vessels are involved [6, 32–34]. Absence of LSR to stimulation during MVD was similar in both groups. In our previous study, we found that patients with residual LSR had higher chance of persistent spasm [5]. We found a significantly higher number of patients in group I to have a vein as the offending vessel at the REZ in comparison to group II. These results suggest that patients with residual LSR should undergo exploration for a vein to prevent persistent HFS. Based on our previous published results [5] and the current study, the use of LSR in redo MVDs may be very valuable.

The common operative complications of MVD for HFS include new or worsening facial weakness, hearing loss, diplopia, headache, balance problem, CSF leakage, and dizziness/vertigo [29, 35, 36]. In our study, there was no significant difference in postoperative hearing decrease and/or loss, diplopia, headache, balance problem and CSF leak, and dizziness and vertigo between two groups and those previously reported. Our hearing outcomes after MVD were similar to our previously reported study [14]. It is possible that experience with reoperations and monitoring the facial and auditory nerve might have contributed to the insignificant difference in the outcomes between the two groups. However, facial weakness was noticed in 12.12 % of the patients after surgery in group I. The following points are important for higher percentage of facial weakness: (1) Repeat MVDs require more time for a careful exposure due to the presence of adhesions which increase the chance for facial weakness. (2) The number of patients who underwent redo MVD is limited, and it could represent the low sample size. (3) The higher incidence of facial palsy in group I could be related to higher botox usage. Although wound infection difference in our study did not reach statistical significance, we found the rate of 6.06 % is higher than that have been reported previously for HFS patients [37]. Given this, what the surgeon needs to do is the dura and muscle should be closed carefully. In our study, an insignificant number of patients had recovery after discharge in both groups I and II. In addition, majority of the patients had very good results. Engh et al. [9] showed that early reoperation was also associated with good results. Based on these data, we think it is important to undergo repeat MVD for failed HFS relief irrespective of the timing of the operation. One limitation of the study is that most of patients in whom a MVD was redone had their first surgery at other medical centers. So, no descriptions of the surgical findings at the first operation were available to better understand the putative cause of the failure from the study.

Conclusion

Repeat MVD for HFS is an effective and safe procedure. No specific preoperative clinical characteristics were identified in patients with repeat MVD. Intraoperative monitoring with LSR is an effective tool to evaluate adequate decompression. In patients with persistent LSR at the end of the procedure, facial nerve compression from a vein should be examined. We believe that it is important to undergo a repeat MVD for failed HFS relief irrespective of the timing of the operation.