Introduction

Hemifacial spasm (HFS) is a cranial nerve disease characterized by involuntary contractions of muscles innervated by the ipsilateral facial nerve [1, 25]. Jho and Jannetta raised that the neurovascular compression (NVC) at the root exit zone (REZ) of the facial nerve is the main cause of this disease in the last century and this idea is currently widely accepted [14]. Microvascular decompression (MVD) of the facial nerve is the most common surgical procedure carried out today, with success rates of more than 90 % in some series [2, 12]. About 2.9–50.3 % of patients are reported to obtain relief gradually rather than cured immediately after MVD. This process may take several months or even years, so it is debated what is the suitable time for patients with persistent hemifacial spasm after their first MVD to receive MVD again [7, 15, 22, 33]. In our study, we retrospectively analyzed the clinical features, surgical findings, outcomes, and complications of the repeat MVD for the patients who are not spasm-free more than 2 years after their first surgery.

Materials and methods

Research population

This study involved 26 patients who underwent repeat MVD from January 1, 2011 to December 31, 2015 in Nanjing Drum Tower Hospital. All of the 26 patients underwent a failed MVD previously, and 23 of them receive their first MVD at other institutions. In this study, failed MVD was defined as the patients who suffered from persistent hemifacial spasm after their first MVD and the spasm did not improve for at least 2 years during postoperative follow-up.

Surgery

All of the patients underwent repeat MVD by one chief physician and the surgery was performed under the electrophysiological monitoring of abnormal muscle response (AMR). Under general anesthesia, the patient was placed in lateral decubitus position. After durotomy, cerebrospinal fluid was drained sufficiently and the root exit zone (REZ) of the facial nerve was exposed under the microscope. The conflict site of first MVD was reexamined and every suspected offending vessel was separated from the facial nerve root with shredded Teflon implants. After confirming that there was no further compression, the dura mater could be closed, followed by cranioplasty and incision closure.

Surgical outcomes and complications

All the patients were followed up continuously and the following items were recorded: (i) Outcomes of the surgery: cured immediately, delayed relief [15], invalid and relapse. (ii) Complications: including facial paralysis, cerebrospinal fluid leakage, tinnitus, deafness, hearing loss, intracranial hematoma, cerebral infarction, and so on.

Results

Clinical features

This study included nine males and 17 females, aged from 37 to 71 years old when they received the repeat MVD. All of them underwent their first MVD more than 2 years ago. The time between the two operations lasted from 24 months to 67 years (average, 38.7 months). The spasm symptom was located on the left side in 14 and right side in 12 patients (Table 1).

Table 1 Results of the study group
Full size table

Findings in redo MVD

During the reoperations, we observed that the most common reason for the failure of a patient’s first MVD was misidentification of the exact offending vessel, which accounted for 15 (57.7 %) cases. The Teflon was placed between the REZ and the irrelevant vessels or just one branch of the offending vessels while the most important vessels were missed. The facial nerve root had been decompressed insufficiently in six (23.1 %) patients because the vertebral arteries were too close to the offending vessels or the vertebral arteries were one part of offending vessels. In four (15.4 %) patients, the Teflon implants were placed between the facial nerve and the offending vessels while neurovascular compression (NVC) at the root exit zone (REZ) still existed after the first surgery. Too many Teflon implants were observed in one (3.8 %) patient thus a new compression may have occurred. The AMRs disappeared in all the patients when their operations were finished (Table 1).

Surgical outcomes

The patients were followed up 8 to 65 months after the operation, with an average time of 43.4 months. Twenty-four (92.3 %) patients were cured immediately after the redo MVD. Delayed relief was found in two (7.7 %) patients. It took 6 days and 2 weeks for them to obtain complete relief. No recurrence was found during follow-up. Surgical complications included three (11.5 %) facial paralyses and one (3.8 %) hearing loss. No serious complications like intracranial hematoma or death occurred in these patients (Table 1).

Discussion

When to perform repeat MVD

Currently, neurovascular compression at the root exit zone of the facial nerve was believed to be the etiology of HFS, and MVD was recommend to be an effective treatment as postoperative cure rate was satisfying, ranging from 86 to 96 % in some reports [2, 4, 12, 34]. However, not all of the patients were cured immediately after the surgery, some patients may experience delayed relief for several months or even years [15, 20, 24]. Ishikawa et al. [13] studied 175 HFS patients and 88 of them had spasms after MVD. Residual spasms disappeared after 1 week in 25 %, after 1 month in 50 %, and after 8 months in 90 % of cases. Li [20] reported that 41 of 545 patients (7.5 %) had residual spasms after MVD but 37 achieved complete resolution within 1 year. Jo et al. [15] also reported that 1.4 % of the cured patients experienced delayed resolution at 1–3 years after treatment. Two patients got delayed relief for 3.5 years in Sindou’s reports, which was the longest time to date [33].

The etiology of delayed resolution has been debated. When the spontaneous or ectopic excitation by the pulsatile compressive force of the offending vessel disappeared after the surgery, HFS patients got an immediate resolution. In some cases, the complete regeneration of the microinjury of the facial nerve or the gradual stabilization of the facial motor nucleus cost some time, thus these may be the reason for delayed resolution [8, 13, 29, 32]. Some researches [32] suggested that the longer duration of HFS and the severer symptoms would lead to the longer time of delayed resolution.

Due to the existence of delayed resolution, early reoperation for the patients is not recommended by most of the scholars. Hyun [12] believed that even if the residual spasm existed after the first MVD, no judgment and decision regarding retreatments should be done before 12 months after the original surgery. Sindou et al. [33] and some other authors suggested that the treatment of persistent or recurrent patients with HFS should be more than 1 year after the previous MVD [4, 9]. Jo et al. [15] suggested the if the surgeon can confirm intraoperative resolution of the lateral spread response (LSR) and severe indentation, reoperation can be delayed until 3 years after the original MVD and he defined the patients who had residual or recurrent spasms at 3 years after first surgery as failed group.

Surgical outcomes and complications

In our studies, the 26 patients with persistent hemifacial spasm after their first MVD for at least 2 years and we defined their first MVD as a failure. During the reoperations, we found the failure may be attributed to misidentification of the exact offending vessels and improper operations in their first MVD. Without good exposure, it was not easy to inspect REZ of the facial nerve. When the vertebral artery combined with small vascular compressions, the accompanying small vessel, beneath the vertebral artery, might be the actual responsible vessel after shifting the vertebral artery [3]. Sometimes, CSF drainage may induce an anatomical shift of neurovascular relationship [26]. All of these causes could affect the surgeon in finding the exact offending vessels. Surgeons without much experience were more likely to miss the most important offending vessels or place the Teflon in a wrong way, resulting in insufficient decompression.

When the persistent spasm lasted more than 2 years after original MVD and there was no improvement, we suggest that compression has not been completely relieved and the first MVD failed. This is the time to make the decision for a reoperation. In our study, 24 (92.3 %) patients were cured immediately and the other two patients obtained spasm relief in 6 days and 2 weeks. The cure rate was excellent. Wang et al. [35] reported that repeat MVD had a cure rate of 85 % in the patients with a first-time MVD failure. Also, Engh et al. [5] suggested that late repeat MVD for HFS is a reasonable treatment option and it is effective in experienced hands.

Facial paralysis, hearing loss, and some other complications like balance problems and CSF leakage are common in MVD for HFS patients [17, 19]. Patients who had repeat MVD did not suffer a higher rate of complications than those who received the surgery only once [5]. In our study, we observed three (11.5 %) facial paralyses and one (3.8 %) hearing loss in the 26 redo MVD patients. Compared to some previous reports [5, 35], there is no significant difference. However, when repeat MVD was performed, local arachnoid’s adhesions and unclear brain tissue due to the first MVD will make it difficult for the surgeon to re-explore the REZ of facial nerve. Payner et al. [27] observed that the risk of hearing loss was greatest in patients who underwent repeat MVD. Thus, in addition to AMR, intraoperative monitoring of brainstem auditory evoked potentials (BAEPs) is necessary in order to decrease the danger of hearing loss [23, 28]. Surgeons should operate more gently and carefully during the redo MVD in order to not damage cranial nerves or blood vessels.

The usage of AMR monitoring

Intraoperative monitoring of AMR is highly useful as an index for the effectiveness of decompression [18, 30]. Sekula et al. [30] reported that the chance of resolving HFS if the AMR is abolished during surgery was 4.2 times greater than if the AMR persisted. AMR can help indicate the identification of offending vessels and confirm whether an adequate decompression has been achieved [6]. Thus, surgeons can avoid unnecessary operative time and reduce postoperative complications. However, it should be noted that the disappearance of AMR during surgery does not mean being definite spasm-free after surgery, and some patients may obtain spasm relief though AMR remains after decompression [10, 21, 34]. Kong et al. [18] reported that AMR was observed during intraoperative facial EMG monitoring in 263 of 300 patients, and 33 patients showed persisting AMR despite decompression. However, 66.7 % of the 33 patients were spasm-free after surgery. Joo et al. [16] also reported that 81.3 % of the patients who had persisting AMR after decompression were cured during the follow-up period. Besides, some patients had remaining spasms even though AMR disappeared after decompression [10, 11]. Thus, there is some limitation of predicting clinical outcome by monitoring AMR in HFS patients. In our study, the AMR disappeared in all of the patients when the operations finished, and the patients have a satisfying surgical outcome. So, we suggest that intraoperative AMR monitoring is helpful in a redo MVD but the most important factor is the surgeons’ experience.

Undoubtedly, HFS will affect the patients’ quality of life. Shibahashi et al. [31] suggested that the patients who were cured immediately and had no postoperative complications had higher quality-of-life scores. Patients suffering from persistent hemifacial spasm after their first MVD may become anxious and the spasm symptoms may get even worse due to their negative emotions. These patients should be comforted and reassured that most of them will obtain relief during the follow-up. Even if the spasm has persisted for more than 2 years, performing repeat MVD is a good choice.

Conclusions

In conclusion, the postoperative course of MVD for HFS patients can vary. Some patients obtain relief gradually after the original MVD. We insisted that at least 2 years of follow-up is necessary before performing repeat MVD in patients with a failure. Repeat MVD is effective when done by experienced surgeons and should be performed with the aid of intraoperative monitoring of AMR.