J.Sales-Llopis

Neurosurgery Department, University General Hospital of Alicante, Foundation for the Promotion of Health and Biomedical Research in the Valencian Region (FISABIO), Alicante, Spain

It is a condition characterized by spontaneous deterioration of facial nerve function in a patient who has otherwise normal or near-normal facial function in the immediate postoperative period. This delayed paralysis is generally reported to occur in the first few days postsurgery, with the majority of patients eventually recovering their immediate postoperative facial function. However, infrequently, it can also occur more than 1 week after surgery (so-called late-onset facial nerve palsy) ¹⁾.

Even those patients with normal to near-normal function in the early postoperative period remain at risk for delayed facial palsy (DFP) after vestibular schwannoma surgery.

Patients with a gross total tumor resection or undergoing a retrosigmoid approach may be at higher risk of DFP. The prognosis is favorable, with patients likely recovering to normal or near-normal facial function within 1 month of onset ²⁾.

It appears that DFP is an uncommon consequence of surgery for VS. Although excellent recovery of facial nerve function to its original postoperative status nearly always occurs after DFP, the magnitude and time course of the disorder were not predictors for subsequent recovery of facial nerve function ³⁾.

Delayed facial nerve dysfunction after vestibular schwannoma surgery is a poorly understood phenomenon that has been reported to occur in 15 to 29% of patients undergoing microsurgery ⁴⁾.

Although patients with DFP did not exhibit any distinguishable characteristics when compared with patients without postoperative facial palsy, a analysis identified significant differences in patients with palsy presenting immediately postoperatively. Further study of patients with DFP should be undertaken to predict its incidence following VS resection ⁵⁾.

Herpes simplex virus (HSV) or varicella zoster virus (VZV) reactivation might be responsible for most cases of DFPs, thus suggesting the usefulness of immediate steroid and aciclovir administration to obtain total recovery. The viral reactivation mechanism is comparable to that already suspected in DFP occurring with the same delay in middle ear surgical procedures ⁶⁾.

In the patients who had a less favourable recovery, the palsy was more severe, and the onset occurred after some time. This seems to agree with those who are of the opinion that the complication is due to viral reactivation. In these patients, it is advisable to start aggressive medical therapy with antiviral agents such as acyclovir as soon as possible ⁷⁾.

A retrospective cohort study of 489 patients who underwent vestibular schwannoma resection at our institution between 2000 and 2014. Delayed facial palsy was defined as deterioration in facial function of at least 2 House-Brackmann (HB) grades between postoperative days 5 to 30. Only patients with a HB grade of I to III by postoperative day 5 were eligible for study inclusion.

One hundred twenty-one patients with HB grade IV to VI facial weakness at postoperative day 5 were excluded from analysis. Of the remaining 368, 60 (16%) patients developed DFP (mean 12 days postoperatively, range: 5-25 days). All patients recovered function to HB grade I to II by a mean of 33 days (range: 7-86 days). Patients that developed DFP had higher rates of gross total resections (83% vs 71%, P = .05) and retrosigmoid approaches (72% vs 52%, P < .01). There was no difference in recovery time between patients who received treatment with steroids, steroids with antivirals, or no treatment at all (P = .530).

Patients with a gross total tumor resection or undergoing a retrosigmoid approach may be at higher risk of DFP. The prognosis is favorable, with patients likely recovering to normal or near-normal facial function within 1 month of onset ⁸⁾.

Of the 104 patients who underwent VS resection, 25.0% developed DFP and 8.6% demonstrated IFP postoperatively. The DFP group did not differ significantly in any measure when compared with patients with no postoperative facial palsy. However, patients with DFP presented with significantly smaller tumor sizes than patients with IFP. This IFP group averaged significantly smaller intraoperative facial nerve responses than patients without facial palsy, and larger tumor sizes than both the DFP and no facial palsy groups. Within the DFP group, patients with late onset DFP showed diminished intraoperative facial nerve responses when compared with the total DFP patient population. In total, 25 (96.2%) of 26 patients with DFP and 7 (77.8%) of 9 patients with IFP recovered to normal or near-normal facial function (House-Brackmann Grade I or II) at longest clinical follow-up ⁹⁾.

Retrospective study on 8 cases from a series of 348 patients operated on of a VS between 1996 and 2002. Seven of the eight patients were given intravenously acyclovir (30 mg x kg(-1) x d(-1) for 5 days) and methyl-prednisolone (2 mg x kg(-1) x d(-1) for 7 days). A serologic testing looking for specific anti-herpes simplex viruses type 1 and 2 (HSV-2) and varicella-zoster virus (VZV) antibodies at the onset of the FP and 2 weeks later could be done in only 3 cases.

Mean delay of FP onset was 8.75 days. Mean time for recovery with intravenous treatment was 90 days. All treated patients had a House and Brackmann grade 1 recovery. The last one had only a grade 3 after 400 days of evolution: he could not be treated because of postoperative transient psychiatric problems. Serologic testing revealed in those patients in whom it could be done either a high level of anti-HSV or -VZV antibodies at the time of onset or a dramatic increase in anti-HSV or anti-VZV antibodies between the two samples, strongly suggesting a HSV or VZV reactivation.

HSV or VZV reactivation can be evocated in most cases of delayed FPs arising in the postoperative course of VSs, suggesting usefulness of emergency-given steroid and acyclovir intravenous regimen to block virus replication and fight secondary oedema and inflammation causative of nerve lesions. Evoked reactivation mechanism is comparable to that already suspected in delayed FP arising with the same delay in middle ear surgical procedures ¹⁰⁾.

Between 1990 and 2001 seven patients were identified who received vasoactive treatment for preservation of hearing and developed a delayed facial nerve paresis after termination of medication. Intraoperative facial nerve EMG activity was analyzed in six patients.

All patients developed a delayed facial nerve paresis between 2-5 days following termination of a 10 day treatment consisting of HES and nimodipine. Medication was re-initiated and the facial nerve paresis improved in all patients. In two patients intraoperative EMG signals revealed “A trains” waveform patterns, which are highly suggestive for an immediate postoperative facial nerve paresis, whereas in four patients no pathognomonic EMG patterns could be recorded.

The delayed onset of a facial paresis following termination of vasoactive treatment points to a disturbed microcirculation of the nerve as the main pathophysiological feature. Two groups could be identified on the basis of intraoperative EMG activity. In one group with presence of “A-trains” medication apparently masked the onset of an immediate postoperative facial nerve deficit. Four patients without “A-trains” did not develop a typical delayed facial nerve paresis during vasoactive treatment, but thereafter. The time lag between termination of treatment and onset of a delayed palsy points to a protective effect due to improved microcirculation ¹¹⁾.

The deterioration in the facial function was found to be delayed in 25 of the 98 patients (26%); of these, it occurred in the first 5 days after surgery in 11 cases, between 6 and 13 days in 10 cases, and after 15 days in 14 patients. The incidence rate of the delayed facial dysfunction was not influenced by age, sex, or the size of the tumour. The prognosis of the facial dysfunction was favourable in the majority of cases, and, in fact, there were only five grade III to IV cases 1 year later. Facial dysfunction was over grade III in the majority of the latter five cases, and the period of recovery was long.

Eighty percent of our patients with delayed facial palsy following vestibular schwannoma resection were classified as having excellent or good function. In the remaining patients who had a less favourable recovery, the palsy was more severe, and the onset occurred after some time. This seems to agree with those who are of the opinion that the complication is due to viral reactivation. In these patients, it is advisable to start aggressive medical therapy with antiviral agents such as acyclovir as soon as possible ¹²⁾.

Seven patients developed DFP after acoustic neuroma surgery, while the remaining 13 patients did not. There was no difference in preoperative and 3-week postoperative IgG titers for any of the 3 viruses tested. However, IgM titers were much higher postoperatively in DFP patients for all 3 viruses tested. The average HSV-1 IgM titer rose 92% in DFP patients compared with only 4.5% in the patients who did not develop DFP. Average HSV-2 IgM titers rose 70% compared with a decline of 8.5% in non-DFP patients. Most strikingly, VZV IgM titers rose an average 495% postoperatively among DFP patients compared with a decline of 14% in the non-DFP patients.

Elevation of the IgM titers of the viruses measured in this study implies that recrudescence of the virus has occurred. The absence of this rise among patients who did not develop DFP implies that viral recrudescence plays a role in the etiology of DFP. These findings support treatment or prophylaxis of DFP with antiviral therapy ¹³⁾.

Grant et al. identified 15 (4.8%) patients from a consecutive series of 314 who underwent surgery for VS between 1988 and 2000, and in whom DFP developed. Delayed facial palsy was defined as a deterioration of facial nerve function from House-Brackmann Grades 1 or 2 more than 3 days postoperatively.

All patients underwent intraoperative neurophysiological monitoring of facial nerve function. The average latency of DFP was 10.9 days (range 4-30 days). In six patients (40%) minor deterioration (< or = two House-Brackmann grades) had occurred at a mean of 10.2 days postsurgery, whereas in nine patients (60%) moderate deterioration (> or = three House-Brackmann grades) had occurred at a mean of 11.8 days postoperatively. Five (33%) of 15 patients recovered to Grade 1 of 2 function within 6 weeks of DFP onset. Of the 15 patients with DFP, 14 had completed 1 year of follow up at the time of this study. Twelve (80%) of these 15 patients recovered to Grade 1 or 2 function within 3 months, and 13 (93%) of 14 patients recovered within 1 year. In all cases, stimulation of the seventh cranial nerve on completion of tumor resection revealed the nerve to be intact, both anatomically and functionally, to proximal and distal stimulation at 0.1 mA. A smaller tumor diameter correlated with greater recovery of facial nerve function. There was no correlation between the latency or severity of or recovery from DFP, and the patient's age or sex, the surgical approach, frequency of neurotonic seventh nerve discharges, anatomical relationship of the facial nerve to the tumor, patient's history of tobacco use, or cardiovascular disease ¹⁴⁾.

Eight of the 67 patients developed a worsening of facial function after the first postoperative day. There were three males and five females with an age range of 29-73 years (mean, 53 years).

DFP should be defined as any deterioration of facial function after vestibular schwannoma surgery ¹⁵⁾.

Late-onset facial dysfunction was seen in 13 patients (2.1%). All of these had significant deterioration in facial nerve function between 1 and 4 weeks postoperatively, and all showed improvement by 1 year ¹⁶⁾.

Delayed facial paralysis occurred in 62 (24.3%) patients; 90% ultimately recovered to their initial postoperative HB grade, and 98.3% recovered to within one grade of their initial HB level. Paralysis occurred at an average of 3.65 postoperative days (range, 1-16 days). The average time to maximal recovery for those with changes of 1, 2, 3, and 4 HB grades was 5.6, 21.5, 39.8, and 50.5 weeks, respectively. The early onset of paralysis (< 48 h after surgery) resulted in shorter average recovery times. Of patients who demonstrated nerve deterioration to grades IV-VI, 20 of 38 required tarsorrhaphy or gold-weight placement. We conclude that the over-whelming majority of patients with delayed facial paralysis after acoustic neuroma surgery do eventually recover to their postoperative HB grade. The magnitude and timecourse of delayed facial paralysis are predictive factors for subsequent recovery ¹⁷⁾.

A patient developed delayed facial nerve palsy at the level of House-Brackmann grade I to grade III 10 days after vestibular schwannoma surgery by the suboccipital transmeatal approach. The palsy had completely recovered after one month. Immunological study showed reactivation of herpes simplex and magnetic resonance (MR) imaging demonstrated an abnormal enhancement pattern of the facial nerve; intense enhancement of the distal intracanalicular segment and labyrinthine segment, similar to the MR findings for Bell's palsy. A prospective control study on the enhancement pattern of the functionally preserved facial nerve after vestibular schwannoma surgery in six cases showed a similar pattern to that of the normal facial nerve. Based on these findings, we propose the hypothesis that herpes simplex reactivation is an underlying cause of delayed facial palsy after vestibular schwannoma surgery ¹⁸⁾.

A patient who underwent translabyrinthine resection of an intracanalicular acoustic neuroma and in whom developed otalgia, vesicles on the ear canal and the ipsilateral buccal mucosa, and progressive facial palsy the week after surgery. Serologic evaluation confirmed the diagnosis of herpes zoster oticus. Reactivation of latent virus apparently occurred as a result of surgical manipulation of the facial nerve. This parallels viral reactivation seen in trigeminal nerve surgery. We propose a new theory for an additional cause of delayed facial palsy after acoustic neuroma resection-reactivation of latent herpesvirus resulting from surgical trauma. Acyclovir should be evaluated in clinical trials for a prophylactic role in patients undergoing acoustic neuroma resection or a therapeutic role in patients in whom a delayed postoperative facial palsy develops ¹⁹⁾.

A patient who underwent surgery for a left sided acoustic neuroma. Facial nerve function was preserved without clinical or electrophysiological deficit during the immediate postoperative phase and then lost almost totally over two days. This began on the seventh postoperative day. Recovery became visible under hypervolemic hemodilution and glucocorticoids a few days later and was complete after three months. Vasospasm in combination with neural edema might account for this uncommon delay of postoperative facial nerve deterioration ²⁰⁾.