Thirty-five year experience of an aggressive surgical intervention for treating post-paralytic facial synkinesis: an outcome study

Aim: Combined neurectomy & myectomy and functioning free muscle transplantation is proposed as an aggressive surgical intervention for postparalytic facial synkinesis (PPFS) to effectively resolve the problem since 1985 and this treatment continues to be the standard. We aim to describe our experiences with 103 PPFS patients who underwent the surgical treatment. Methods: A total of 103 patients with PPFS were investigated (1985-2020), but 94 were selected with all having at least one year of postoperative follow-up. Among them 50 were Type II and 44 were Type III PPFS. All patients underwent extensive removal of the synkinetic muscles and triggered facial nerve branches in the cheek, nose and neck regions, followed by gracilis transplantation for facial reanimation. Results: The incidence of receiving the aggressive surgical intervention increased from 15% prior to 2012 up to 24%. The mean postoperative follow-up period was 10 years. Young adult (79%) and female patients (63%) were the dominant populations, showing their great ambition for a treatment. Results showed a significant improvement in facial smile with more teeth visible, and a significant decrease in facial synkinesis. About 96% (90 patients) did not require botulinum toxin A injection after surgery. Revision surgery for secondary deformity was approximately 53%. Conclusion: Treatment of PPFS is primarily reconstructive. Combined myectomy & neurectomy and functioning free muscle transplantation for Type II and III patients are well accepted, and leads to promising and long-lasting results despite higher revision rates. Refined techniques to decrease revision rates are needed. Page 2 of 14 Chuang. Plast Aesthet Res 2021;8:5 I http://dx.doi.org/10.20517/2347-9264.2020.190

With advanced micro-neuro-vascular surgery, nerve transfer and functioning free muscle transplantation (FFMT) is now becoming the mainstream surgery for long-standing facial paralysis reconstruction [6][7][8][9][10][11] . However, the treatment of PPFS remains challenging and controversial, although the mechanism of PPFS has been well studied and proposed, and the hypothesis of aberrant reinnervation is widely accepted [5] .
The author (DCC Chuang) started his clinical practice for facial paralysis reconstruction at Chang Gung Memorial Hospital in 1985. In the first 10 years (1985)(1986)(1987)(1988)(1989)(1990)(1991)(1992)(1993)(1994) [ Figure 1] conservative treatments by botulinum toxin A (BTX-A), neurectomy, myectomy, and/or rehabilitation program were performed, but were ultimately deemed impractical and insufficient. In the subsequent 25 years (1995-2020) [ Figure 1] after accumulation of experience with PFP reconstruction, the author shifted the treatment to a more aggressive surgery to resolve the problem, which now becomes our standard practice. The aim of this article is to give a comprehensive and prospective review of PPFS patients who were treated by combined myectomy, neurectomy and gracilis FFMT, an aggressive surgical intervention.

METHODS
A special "Patient Examination Sheet" was designed in 1995 to record all facial paralysis patients [Supplement Digital Content 1]. The physical examination of the face is performed at repose and movement. There are five basic facial movements for recording: forehead raise, eye closure, smile with teeth exposure, snarl with lower lip pulled down, and whistle with lip pout. For PPFS, six common facial synkinesis movements induced by the individual basic movement are recorded, including frontal synkinesis with corrugator hypertrophy and tension face, eye synkinesis with eye narrowing, upper lip synkinesis with mouth commissure/upper lip twitching or elevation, lower lip synkinesis with chin skin dimples, and neck synkinesis with platysma neck bands and lower lip eversion.
Following examination with the "Patient Examination Sheet", the PPFS patients were divided into four "Types" (or Patterns) based on the quality of smile and degree of synkinesis. Quality of smile was basically dependent upon the number of teeth exposure during smiling as well as smile symmetry. Severity of synkinesis was recorded from none (0), mild (+), moderate (++), or severe (+++), based on different synkinetic movements of the face following orders of individual five basic facial expressions [ Table 1].

Type I: good smile but mild synkinesis
Type I PPFS was seen in 4% of the population [26] . The patients always show at least 4 teeth with a big smile and good appearance. Mild tension on the affected side of the face at repose and intermittent perioral twitching following eye blinking are always noted. Patients usually close the eye without lagophthalmos, but the movement is slow and weak. Good smile quality is associated with less attention to synkinesis. Usually, most patients are aware of their altered facial expression, but observers including physicians might not fully recognize the problems due to their mildness. In addition, some patients are themselves not aware that they have synkinesis. Type I patients are mostly concerned more with cosmetic issues than synkinesis.
Type II: acceptable smile but moderate to severe synkinesis Type II PPFS was seen in 48% of the population [26] . In Type II, the patient's biggest smile shows 3-2 teeth with an acceptable smile. Their bizarre facial expression sometimes requires wearing a facemask to cover up  for the differences seen. Lagophthalmos with eye incompetence is often seen. The patients have moderate to severe synkinesis. Dominant narrow eye, deep nasolabial fold and dominant neck bands are commonly seen on the affected side. Some patients might have a bizarre facial expression, such as lip pouting or spitting inducing eye closure on the affected side. Facial disfigurement with synkinesis is the major reason of these patients presented for treatment [ Figure 2A].

Type III: unacceptable smile and moderate to severe synkinesis
Type III PPFS was seen in 47% of the population [26] . In Type III, patients have a smile with one or no teeth visible, which leads to an unacceptable smile. Lagophthalmos is often noted. They are easily recognized given the poor smile and severe synkinesis. Like Type II, facial disfigurement with synkinesis is the major reason to seek for treatment. Patients in Type III demonstrate a greater ambition to accept aggressive procedures to correct their disfigurements [ Figure 2B].

Type IV: poor smile and mild synkinesis
Type IV PPFS was seen in less than 1% of the population [26] . In Type IV PPFS, patients have near complete facial paralysis with mild synkinesis. Lagophthalmos is always noted. The patient's chief complaints for requesting a treatment are related to the paralytic face, less on the synkinesis.
A retrospective review was performed between January 1985 and August 2020 (35 year period, Figure 1). A total of 430 facial palsy patients were treated by gracilis-FFMT for smile reconstruction. Of these, 103 (24 %) were PPFS patients. The rate of treating PPFS patients with a surgical approach increased from 15% prior to 2012 [26] up to 24%.
After exclusion of 9 patients in 2020, a total of 94 PPFS patients were selected and enrolled in this study. They all had at least one year of postoperative follow-up. Patient characteristics including gender, age at the time of surgery, etiology, affected side, type of PPFS, and FFMT neurotized by different motor neurotizers (CFNG, cross-facial nerve graft; XI, spinal accessory nerve; V3, masseter nerve) were recorded [ Table 2].
In the selected 94 PPFS patients, 50 patients were Type II, and 44 patients were Type III. They all underwent an aggressive surgical intervention, which included combined myectomy & neurectomy, and smile reconstruction with gracilis FFMT. This was performed during the first stage of XI-gracilis and V3-gracilis, or the 2nd stage of CFNG-gracilis. Four incisions were made on the affected side of the face [ Figure 3A]. The first incision was on the white line of the upper lip. The incision went directly and vertically down to the above-mucosa layer, leaving the orbicularis oris muscle on the lip side. If the CFNG was performed in the first stage, the short CFNG was identified, mobilized, and protected in the lip wound.
Three to four 4-0 Dexon anchoring sutures were placed behind the orbicularis oris muscle for a later gracilis aponeurosis attachment. The second incision was 1cm in length on the white line of the lower lip for lower lip suspension with gracilis tail. The third incision was a long preauricular incision from the temporal hair and down to the mandibular body for a subcutaneous face-lift. Under the elevated skin flap, a trapezoid with four sides was drawn from the zygomatic arch down to the upper lip [ Figure 3B]. Dissection was performed along the surface of the masseter fascia and muscle down to the inferior edge of the drawn trapezoid area. All tissues in this trapezoid-marked area, including the upper lip elevators (zygomatic muscles, levator labii superioris, buccinator, caninus), some of the nose muscles (nasalis), and zygomatic and buccal facial nerve branches were removed en bloc [ Figure 3C]. Parotid duct (Stensen's duct) was protected and preserved to avoid injury and saliva leak. An assistant typically inserted a finger into the mouth to guide soft tissue removal and avoid penetration into the oral mucosa. The buccal fat pad was routinely removed subtotally to decrease the volume and to avoid its herniation in the future. A subcutaneous space was created [ Figure 3D]. A 3-cm transverse incision down to the zygomatic arch bone was made. The incision could be extended medially to the zygomatic body, and laterally but not beyond the line of the pathway of the temporal facial nerve branch. The temporal facial nerve branch was protected to avoid lagophthalmos. Three to four 3-0 Dexon stitches were anchored on the suprazygomatic margin periosteum for a later gracilis muscle attachment. Corrugator muscle, if preoperatively noted to have hypertrophy, was removed through another small suprabrow incision. On the mandible wound, the wide insertion of the platysma on the mandible was totally exposed and dissected down to its insertion on the clavicle. Because of its breadth, a 4th supraclavicular access incision was always required. Total removal of the platysma was required because it was a major synkinetic muscle [ Figure 3A and C]. Under the platysma, the mandibular facial nerve branch crossing the facial artery and vein was found and protected to preserve lower lip movement. The facial vessels were dissected and prepared for later vessel anastomoses to the transferred muscle's vessels. A suction drain was always placed in the neck wound because oozing was much more than the surgery for PFP. Smile reconstruction with gracilis FFMT Dissection and harvest of the gracilis from the contralateral thigh was performed simultaneously with the surgical ablation described above. The indication and surgical details for three methods have been reported in prior publications [10,11] .

Postoperative care and rehabilitation protocol after FFMT
After ablative and reconstruction surgery, the midface became nearly paralytic. Postoperative care and rehabilitation were principally the same as CFNG-, XI-and V3-gracilis transplantation for smile reconstruction, including massage, electric stimulation, and "Induction Exercise" [10,11] .
Two extra rehabilitation procedures, which were proposed and advocated in recent years, were followed: (1) "upper lip stretching exercise" [Video 1] on the reconstructed side was usually performed 6 months postoperatively, when the gracilis muscle started to move (M1). The "stretching exercise" was performed daily and continued forever to avoid potential lip contracture; and (2) smiling training in front of the mirror was performed once the gracilis muscle function reached to M2 (two teeth visible). This usually occurred one year after FFMT. Smiling on the reconstructed side is strictly triggered by the healthy side and avoided by the reconstructive side to avoid recurrent synkinesis.

Outcome assessment
Postoperatively, all patients were evaluated clinically to confirm the function of FFMT without EMG. All patients were closely followed every 2 months for at least one year. Patients were photographed and videotaped during pre-and postoperative visits in a standardized fashion, including face at repose, face at smile (mild, moderate and maximal smile), and degree of residual synkinesis of the eye, mouth and neck. Outcomes were assessed in terms of success of decreasing synkinesis and achieving facial symmetry at rest and smiling, which were recorded in the special chart.

Statistical analysis
Statistical analysis was performed using SPSS 22.0 software (SPSS, Inc., Chicago, Ill). Data were presented with mean and ± standard deviation. Mann-Whitney U test was used for comparison between two independent samples and Kruskal-Wallis test for more than two independent groups. Nonparametric test was applied for unequal patient numbers in different groups with skewed distributions. P values less than 0.05 were considered statistically significant.

RESULTS
Young adults (79%) and female patients (63%) were dominant in the study population, showing their great ambition for receiving treatment.
The 90 (96%) patients felt that their synkinesis was significantly decreased after surgery and had no need for a BTX-A injection. The improvement of synkinesis started on the first day postoperatively and continued to improve throughout the long-term follow up and smiling training. The most improved sites were the eye and neck areas with absence of eye narrowing and neck bands. Many patients also showed disappearance of synkinesis in the clinic follow-up. The synkinesis became subclinical.

Complications
Acute complications such as flap reopen, flap loss, hematoma, saliva leak or wound dehiscence were rare. The most significant complication was a high revision rate (approximately 53%) in the long-term followup after FFMT reconstruction. Major secondary deformities [ Table 3] included: (1) deep fold on the upper lip and/or mouth angle [ Figure 10A]; (2) wide lip deformity [ Figure   Four patients (4%) required BTX-A injection due to a recurrent lip contracture from the underling gracilis after revision surgery. One patient who failed the CFNG-innervated gracilis received a masseter-innervated 2nd gracilis for reconstruction one year after the 1st gracilis removal and showed improved results.

Philosophy of PPFS treatment
Aberrant reinnervation may occur in motor and sensory distributions following peripheral nerve degeneration and regeneration. Motor recovery with aberrant reinnervation accompanies synkinesis (or co-contraction), which especially occurs in PPFS and obstetrical brachial plexus palsy [27,28] . Only aggressive surgical intervention can effectively resolve the problem, including muscle transfers and FFMT [26][27][28] .
Once a synkinesis, always a synkinesis. Treatment of PPFS is a warto patients, surgeons and rehabilitation therapists, just like a scar is a war. It requires team work and more patience. The treatment should include a good preoperative explanation, an aggressive surgical intervention, postoperative long-term rehabilitation and close clinical follow-up, and the abilities to correct any potential deformities. Our results were longlasting and 96% of treated patients did not require an additional BTN-A injection.

Surgical procedure for PPFS and PFP
Surgical procedure for PPFS is quite different from PFP: (1) Soft tissue removal: For incomplete PFP, only soft tissues above the facial nerve branches between the zygoma bone and upper lip are removed to create a shallow trough for a muscle inset. For complete PFP, the soft tissues above the masseter muscle are removed to create a bigger space. For PPFS, the soft tissue removal is similar to complete PFP, but is more extensive and includes the midface facial mimetic muscles, neck platysma, corrugator and some nasal muscles. But temporal and mandibular facial nerve branches should be preserved to prevent lagophthalmos and lower lip asymmetry; and (2) the gracilis FFMT is trimmed longer than when used in PFP to avoid subsequent contracture.

Higher revision rates
There are several potential reasons: (1) the harvested gracilis might be too short; (2) reversal of the gracilis puts the aponeurosis into the movable upper lip; and (3) the residual synkinesis. Although the midface mimetic muscles and the zygomatic and buccal facial nerve branches are extensively removed, there are still residual muscles in the perioral region, around 30%. Those residual muscles are still functionally innervated by the superior temporal and inferior mandibular facial nerve branches. A refined gracilis harvest and inset, and intensive postoperative rehabilitation are ongoing in recent years to decrease the revision rates.  PPFS surgery is reconstructive, not cosmetic. The treatment requires team work, including comprehensive preoperative explanation, postoperative long-term rehabilitation and close clinical follow-up, and surgical skills to treat the second deformity. The results for the Type II and III PPFS treatment is effective and longlasting. They were satisfied with the results despite a high revision rate.
In conclusion, treatment of PPFS is primarily reconstructive, not aesthetic. Combined myectomy & neurectomy and FFMT are aggressive. The patients, Type II and III PPFS, accept it because of the promising and long-lasting results, despite a high revision rate. Refined techniques to decrease the revision rates are needed.