Two‐step highly selective neurectomy for refractory periocular synkinesis

Botulinum toxin type A (BTA) has become a gold standard of treatment for facial synkinesis, and now is an integral part of the treatment paradigm for patients with this sequela. In our center, we routinely treat 30 patients per week with botulinum toxin. Addressing synkinesis via a multidisciplinary approach, we begin with physical therapy and later add chemodenervation to the regimen.1 Injections of BTA (onabotulinumtoxinA, Botox Cosmetic, Allergan Inc, Irvine, CA) are repeated every 3 to 6 months, if they offer symptom relief. Repeated BTA treatments can lead to decreased effectiveness through the well-described phenomenon of antibody development.2 Once BTA has lost its effectiveness, despite increased doses, we advance to type B botulinum toxin (BTB) (rimabotulinumtoxinB, Myobloc, Solstice Neurosciences LLC, Louisville, KY). We currently treat 390 patients with BTA injections; only 10 are receiving BTB. The same loss of effectiveness that occurs with BTA can occur with BTB; historically, these patients have continued their physical therapy but experience worsening of synkinesis symptoms in the absence of chemodenervation. To address this problem in the periocular area, we have refined the decades-old technique of selective neurectomy. Traditionally, the procedure did not identify specific nerve branches responsible for the synkinesis symptoms, resulting in either unsatisfactory relief, worsening of facial weakness, or a combination of both. Therefore, we resolved to avoid these problems through precise titration of the neurectomy. To this end, the procedure was divided into two steps: 1) Facial nerve dissection is done under general anesthesia; 2) The patient is awakened and recovered, and then the neurectomy is performed, achieving the exact degree of orbicularis oculi weakening required to decrease ocular synkinesis while avoiding lagophthalmos. Between August 2009 and August 2011, three patients underwent two-step highly selective neurectomy (2HSN) for facial synkinesis refractory to botulinum toxin chemodenervation. All three had been previously treated with physical therapy and BTA, followed by BTB. Facegram software, described and validated in 2010, was used to obtain palpebral fissure width measurements from photographs taken before and after surgery.3 During the first step of the operation, a preauricular incision is made on the synkinetic side and a flap elevated along the parotidomasseteric fascia. Dissection continues beyond the anterior border of the parotid gland until branches of the facial nerve are encountered. Using a Montgomery stimulator (Boston Medical Products, Worcester, MA), the branches innervating the orbicularis oculi are identified and isolated with vessel loops (Fig. 1). The dissection must be carried distally enough to locate four to six branches to the muscle. Once the dissection is complete, stab incisions, oriented parallel to the relaxed skin tension lines, are made overlying the branches; and the vessel loops are delivered through the incisions and then secured to the skin (Figs. 2 and 3). The preauricular incision is closed and the patient awakened from anesthesia. A facial flap is elevated, and facial nerve branches supplying the orbicularis oculi are isolated with vessel loops. Stab incisions are made overlying the facial nerve branches. The incisions are oriented within relaxed skin tension lines. The vessel loops are delivered through the stab incisions and secured to the skin. In the second step, the patient is awake and participates in the operation. As he or she repeatedly smiles on command, each nerve branch is delivered through its stab incision and divided in turn until the surgeon observes a suitable decrease in synkinetic closure of the eye without lagophthalmos (Fig. 4). A 1-cm segment of each nerve branch is removed in order to prevent axons from the proximal stump from penetrating into the distal stump during healing. Once the desired effect is achieved, the remaining branches are released into the wound, in continuity. No local anesthesia is required for this stage as there are no sensory fibers in the facial nerve branches. A schematic of procedure showing preauricular incision, stab incisions overlying nerve branches to orbicularis oculi, retraction and resection of a nerve branch. Three patients aged 35, 49, and 50 with periocular synkinesis underwent 2HSN, performed by the senior author (TAH). All three patients had photographs taken at initial consultation, as well as at 2 week and 2 month postoperative visits. The first step of the surgery was performed under general anesthesia in the main operating room, and in all cases this took approximately 2.5 hours. The second step was performed after the patient had been released from the postanesthesia care unit, approximately 1 hour later. Patients were transported to the procedure room in the outpatient clinic for this phase, which lasted only 10 minutes. There were no intraoperative complications and no postoperative complications. Since surgery, no patients have complained of lagophthalmos, xerophthalmia, or ectropion. Using Facegram software, improvements in palpebral fissure height were documented in all three cases (Fig. 5). The first patient had the smallest degree of change; the palpebral fissure width of the eye on the paralyzed side increased from 72% of the palpebral fissure width on the nonparalyzed side to 85% after surgery. The second and third patients improved from 28% to 96% and 43% to 85%, respectively (Fig. 6). All patients have had been evaluated between 6 and 24 months postoperatively, and all have experienced stability of palpebral fissure width correction. None of the patients is receiving further periocular botulinum toxin treatment. Preoperative (top) and postoperative (bottom) photographs of all three patients in this series. Note an increase in palpebral fissure width on the paralyzed side relative to the nonparalyzed side. All photographs were taken with the patient making a large smile. The preoperative and postoperative palpebral fissure widths on the paralyzed sides are expressed as percentages of the width of the palpebral fissures on the corresponding nonparalyzed sides, demonstrating improvements of 13%, 68%, and 42%, in patients 1, 2, and 3, respectively. Facial paralysis manifests itself in myriad ways, ranging from mild segmental weakness to full bilateral flaccid paralysis. The etiologies are numerous, including infectious, traumatic, neoplastic, autoimmune, congenital, and idiopathic. The most common type of facial paralysis is Bell palsy, which affects 20 to 30 people per 100,000, and accounts for approximately two-thirds of all new facial paralysis cases.4 Approximately 80% of Bell palsy patients recover spontaneously; and over 90% will improve to a House-Brackmann grade I or II if given steroids. The role of antiviral medication remains controversial.5 Despite the generally favorable prognosis of Bell palsy, many patients do not recover fully, particularly those with either complete flaccid paralysis or prolonged recovery periods. The most common sequela of facial paralysis without transection of the seventh nerve is synkinesis, occurring in 10% of patients.6 Aberrant regeneration of facial nerve axons occurs after an insult resulting in axonal injury with damage to the endoneurium, perineurium, or epineurium. In these cases, there is Wallerian degeneration of the axons beginning at the node of Ranvier immediately proximal to the site of injury, often in the geniculate ganglion, and proceeding distally.7 While the neuronal cell bodies usually remain viable, the axons must resprout, each one extending multiple growth cones and attempting to reestablish continuity with its target muscle. The large number of growth cones, travelling down a nerve trunk designed for only 7,000 fibers, often will result in misrouting and multiply terminal axons.8 Simple axonal misrouting may manifest as an axon intended for the zygomaticus major, incorrectly innervating the orbicularis oculi. This phenomenon is relatively treatable using neuromuscular retraining through physical therapy. The more complicated situation arises when multiple axonal extensions from a single neuron establish connections with multiple distal target muscles, resulting in involuntary movement simultaneous with voluntary movement (synkinesis). In this case, physical therapy must be supplemented with chemodenervation treatments, or neurectomy. Selective peripheral neurectomy for spastic facial palsy was first described in 1950 by Marino and Alurralde.9 In 1986, Dobie and Fisch reported a procedure involving complete extirpation of all branches of the facial nerve innervating the involved muscles.10 The majority of their patients had hemifacial spasm or blepharospasm, with only 13% suffering from synkinesis. Patients did well in general, with only a small number requiring tarsorrhaphy, and none complaining that postoperative facial weakness was worse than preoperative spasticity. This series differs significantly from ours in that the other patients had not undergone chemodenervation and physical therapy prior to surgical treatment, and their entire operation was conducted as a single stage under general anesthesia. At that time, phenol injection neurolysis was the preferred method of chemodenervation, although this was more painful than botulinum toxin, and its efficacy was limited by the scarring that accumulated after repeated injections every 3 months. Because the novel 2HSN is performed on the awake patient, the likelihood of a successful result is maximized. The surgeon is much less liable to under-resect, leaving persistent synkinesis, or over-resect, causing problematic periocular weakness. We have performed three of these procedures at our institution in the last 2 years, with success in all three cases as determined by computerized measurements of improvement in palpebral fissure height postoperatively. There have been no complications in our small series. Additionally, there is no visible scarring from the stab incisions made for delivery of the nerve branches (Fig. 7). Thus far there has been no recurrence of periocular synkinesis. Patients have described lasting improvements in not only the periocular area, but also in the lower zones of the face, suggesting that the periocular region may be a central focus of synkinesis so that relaxation, once achieved there, can propagate to the remaining facial muscles. Postoperative view of patient 1, demonstrating lack of scarring where stab incisions were made for delivery of nerve branches. The 2HSN procedure described in this article is a straightforward operation with reproducible results in a small cohort of patients. It can be applied to the small subset of chronic facial paralysis patients who have benefitted from physical therapy and chemodenervation in the past, but whose synkinesis has become refractory to botulinum toxin due to long-term repeated exposure. Our reported series herein is small; however, we have continued to treat patients with 2HSN and look forward to presenting a more comprehensive series in the long term.