Introduction. Benign tumour surgery of the parotid gland may cause iatrogenic injury of the facial nerve, with results of postoperative treatment depending on the type of injury. The study aimed to clarify the mechanism of facial nerve injury after benign tumour surgery of parotid gland.

Material and Methods. The effectiveness was verified preoperatively and 1, 3, 6 and 17 months postoperatively. House-Brackmann scales, electroneurography, blink reflex study and needle electromyography were performed. Pharmacological treatment (Galantamine, Cocarboxylase, Dexamethasone, Triamcinolone) and supervised physiotherapeutic procedures (Facial-Oral-Tract-Therapy, Proprioceptive neuromuscular facilitation) were applied for six months.

Results. Tumour removal led to the total paralysis of the left facial nerve, IV, III and III House-Brackmann grades were ascertained at the subsequent 3rd–5th periods of observation. In postoperative studies, electroneurography results showed full functional recovery of the frontal branch and incomplete regeneration in the marginal mandibular branch. Blink reflex examination showed proper parameters of evoked potentials only during preoperative and the last observation period. Residual voluntary activity of the frontal muscle and weak voluntary activity of orbicularis oris muscle were recorded in the needle electromyography examination. Contracture of mimic muscles at rest and improvement of their voluntary activity on the left side was observed six months after surgery compared to the early period of observation.

Conclusion. Consecutive studies showed the predominant axonal type of injury in the marginal mandibular branch and neuropraxia effect of the facial nerve, allowing the creation of a rehabilitation programme optimal for the functional recovery of the nerve.

Iatrogenic injury of the facial nerve accompanied by benign tumours surgery occurs rarely [1], with the size and location of the tumour within the parotid gland determining the scope and duration of surgery which may contribute to postoperative complications [2]. Nerve oedema and ischaemia during surgery, as well as mechanical and thermal manipulations within surrounding tissues may result in persistent or temporary facial nerve injuries in the postoperative period [3,4], which can lead to axonotmesis or neuropraxia, conditioning the degree of facial nerve damage. Neuropraxia is present when a conduction block appears without macroscopic nerve damage, which may reflect a demyelinating lesion without a Wallerian degeneration distal to the injury. In the case of axonotmesis, the nerve fibers, including motor axons are damaged, but the epineurium and perineurium remain intact, with the Wallerian degeneration occurring distal to the zone of injury [5]. Data regarding the type of iatrogenic injury, neuropraxia or axonotmesis dominating during parotid gland tumour surgery is lacking. Elmadawy, Hegab, Alahmady and Shuman [6] pointed to electromyographic (EMG) and electroneurographic (ENG) assessment as valuable tools for the evaluation of facial nerve function after different surgeries. Axonotmesis and conduction block in motor fibers may lead to degenerative changes in muscles which can be detected in needle electromyographic recordings. Conduction block as result of neuropraxia can be functionally reversible. Similarly, in the case of axonotmesis, the regeneration process is possible due to the surviving proximal stump of the nerve. The improvement of motor function in facial nerve and face muscles after 12 weeks of conservative treatment may indicate a functional recovery [6,7]. Additionally, late complications related to improper functional regeneration of the facial nerve are crucial for successful treatment [8,9].

This study aimed to explain the mechanism of facial nerve iatrogenic injury by means of axonotmesis versus neuropraxia taking into consideration results of clinical and neurophysiological observations performed 1, 3, 6 and 17 months after benign parotid gland tumour surgery. Also, conservative treatment results with pharmacotherapy which may influence nerve functional recovery in one patient with the iatrogenic facial palsy are presented.

The pilot study included thirteen patients with a benign parotid gland tumour, which was recognised based on clinical and ultrasonographic examinations. The study group consisted of thirteen patients, including six women and seven men aged 16 to 67 (the average age was 47), with tumours of small or medium size, from 1.5 cm to 3.5 cm, located in the superficial lobe of the parotid gland (N = 12) except one patient, where the tumour was located in both lobes. Results of histopathological examination, surgical treatment as well as the clinical and neurophysiological studies of this particular patient are presented and discussed in detail in this paper. The surgeon removed the tumour located in the lower pole of the left parotid gland in this particular case. Ultrasound Doppler imaging revealed poorly demarcated and rich vascularised pathological structure 1.5 cm under the skin, 35 x 30 mm in size measured transversally and 40 mm sagittally in computer tomography, without signs of the destruction of the mandible and mastoid process (Figure 1). The anteromedial part of the tumour closely adhered to the left pterygoid muscle; a clear border between them was not visible in several scans. The tumour did not adhere to the masseter muscle, external auditory meatus and lymph nodes, as well as other salivary glands showed no pathologies.

Observations including clinical and neurophysiological examinations of the facial nerve function were performed preoperatively and 1, 3, 6 months postoperatively. Additionally, a patient with deep localisation of the tumour and facial nerve palsy underwent a neurophysiological examination of the facial nerve 17 months after nerve damage. The investigation took place as part of the routine diagnostics ordered by the physician. The stage of injury was also determined by the House-Brackmann (H-B) scale [10], which is used for the clinical evaluation of the degree of nerve damage in facial nerve palsy. The measurement is determined by analysis of the face muscle movement, symmetry, and muscle tone, with the presence of deficits like synkinesis, contracture of face muscle, or hemifacial spasm also determining the degree of nerve dysfunction. The scale is based on functional impairment, ranging between I (normal function) and VI (no movement, total paralysis).

Neurophysiological tests were recorded with a Keypoint system (Medtronic A/S, Skovlunde, Denmark). Standard (AgCl) 5 mm2 surface electrodes were used for recording compound muscle action potentials (CMAP) from frontal and orbicularis oris muscles (Figure 2 A and B) during ENG examination. The active electrode (cathode) was placed on the muscle belly, with the reference electrode positioned on contralateral frontal muscle or the chin, the ground electrode on the neck. A stimulating bipolar electrode was applied to the stylomastoid foramen area along the anatomical passage of the facial nerve [11]. The time base was set at 5 ms/D, recording sensitivity at 2 mV/D, 20 Hz upper and 10 kHz lower filters of recorder amplifier were used during ENG.

The intensity of the single electrical stimuli (with 0.2 ms duration and 1 Hz frequency) was increased from 30 mA to a value evoking the potential with the largest amplitude. Values of amplitude [12] and standardised latency [13-15] were analysed in CMAP recordings. The decrease of CMAP amplitude was the result of axonal loss (axonotmesis), whereas an increase in the standardised latency parameter was initiated by the demyelination process (neuropraxia). Twenty-one healthy volunteers (fourteen females and seven males) aged 21 to 51 years (mean age was 28) underwent neurophysiological tests to establish reference values.

The Blink reflex examination (BR) was used as a supplementary method for facial nerve testing. The blink reflex test is an essential element of a comprehensive neurophysiological assessment of the facial nerve peripheral damage with 81% sensitivity and 94% specificity. The improvement in parameters of the recorded potentials to the reference values suggests that the nerve impulse conduction disorder concerns the myelin sheath of the nerve without signs of axon damage. The blink reflex study is also an indicator of the improvement in the function of the facial nerve fibers. Recording electrodes were placed bilaterally on the orbicularis oculi muscles and a ground electrode on the neck. Electrical stimulus was applied over the supraorbital nerve (Figure 2 C) and required rectangular stimulus at 1 Hz with intensity of 20 mA and 0.5 ms duration. The short-latency ipsilateral R1 response and two long-latency R2 responses were analysed in recordings both ipsi- and contralaterally [11,16].

Statistical calculations for neurophysiological tests were performed using the STATISTICA version 9.1 programme (StatSoft Inc., 2009). The compliance of parameters with the normal distribution was checked by the Shapiro-Wilk test. Descriptive statistics of measurable parameters concerned minimum and maximum values, arithmetic means and standard deviation values. Simple comparison of parameters was used in all observation periods.

Confirmation of axonal changes in the facial nerve required needle EMG examination from the frontal and orbicularis oris muscles [11]. The study included the analysis of the muscle spontaneous activity at rest and evaluation of twenty motor unit action potentials (MUAPs) parameters (amplitude in mV, duration in ms and Size Index in mV/ms) recorded during weak voluntary contraction [15]. Two parameters of MUAPs with values greater than reference values determined the neurogenic advancement of their injury and reinnervation process in examined muscle. The last stage of EMG examination included measurements of amplitude and frequency of motor unit recruitment during maximal voluntary contraction lasting 5 seconds [17]. Conducting the two last stages of EMG examination was possible only in face muscles presenting voluntary activity.

The modified Blair’s skin incision was applied, with patients (N = 12) treated surgically by the extracapsular tumour removing technique. General anaesthesia with volatile anaesthetics was applied during the surgery in all patients, with no complications in patients during anaesthesia. Superficial parotidectomy with tumour removal was performed in one patient where the tumour was located in the lower parotid pole in the superficial and deep lobes. The facial nerve was identified and dissected using an anterograde technique due to its location in the parotid gland. Facial nerve branches were located on the surface of tumour and were “peeled off” during the intervention which lasted 180 minutes.

In one patient with the tumour located in both lobes, the pharmacological treatment included injections with Galantamine (5 mg per day, for 30 days), Cocarboxylase (50 mg per day, for 10 days), Dexamethasone (4 mg twice a day, for 10 days) and Triamcinolone (orally 4 mg twice a day, for 3 weeks). The above conservative treatment, as well as physiotherapy, was used because of iatrogenic facial nerve palsy which occurred in this particular patient.

Physiotherapeutic treatment was introduced two days after the surgery, performed twice a day for eight days. Two therapeutic procedures were applied, the Facial-Oral-Tract-Therapy (F.O.T.T.) [18], which stimulated muscle receptors on the paralysed side together with relaxation and stretching, as well as with muscle tension decrease on the asymptomatic side, and Proprioceptive neuromuscular facilitation (PNF) [19], which included activation of facial muscles on both sides to evoke their synergy, similar to the conditions of their normal innervation. Proper face muscle function on the asymptomatic side influences contractile properties of muscles on the paralysed side. The patient continued the exercises at home (2–3 times a day for 10 minutes) using a facial mirror biofeedback technique, focusing on the symmetry of facial movements. Supervised treatment was continued for the next 6 months.

The remaining patients did not receive the treatment, as mentioned earlier, due to the lack of complications after surgery in the form of facial nerve palsy.

Ethical considerations were according to the Helsinki Declaration. The research was approved by the Bioethical Committee of the University of Medical Sciences including studies on healthy people. Each patient was informed about the aim of study and provided written consent for examinations and data publication. Statistical analysis included comparison of mean values and ranges of parameters with reference to the tumour size and analysed results of neurophysiological studies.

In nine out of thirteen patients, the histopathological study revealed pleomorphic adenoma, adenolymphoma (Warthin’s tumour) in three patients and myoepithelioma was recognised in the preliminary histological study in one patient. The type of tumour was indicated by the immunohistochemical examination. Tumour was not malignant; its total excision was not confirmed. The anatomical tumour-nerve relation and surgery difficulty were factors determining the iatrogenic facial nerve injury.

In twelve patients, the clinical examination using the H-B scale showed the proper function of the facial nerve bilaterally (H-B I) in all periods of observations. One of the co-existing symptoms was analgesia of auricle; this symptom remitted 6 months after surgery.

The data presented in Table 1 concerns one patient with a tumour located in both lobes of the parotid gland, which shows that before the surgery, despite proper facial nerve function (H-B grade I), the only weakness of the orbicularis oculi muscle was found on the symptomatic side. Xerostomia, as well as numbness and itching of the left cheek, were reported by the patient. Tumour removal led to the total paralysis of left facial nerve fibers (grade VI of H-B scale), which was accompanied by additional symptoms such as hypoesthesia of auricle and numbness. Recovery of the facial nerve function (H-B grades IV, III and III, respectively) was found at the subsequent 3rd–5th periods of observation. Late complications such as Frey’s syndrome and oral-ocular synkinesis were detected 3 months after surgery (Table 1). The H-B score was IV, although the improvement of facial nerve function was detected in ENG (Figure 6) and BR examination (Table 3), as well as in the residual bioelectrical activity of mimic muscles (Table 4). Following (4th and 5th) periods of observation, the H-B score was III with further improvement of nerve function and muscle activity. Moreover, contracture of mimic muscles on the symptomatic side was observed in these observation periods. The patient reported regression of auricle hypoesthesia in the 5th period (Table 1).

Table 2 presents the normative values for clinical neurophysiology studies in healthy volunteers.

The results of the ENG and BR examination in twelve patients were within physiological limits (Table 2), with no significant differences between the symptomatic and asymptomatic sides in the four periods of observation.

In the case of one patient, data in Figure 3 shows that the amplitudes of CMAP evoked after stimulation of the frontal branch were comparable to the reference values at all periods of observation. In postoperative studies, CMAP amplitudes recorded after the marginal mandibular branch stimulation were always lower compared to reference values. Additionally, the dispersion of CMAP potential duration recorded from the orbicularis oris muscle was noted 6 months after surgery, which coexisted with mimic muscles contracture on the symptomatic side (Table 1). Values of CMAP standardised latencies after stimulation of both nerve branches were higher when compared to the controls, indicating decreasing conduction velocity of the nerve impulses in subsequent postoperative periods. Values of the standardised latency were normal 17 months after surgery only in studies of the frontal branch.

The value of CMAP amplitude evoked after stimulation of the frontal branch on the symptomatic side was 38% lower compared to the asymptomatic side one month after surgery. Despite the lack of difference in amplitude, above 50% between sides, the result of H-B score (VI), the presence of denervation signs and no voluntary activity in EMG examinations from frontal muscle confirmed an axonal injury within the frontal branch of the facial nerve. Furthermore, values of CMAP amplitudes were comparable to those recorded on the asymptomatic side in subsequent periods of observation. Amplitudes of CMAPs after stimulation of the marginal mandibular branch were lower at 86%, 58%, 47% and 27% in subsequent postoperative observation periods in comparison to the values recorded on the asymptomatic side.

According to Esslen [20], CMAP amplitude value reflects a 1:1 proportion of active axons within the examined nerve branch. In the case of one patient, we ascertained the progression of nerve regeneration based on the results of CMAP amplitude variability (expressed in percentages) after marginal mandibular branch stimulation in subsequent postoperative observation periods compared to the preoperative recording. These values were 14%, 43%, 49% and 68%, indicating the advancement of the regeneration process which was most prominent at 29% between the first and third month after surgery.

The Blink reflex examination showed proper parameters of evoked potentials only during the preoperative and the final observation periods (17 months after surgery; Table 3, Figure 4 AB1, 4). One month after tumour resection, R1 and R2 ipsilateral evoked potentials were not recorded on the symptomatic side (Figure 4 A2), nor was the R2 contralateral response after stimulation on the asymptomatic side (Figure 4 B2). This reflects the consequences of injury in the efferent pathway of the reflex arc (the facial nerve on the symptomatic side). In addition, low amplitude potentials with prolonged latencies in the BR study were recorded three months after surgery (Figure 4 AB3), indicating the recovery of nerve impulses transmission.

Data presented in Table 4 for a case of one patient proves the recovery of mimic muscles activity, which was observed 3 months after surgery. Residual voluntary activity of the frontal muscle and weak voluntary activity of the orbicularis oris muscle were recorded in the EMG examination. Analysis of MUAP parameters 6-months after surgery showed properties of neurogenic changes secondary to nerve injuries. Positive sharp waves and fibrillation potentials indicating the denervation of face muscles were recorded in the remaining postoperative observation periods. This activity was particularly frequent in the 1st and 3rd months after surgery, thus confirming the acute neurogenic injury. Signs of denervation process were weaker in other postoperative observation periods (Figure 5).

Assessment of EMG spontaneous activity 6 and 17 months after surgery was particularly complicated in terms of the orbicularis oris muscle due to the coexistence of oral-ocular synkinesis and facial muscles contracture (Figure 6). The synkinesis on the symptomatic side appeared after 3 months, whereas muscles contracture was present 6 months after the parotid gland tumour surgery (Table 1).

A decrease of asymmetry between the symptomatic and asymptomatic side was observed in the long-lasting postoperative observation, being visible during the total relaxation of facial muscles. During attempts of voluntary contraction, the asymmetry increased particularly in the lower part of face. Asymmetry was deepened by the muscle contracture on the symptomatic side six months after surgery. Weak voluntary activity and the clearly visible narrowing of the eyelid fissure during activation of orbicularis oris muscle (oral-ocular synkinesis) were observed.

Differentiation of the facial nerve injury type presented in this study based only on ENG results is difficult. CMAP amplitude may be reduced in the case of axonotmesis and conduction block. The latter refers to neuropraxia with decreasing axonal conductive properties at the site and above compression. Possibilities of regeneration differ in these types of injuries, according to Hughes et al. [21], 50% reduction of CMAP amplitude recorded proximally to the site of nerve block confirms this phenomenon. If the distal point of facial nerve stimulation overlaps with the site of parotid gland tumour surgery, ascertaining the conduction block is not possible. In our study, the analysis of facial nerve regeneration progress with its clear exponent of CMAP amplitude increase in the period between the 1st and the 3rd month suggests that the most probable mechanism of nerve injury was coexistence of conduction block and mild axonotmesis. Diminished conduction block in the 3rd month after surgery resulted in a significant increase of CMAP amplitude which, in the case of frontal branch examination, reached a value comparable to normal. The increase in CMAP amplitude recorded from the marginal mandibular branch in all periods of observation was also recorded, but this value has never reached the reference, confirming mild axonotmesis. The advancement in facial nerve regeneration was also observed in the results of the blink reflex examination, validated by recordings of R1 and R2 responses on the symptomatic side 3 months after surgery.

The rate and dynamics of changes recorded in the frontal and orbicularis oris muscles during EMG studies also indicate the possibility of coexistence of both axonotmesis and neuropraxia. The spontaneous activity at rest showing denervation and lack of muscle voluntary activity was observed 1 month after surgery. The appearance of voluntary activity of face muscles, which was the symptom of nerve regeneration with still existing denervation, were recorded 3 months after surgery. Similar to observations of Martin and Helsper [22], in our study, the first symptoms of reinnervation were better expressed in muscles located closer to the injured area. At the same time, oral-ocular synkinesis was also observed in EMG, suggesting an axonotmesis mechanism of nerve injury and reflects the abnormal proliferation of axons during regeneration. Considering the location of the tumour in the parotid gland and possible mechanical and thermal reasons of injury during surgery, it is likely that axonotmesis involved the marginal mandibular branch to a greater extent than the whole nerve trunk. This can be validated by a better function of the frontal than marginal mandibular branch ascertained in the neurophysiological examination. The marginal mandibular branch is located nearest to the area of surgery, hence it may be exposed to more damage than another branch of the facial nerve. Moreover, the marginal mandibular branch has the least anatomical connections between other branches of the facial nerve, so the regenerative capacity may be limited. Similar results pointing to the iatrogenic injury of the marginal mandibular branch during parotid gland tumour surgery were described in other studies [2,23].

The results presented in this study and findings of Reddy et al. [24] indicate the importance of physiotherapeutic treatment introduced immediately after surgery for creation of the appropriate conditions for facial nerve regeneration and muscle atrophy prevention. Spontaneous regeneration is successful when nerve branches undergo intensive afferent and efferent functional stimulation, consequently accelerating neurogenesis and activating the cortical centres in a feedback way. Exercises to activate face muscles promote regenerating axonal endings to create new connections at neuromuscular junction [25].

The study findings also suggest that precautions undertaken during surgery, including intraoperative monitoring of facial nerve function, and application of appropriate conservative treatment are essential for the final nerve regeneration. Neurophysiological examination performed 1, 3, 6 and 17 months after surgery allowed for objective determination of the facial nerve injury type, the progress of its regeneration and face muscles reinnervation. Comparative studies showed the predominant mild axonotmesis, especially with reference to the marginal mandibular branch. This type of injury was also confirmed by detecting post-regenerative deficits previously mentioned by other authors [26-29]. In this study, complications after surgical removal of the parotid gland tumour affected only one patient. The rarity of this phenomenon in the case of benign parotid tumours prompted the authors to present this case and propose optimal therapeutic and diagnostic processes.

In the case of axsonotmesis, spontaneous regeneration of the nerve may be incomplete, therefore, a detailed neurophysiological diagnosis is non-invasive and a valuable diagnostic and prognostic tool for assessing facial nerve function.