Main prognostic factors and physical therapy modalities associated with functional recovery in patients with peripheral facial paralysis

ABSTRACT Patients with peripheral facial paralysis (PFP) have some degree of recovery. The aim of this study was to evaluate prognostic factors and physical therapy modalities associated with functional recovery in patients with PFP. This is a cohort study with 33 patients. We collected the following variables of patients who underwent treatment at the rehabilitation center: age, sex, risk factors, affected side, degree of facial paralysis (House-Brackmann scale), start of rehabilitation, and therapy modality (kinesiotherapy only; kinesiotherapy with excitomotor electrotherapy; and kinesiotherapy with excitomotor electrotherapy and photobiomodulation therapy). The outcomes were: degree of facial movement (House-Brackmann) and face scale applied 90 days after treatment. Degree of PFP was associated with functional recovery (RR=0.51, 95% CI: 0.51-0.98; p=0.036). The facial movement was associated with the time to start rehabilitation (r=−0.37; p=0.033). Lower facial comfort was observed among women, worse ocular comfort was associated with diabetes mellitus, worse tear control with prior PFP, and worse social function with the degree of PFP. Our results indicate that the all modalities present in this study showed the same result in PFP. Recovery of PFP was associated with degree of nerve dysfunction, the length of time to onset of rehabilitation, female sex, hypertension, diabetes mellitus, and previous PFP, all of which were associated with worse outcomes on the face scale.


INTRODUCTION
Peripheral facial paralysis (PFP) is characterized by reduction or interruption of nerve conduction of the seventh cranial nerve to the muscles of the face, resulting in partial or complete paralysis of these muscles 1,2 . The incidence of PFP in the world population is approximately 20 to 30 cases per 100,000 people and is more prevalent in older adults [3][4][5][6][7] .
PFP is often idiopathic (i.e., Bell's palsy), but infection, inflammation, trauma, surgery, and tumor are additional causes [8][9][10] . The main risk factors for PFP are diabetes mellitus, pregnancy and postpartum, cold weather, hypertension, age, and immunodeficiency syndromes. Diagnostic criteria for PFP include sudden onset of unilateral complete or partial paralysis of motility of the facial muscles with no signs or symptoms of central nervous system (CNS) injury. PFP may be accompanied by disturbances in taste, salivation, and tearing; hyperacusis and hypoesthesia in the external auditory canal; and functional deficits in oral functions such as speech, chewing, sucking, swallowing, and lip hold. The degree of motor impairment is assessed via the House-Brackmann scale to quantify the degree of paresis at diagnosis [11][12][13][14][15][16][17][18] .
The prognosis for a patient with PFP depends on several factors, such as age, type of injury, etiology, nerve nurturing, neuromuscular impairment, therapeutic use, and precocity of rehabilitation treatment 19 . In general, all patients exhibit some degree of recovery, even without treatment, but treatment may accelerate and optimize recovery. Most cases of PFP require treatment with medication and physical and speech therapy. Studies suggest that treatments with antivirals, steroids, acyclovir, botulinum toxin, electrotherapy, and vitamin B12 are effective. There is also some evidence for the effectiveness of physical therapy or electrotherapy, but there are a few clinical trials for patients with PFP [19][20][21] .
Several patients with PFP perform different treatment modalities and present different associated risk and clinical factors. There is a gap in the literature in establishing which factors (treatments and clinical factors) are decisive in predicting worse or better outcome in the long term after PFP. Thus, the aim of this study was to evaluate the main prognostic factors and physical therapy modalities involved in the functional recovery of patients with PFP. The main hypothesis was that different types of treatment, comorbidities, anthropometric and clinical characteristics may interfere with long-term functional outcomes. edad, sexo, factores de riesgo, lado afectado, grado de parálisis facial (escala de House-Brackmann), inicio de la rehabilitación y modalidad de terapia (kinesioterapia, kinesioterapia asociada a la electroterapia, kinesioterapia asociada a la electroterapia y terapia de fotobiomodulación). Los resultados fueron: grado de movimiento facial (House-Brackmann) y face scale, aplicados 90 días después del tratamiento. El grado de PFP se asoció con la recuperación funcional (RR=0,51,95% IC: 0,51-0,98; p=0,036).

Study design, setting and participants
This study involved a retrospective and prospective cohort of patients admitted to the rehabilitation center of the Hospital of the Botucatu School of Medicine (HC-FMB) from January 2010 to January 2015. The study protocol was approved by the Research Ethics Committee of the Botucatu School of Medicine, UNESP (CAAE: 50507215.7.0000.5411). All participants or their legal representatives were aware of the study objectives and provided written informed consent.

Eligibility criteria
We evaluated individuals diagnosed with idiopathic PFP. Patients who presented with PFP secondary to cerebrovascular disease, brain tumor, Guillain-Barré syndrome, or parotitis were excluded. In addition, patients with otogenic facial paralysis due to otitis media or mastoiditis and patients with PFP due to trauma, surgery, or tumor were also excluded.

Measurements
We collected several variables from the electronic medical records of patients with PFP who underwent treatment at the rehabilitation center, including age, sex, risk factors, affected side, degree of facial paralysis (House-Brackmann scale), onset of PFP, onset of rehabilitation, and physical therapy modality.
The primary outcome was the degree of full or partial return of facial movement as measured by the House-Brackmann scale. The House-Brackmann scale grades the level of facial nerve injury in patients with PFP by evaluating the frontal muscles, the orbicularis of the eyes and mouth, the Risorius zygomatic muscle, and muscles of the upper lip and nose. The House-Brackmann scale ranges from grade I to grade VI, with grade VI representing the most severe facial nerve paralysis 22 .
Secondary outcomes were the scores on the six areas measured by the face scale 90 days after initial diagnosis. The face scale assesses movements on both sides of the face, the occurrence of certain events, and how the patient feels due to facial problems. The face scale is divided into six areas: facial movement, facial comfort, oral function, ocular comfort, lacrimal control, and social function. The total score ranges from 0 (worst) to 100 (best) 23 .

Independent variables
All participants were included to the facial rehabilitation protocol applied in the Botucatu Medical School. Patients in this study received one of three treatment types: (1) kinesiotherapy only; (2) kinesiotherapy with excitomotor electrotherapy; (3) kinesiotherapy with excitomotor electrotherapy and photobiomodulation therapy.
Kinesiotherapy consisted of active exercises (brow lift, gentle eye closure, open mouse smile, snarl, and lip pucker) and massage therapy in all muscles of the face in the direction of the fiber. Five sets of five repetitions of each exercise were performed with an interval of 1 minute between them to avoid synkinesis. The total of this modality was 20 minutes 24 .
Excitomotor electrotherapy was performed with Bernard's diadynamic currents with the positive pole positioned on the external acoustic meatus and the negative pole positioned on the frontal muscle. The metallic aluminum electrodes used had the standard size of 4.3 cm × 11.3 cm, covered by a damp sponge. The following parameters were used: 3 minutes of diphase fixe (DF) current (frequency of 100 Hz; pulse of 10 ms and no intervals), 3 minutes of simple impulse (MF-monophase) current (frequency of 50 Hz; pulse of 10 ms; and intervals of equal duration), and 3 minutes of short periods (CP) current (MF and DF alternating 1 s). The adjustment of the amplitude of the current for the groups was done in a standardized way, increasing 1 mA of current every second until the participant report strong but comfortable paresthesia. Stimulation started with positive pole with aim to guide fluid removal by endosmosis in acute phase of PFP and the polarities were reversed after 15 days 25,26 .
For photobiomodulation therapy, a gallium-arsenide diode (GaAs) laser (830 nm) which had the following specifications: 830 nm wavelength and 100 mW output power and 0.1 cm diameter, average energy density of 4 J/cm2, frequency of 1 KHz, and a duty cycle of 80% 27 . In all cases, the laser was in direct contact with the superficial roots of the facial nerve on the affected side and was applied for 15s per point for 8 points. The irradiance was 12.7 watts/cm and the energy fluence (energy per unit areas incident on the skin) was 191 joules/cm 27 .
The three treatment types were performed twice per week for three months. All patients received steroids/ antivirals within 48 hours of onset of symptoms of PFP.

Confounding factors
The main confounding factors that were evaluated in this study include age, sex, facial laterality, risk factors (hypertension, diabetes mellitus, smoking habit, alcoholism, hypercholesterolemia, hypothyroidism) previous facial paralysis, facial paralysis upon awakening, degree of paralysis at initial evaluation, and time between diagnosis and start of rehabilitation.

Statistical analysis
The relative risk (RR) of recovery of full movement in six months, as assessed by the House-Brackmann scale, was estimated by a Cox multiple logistic regression model adjusted for the degree of facial paralysis. The association between the different physical therapy treatment groups and the face scale scores was analyzed by the non-parametric Kruskal-Wallis test; no corrections for potential confounders were used in this analysis due to the small sample size. Results with p<0.05 were considered statistically significant. All analyses were performed in SPSS version 21.0 (IBM, Armonk, NY, USA).

RESULTS
In total, 103 patients were screened, of which 70 were excluded (central facial paralysis: 25; parotitis: 17; brain tumor: 5; otitis media or mastoiditis: 13; traumatic facial paralysis: 5; surgical facial paralysis: 3; tumor facial paralysis: 2), thus, 33 patients were included in the study (11 -kinesiotherapy only; 12 -kinesiotherapy with excitomotor electrotherapy, and 10 -kinesiotherapy with excitomotor electrotherapy and photobiomodulation therapy) and completed the clinical follow-up ( Figure 1). The clinical data and outcomes of the total of patients and the comparison between the groups of interventions are described in Table 1.
The RR of full recovery of facial movement at six months was associated with the degree of facial paralysis on admission (RR=0.51, 95% CI: 0.51-0.98; p=0.036).
No other clinical and anthropometric variables were related to the RR of full recovery ( Table 2).
The Cox multiple regression model, which was adjusted for the degree of facial paralysis on admission, revealed no statistically significant difference in the effect of treatment on recovery (Table 3).
: Fisher's exact test; (2) : Kruskal-Wallis test; G1: kinesiotherapy; G2: kinesiotherapy and excitomotor electrotherapy; G3: kinesiotherapy and excitomotor electrotherapy and photobiomodulation.   Table 4 reports the association between clinical and anthropometric characteristics and the face scale score. Longer time between the onset of PFP and the start of rehabilitation was associated with a lower face scale score for facial movement (r=−0.37; p=0.033). Sex was significantly associated with facial comfort such that women but not men exhibited significant improvement in the face scale score (p=0.034). Degree of paralysis was also associated with facial comfort such that a higher degree of paralysis was associated with lower facial comfort (r=−0.44; p=0.010). Patients with hypertension had lower oral function scores (p=0.008), whereas patients with diabetes mellitus had lower ocular comfort scores (p=0.003). Patients with a previous diagnosis of PFP had worse tear control (p=0.045). Degree of facial paralysis also was associated with social function such that a higher degree of paralysis was associated with lower social function (r=−0.56; p=0.01). The total score on the face scale was associated only with the time to onset of rehabilitation such that longer time between the onset of PFP and the start of rehabilitation was associated with a lower total face scale score (r=−0.37; p=0.035). There was no association between the type of physical therapy treatment and the scores obtained on each face scale domain.

DISCUSSION
In this study, the prognosis for recovery was significantly associated with PFP degree. This finding is consistent with other studies that have reported an influence of initial paralysis severity on prognosis 28 . One strength of this study is that we assessed the House-Brackmann scale both at the initial visit and after 6 months of treatment to establish the final recovery of the patient. The advantages of the House-Brackmann scale is that it is commonly used during the initial evaluation of PFP, it is easy to use, helps to establish the prognosis, and allows the assessment of the face during both movement and rest. The House-Brackmann scale is comparable to other scales to assess facial paralysis, and is being adopted as the gold standard by the Facial Nerve Disorders Committee of the American Academy of Otolaryngology -Head and Neck Surgery 29,30 . The evaluations of functional outcomes were conducted at 6-month follow-up because it is established as a period of optimal plasticity and used in different clinical studies.
In this study, the effect of treatment was not significantly different across treatment types for either the primary or secondary outcomes. Several physical therapy techniques are used for the treatment of PFP, including biofeedback, self-massage, relaxation, and electrotherapy, to restore trophism, muscle strength, and function. Prior studies have compared the use of electrical stimulation to isolated therapeutic exercise and found that neither treatment type exhibited statistical superiority in terms of complete facial recovery. In addition, the literature has not established the amount of exercise (and thus the amount of facial nuclear hyperexcitability or regeneration of aberrant nerve fibers) that could lead to consequences such as synkinesis and facial spasm 30,31 .
The face scale is a validated instrument and has been used to quantify the association between PFP and quality of life in patients. A lower score on the facial movement domain of the face scale was associated with a higher degree of facial paralysis at initial evaluation. This result is consistent with a previous study, in which higher scores on the House-Brackmann scale were associated with lower facial movement 32 . That study also found that muscle strength at diagnosis was associated with long-term poor motor control.
A study by Kleiss et al. 31 found that women were more psychologically affected by PFP and less satisfied with their appearance, leading to reduced social interaction and psychosocial dysfunction. In our study, women had lower scores on the facial comfort domain, which is consistent with the prior study and is believed to be due to the higher aesthetic demands of women.
A retrospective study observed that lagophthalmos and keratitis (tearing), which is also known as Bogorad syndrome or "crocodile tears", is a major eye complication after PFP. In this study, we found that the score on the tear control domain of the face scale was lower when the patient had previous facial paralysis 32 . Patients were not classified according to the topography of the facial nerve, but we can infer that the amount of facial paralysis may be related to a greater number of topographies in the path of the facial nerve, which is more likely to affect the lacrimal system.
Our study also found that social function was associated with the degree of paralysis such that a higher degree of paralysis was associated with worse social outcomes. Several studies corroborate this finding, highlighting the psychological and social impact of PFP on patients. PFP may affect social relationships, self-esteem, public behavior, work performance, and interpersonal communication, generating conflicting depressive feelings and leading to isolation and poor quality of life 32 .
The influence of diabetes mellitus and hypertension on the PFP prognosis are still unclear, but these clinical factors were associated with face scale domains in our study. Some authors reported a less favorable outcome for PFP in diabetic patients showing a clinical total palsy in 76% of diabetics and 45% of non-diabetics, and more denervation rate. Recent studies showed a no correlation between the duration of diabetes and the clinical severity of the PFP [33][34][35] . Hypertension can be the cause of facial paralysis. The appearance of an idiopathic PFP in combination with psychological stress may give rise to a hypertensive state. Vascular spasm and local edema can, in that case, influence oxygen and nutrient supply to the facial nerve 36 .
The main limitations of this study are its small sample size, the lack of patient randomization, the lack of analysis of electrophysiological findings, and the absence of synkinesis and facial spasm reports after the follow up. One strength of this study is that it is one of the few studies in Latin America to investigate prognostic factors and major treatments associated with the functional outcome of patients with PFP.

CONCLUSION
In conclusion, our results indicate that all modalities in this study showed the same result in PFP. Recovery of PFP was associated with several factors, including degree of nerve dysfunction, length of time to onset of rehabilitation, sex, hypertension, diabetes mellitus, and previous PFP, all of which were associated with worse outcomes in the face scale. Different treatment methods should be further evaluated through randomized clinical trials.