Rehabilitation of facial palsy Updated 2026

Clinical synthesis on peripheral facial palsy (Bell's palsy): recognising the presentation, ruling out a central cause, protecting the eye, and delivering rehabilitation that prevents synkinesis. Every reference has been individually verified on PubMed.

🙂 What are the fundamentals to know about facial palsy?

Before turning to rehabilitation, the physiotherapist must master a few simple but decisive landmarks: what peripheral facial palsy is, how often it will be encountered, what is believed about its mechanism, and above all what the facial nerve demands of us in order to assess, refer and protect the patient. These fundamentals govern both safety (recognising what is not Bell's palsy) and the relevance of the rehabilitation plan.

Definition: what are we talking about?

The Bell's palsy (also called idiopathic peripheral facial palsy, or "a frigore" palsy) is an acute, unilateral involvement of the facial nerve (7th cranial nerve) with no identified cause. It presents as weakness or paralysis of the hemiface that includes the forehead, in the absence of any other neurological abnormality ¹. This last point is no semantic detail: it is the core of the diagnostic reasoning, to which we shall return.

The term "idiopathic" means that, by definition, Bell's palsy is a diagnosis of exclusion : it is confirmed when examination finds neither central signs nor any obvious locoregional cause. It is also by far the most common form of facial nerve involvement ².

Epidemiology: a condition every physiotherapist will encounter

Bell's palsy is the most frequent diagnosis among facial nerve palsies and, more broadly, the most common acute mononeuropathy ². Its annual incidence is estimated at between 11.5 and 53.3 cases per 100,000 people depending on the populations studied, across all ages and sexes ² ; North American data more commonly cite a range of 20 to 30 per 100,000 per year ¹.

These figures translate into a concrete clinical reality: peripheral facial palsy is not a curiosity, it is a situation the physiotherapist will encounter regularly, at various stages, from the acute phase referred early to the chronic patient carrying sequelae established for months.

Pathophysiology: the dominant viral hypothesis

The exact cause remains by definition unknown (hence "idiopathic"), but the dominant pathophysiological hypothesis is reactivation of herpes simplex virus type 1 (HSV-1) at the level of the geniculate ganglion of the facial nerve ². This viral reactivation, centred on the geniculate ganglion, provides a coherent explanatory framework for the inflammatory involvement of the nerve, even though it does not account for every case.

For the clinician, this hypothesis has a twofold practical consequence. First, it sheds light on the, nuanced, place of antivirals, whose benefit remains marginal and never as monotherapy (see the treatment section). Second, it is a reminder that Bell's palsy is the result of acute nerve injury : it is the quality of the recovery and reinnervation that follow which will determine final function and any sequelae.

A favourable spontaneous prognosis… but not for everyone

Typical Bell's palsy carries a good prognosis. More than two thirds of patients recover completely and spontaneously, and this rate reaches up to 90 % in children and pregnant women ¹. Most cases resolve within 4 to 6 months and almost always resolve completely within a year ² ; between 66 % and 85 % of patients recover within the first three weeks ¹.

The largest published longitudinal series refines this trajectory. Across 2,570 peripheral facial palsies followed for 25 years (including 1,701 cases of Bell's palsy), Peitersen reports that recovery begins within 3 weeks in 85 % of patients, and between 3 and 5 months in the remaining 15 %. A normal facial movement is found in 71 % of patients; mild sequelae persist in 12 %, moderate in 13 % and severe in 4 %. Contractures and associated movements (synkinesis) are observed in 17 % and 16 % of patients respectively ⁴.

In other words: while the majority recover without sequelae, close to one patient in three retains a sequela, synkinesis being the most frequent. It is precisely this minority (severe forms, incomplete recoveries, established sequelae), that constitutes the core target of rehabilitation. Knowing this prognosis avoids two symmetrical pitfalls: over-treating patients destined to recover on their own, and underestimating those who will retain a lasting deficit.

Bell's palsy most often resolves on its own; the physiotherapist's task is not to "cure" everyone, but to identify and support those who will retain a deficit or synkinesis.

The share of Bell's palsy among facial palsies

Bell's palsy dominates the landscape of facial nerve involvement, but it is not the only cause. It remains a diagnosis of exclusion : the role of the first professional to examine the patient, physiotherapist included, is to recognise what might not be Bell's palsy. The fundamental distinction, the one that governs emergency referral, sets peripheral involvement against central.

What the physiotherapist needs to understand about the facial nerve

The facial nerve is the motor nerve of facial expression. Three anatomo-clinical points are enough to structure the assessment and make management safe.

1. The peripheral / central red flag: the forehead is the key. The frontalis muscle receives bilateral cortical innervation. Direct consequence: a central lesion (supranuclear, stroke, tumour) paralyses the lower and middle hemiface but spares the forehead, whereas a peripheral lesion (Bell's palsy) affects the whole hemiface, including the forehead ⁵. A patient who can still wrinkle the forehead and close the eye tightly on the affected side must therefore be assessed for a central lesion and referred urgently ¹. This simple test: "can the patient raise the eyebrow on the paralysed side?", is one of the most useful manoeuvres in the physiotherapist's repertoire.

2. The eye is the organ to protect first. The facial nerve innervates orbicularis oculi; its failure leads to lagophthalmos (inability to close the eye) and corneal exposure ⁶. Any patient unable to close the eye must be taught eye protection measures: artificial tears or lubricating gel, and taping the eyelid closed at night ¹. Initial management rests on artificial tears and a lubricating ointment; severe exposure keratitis may be complicated by loss of corneal sensation leading to a neurotrophic corneal ulcer ⁶. Checking and protecting the eye is a non-negotiable reflex, whatever the stage of management.

3. Aberrant reinnervation explains synkinesis. After nerve injury, axonal regrowth may be "misdirected": fibres destined for one muscle reinnervate another. The result is synkinesis: involuntary co-contractions accompanying a voluntary movement, typically eye closure on smiling. These are the most frequent sequelae of peripheral facial palsy ⁴, and they are a major target of specialised facial neuromuscular rehabilitation. Understanding that synkinesis is not a "lack of strength" but a wiring error radically changes the logic of rehabilitation: the aim is to inhibit and re-coordinate, not to strengthen.

Grade in order to monitor. The severity of the involvement and its course are documented with standardised scales. The House-Brackmann scale, described in 1985 and adopted as the reference by the American Academy of Otolaryngology, grades facial nerve function into six grades, from grade I (normal function) to grade VI (complete palsy, no movement) ⁷. In rehabilitation, the Sunnybrookscale, which is more sensitive, is particularly useful for quantifying synkinesis and demonstrating progress.

🚩 Peripheral or central? The triage that saves

Faced with a drooping hemiface, the first decision is not a rehabilitation one: it is diagnostic. Not every acute facial palsy is Bell's palsy, and confusing peripheral involvement (the facial nerve itself) with central involvement (the upper motor neurone, upstream of the nucleus) can mean missing a life-threatening emergency. The physiotherapist, often on the front line or close behind it, must be able to recognise what falls outside the picture and redirect without delay.

Idiopathic peripheral facial palsy (Bell's palsy, or "a frigore" palsy) remains the most frequent cause. Its annual incidence is estimated at between 11.5 and 53.3 cases per 100,000 people depending on the population ², a range that other sources narrow to around 20 to 30 per 100,000 ¹. It is the most common acute mononeuropathy, and its dominant pathophysiological hypothesis is reactivation of herpes simplex virus type 1 around the geniculate ganglion ².

The forehead settles it: the sign that separates peripheral from central

The key to triage lies in one neuroanatomical detail. The frontalis muscle receives bilateral cortical innervation: the lower part of the facial nucleus, which controls the lower and middle hemiface, is supplied only by crossed fibres from a single hemisphere ⁵. Practical consequence: a central (supranuclear) lesion paralyses the lower face but spares the forehead, whereas a peripheral lesion affects the whole hemiface, forehead included.

The clinical rule is explicit: a patient able to close the eye tightly and wrinkle the forehead on the paralysed side must be assessed for a central lesion ¹. Conversely, Bell's palsy is defined as acute unilateral facial weakness involving the forehead (inability to raise the eyebrow or close the eye), in the absence of any other neurological abnormality ¹. This "forehead affected + isolated" pattern is the foundation of the positive diagnosis.

Grading severity: House-Brackmann and Sunnybrook

Once peripheral involvement is confirmed, it must be quantified, to establish a prognosis, guide rehabilitation and monitor progress. The reference scale is the House-Brackmannscale, described in 1985 and adopted as the standard by the American Academy of Otolaryngology: it grades facial nerve function into six grades, from grade I (normal function) to grade VI (complete palsy, no movement) ⁷.

Simple and robust for a global assessment, House-Brackmann nonetheless shows its limits when it comes to demonstrating synkinesis, those involuntary movements accompanying a voluntary movement. This is where the Sunnybrook scale comes in (Sunnybrook Facial Grading System), finer and regional, widely used in rehabilitation to quantify symmetry, voluntary movement and synkinesis ⁷⁸. It is on the Sunnybrook, moreover, that the gains of specialised facial rehabilitation are measured: mime therapy improves symmetry by 20.4 points (95 % CI 10.4-30.4) in long-standing paresis ⁸.

Red flags: when it is not (only) Bell's palsy

Bell's palsy is a diagnosis of clinical exclusion : it presupposes acute unilateral involvement, including the forehead, with no other neurological sign ¹. Anything that goes beyond this definition should raise the alarm and prompt referral.

The most concrete short-term risk is ocular. Loss of orbicularis oculi function leads to lagophthalmos and corneal exposure ⁶. Any patient unable to close the eye must be taught protective measures: artificial tears or lubricating gel, and taping the eyelid closed at night ¹. If neglected, severe exposure keratitis may be complicated by loss of corneal sensation leading to a neurotrophic corneal ulcer ⁶ : signs of keratitis warrant referral to an ophthalmologist ¹.

This eye-protection reflex is, moreover, one of the strong clinical practice recommendations (AAO-HNS), on the same footing as early oral corticosteroid therapy ⁹. It depends on no uncertainty: it is a systematic safety measure as soon as eye closure is impaired.

Reassuring without trivialising

Triage does not mean dramatising. The spontaneous prognosis of typical Bell's palsy is broadly favourable: more than two thirds of patients recover completely without treatment, and this rate reaches up to 90 % in children and pregnant women ¹. In the largest published longitudinal series (2,570 peripheral palsies followed over 25 years, including 1,701 cases of Bell's palsy), recovery begins within 3 weeks in 85 % of patients ⁴.

This favourable natural history has two virtues: it allows reassurance, and it is a reminder not to over-treat. But it must not mask the minority who retain sequelae: close to one patient in three in Peitersen's series, with contractures in 17 % and synkinesis in 16 % ⁴. It is precisely this fringe that will justify, later in management, targeted rehabilitation and early medical treatment.

Diagnostic triage also directly determines the therapeutic window. The medical mainstay, early oral corticosteroid therapy, has a demonstrated effect only when started within 72 hours ¹⁰. Confusing central involvement with Bell's palsy, or being slow to identify a genuine peripheral palsy, therefore means either delaying an emergency or missing the window in which treatment changes the prognosis. The right initial move is not to rehabilitate: it is to triage quickly and well.

📈 What prognosis? What can be said to the patient

This is the patient's first question, often asked with the face still frozen and the fear of "staying like this" deep in their eyes. The honest answer comes in two parts: the prognosis of idiopathic peripheral facial palsy (Bell's palsy) is broadly favourable, but it is not universal. Our role is neither to promise full recovery nor to dramatise: it is to give reliable figures, credible timeframes, and to name the possible sequelae so that the patient knows what to expect, and so that rehabilitation finds its proper place.

A favourable spontaneous prognosis

Bell's palsy is the most common acute mononeuropathy, with an incidence estimated at between 11.5 and 53.3 cases per 100,000 people per year depending on the population ². Its natural history is rather reassuring: more than two thirds of patients with a typical form recover completely, with no treatment at all ¹. Most cases resolve within 4 to 6 months and almost always resolve completely within a year ².

Some profiles do even better: in children and pregnant women, the rate of complete recovery reaches up to 90 % ¹. This figure is important to know, not in order to trivialise, but to avoid over-treating cases destined to recover on their own and to reassure with solid data rather than empty phrases.

Timeframes: when recovery begins

The largest published longitudinal series is the reference here: 2,570 peripheral facial palsies followed over 25 years, including 1,701 cases of Bell's palsy ⁴. It provides a valuable timeline for guiding the patient:

• In 85 % of patients, recovery begins within the first 3 weeks ;
• In the remaining 15 %, it starts later, between 3 and 5 months.

This time to onset is itself a prognostic indicator: an early return of movement, within the first month, generally heralds a favourable course. Conversely, a face that remains completely frozen beyond three weeks calls for closer monitoring and for not promising a return to normal. In practical terms, you can tell the patient: "In the great majority of cases, the first signs of recovery appear within three weeks. If they are slow to come, it does not mean that nothing will move, but recovery will be longer and sometimes incomplete."

Factors that weigh on the prognosis

Several elements modulate the course, and it is useful to share them honestly with the patient rather than let them believe it is a lottery:

• Initial severity. It is graded with the House-Brackmann scale, the historical reference validated by the American Academy of Otolaryngology, which grades facial nerve function from I (normal function) to VI (complete palsy, no movement) ⁷. The more complete the initial involvement, the greater the risk of imperfect recovery; it is this initial grade that shapes follow-up.
• Patient characteristics. Young age and pregnancy are associated with better rates of complete recovery ¹.
• How early medical treatment is started. Oral corticosteroid therapy started within 72 hours concretely alters the prognosis (see below).
• Time to return of movement, described above: an early return is a good sign.

One point is worth repeating to the anxious patient, because it governs safety: Bell's palsy affects the whole hemiface, forehead included. A patient who retains the ability to wrinkle the forehead and close the eye tightly on the affected side does not fit this diagnosis and must be assessed for a central lesion (stroke, tumour), which spares the forehead thanks to the bilateral cortical innervation of the frontalis muscles ¹⁵. This is not a prognostic factor for Bell's palsy, but a red flag that radically changes management.

Sequelae and synkinesis: close to one patient in three

This is the part one is sometimes tempted to leave unsaid, and which on the contrary must be stated clearly. In Peitersen's series ⁴, normal facial movement is found in 71 % of patients. In other words, close to one patient in three retains a sequela. In detail:

The synkinesis is the emblematic sequela of peripheral facial palsy: an involuntary movement accompanying a voluntary one: the eye that closes when the patient smiles, the corner of the mouth that pulls when they blink. It results from aberrant reinnervation of the facial nerve during regrowth: regenerated fibres do not always find their original muscle again ³. It is precisely this late complication that rehabilitation seeks to prevent, then to correct once established. It can be explained simply to the patient: "The nerve will grow back, but it can take the wrong path. That is what produces these involuntary movements; we now know how to limit them."

What treatment changes about the prognosis

The prognosis is not fixed: early treatment shifts it. Oral corticosteroid therapy started within 72 hours is the recognised mainstay. In Sullivan's landmark double-blind randomised trial ¹⁰, complete recovery at 3 months was 83.0 % with prednisolone versus 63.6 % without, and 94.4 % versus 81.6 % at 9 months. The Cochrane review confirms this at the highest level of evidence: 17 % incomplete recovery with corticosteroids versus 28 % without (relative risk 0.63; 95 % CI 0.50–0.80), that is about 10 patients to treat to prevent one incomplete recovery ¹¹.

A frequently forgotten fact, and one directly relevant to the physiotherapist: corticosteroids also reduce long-term motor synkinesis (RR 0.64; 95 % CI 0.45–0.91), that is the very sequela that rehabilitation will then seek to contain ¹¹. Aciclovir alone provides no benefit ¹⁰, and antivirals can only be considered as an add-on to corticosteroids, where they might reduce long-term sequelae (synkinesis, "crocodile tears": RR 0.56; 95 % CI 0.36–0.87) without clearly improving the recovery rate ¹².

On the rehabilitation side, there is a genuine role but the level of evidence must be stated without overselling it: the Cochrane review finds low-quality evidence that individualised ("tailored") facial exercises improve function, particularly in moderate palsies and chronic cases ¹³. A major point for prognosis: these targeted exercises significantly reduced the occurrence of synkinesis after an acute palsy (RR 0.24; 95 % CI 0.08–0.69) ¹³. In long-standing paresis, mime therapy (massage, relaxation, inhibition of synkinesis, coordination and expression) improves facial symmetry by 20.4 points on the Sunnybrook scale (95 % CI 10.4–30.4) ⁸. And once synkinesis is established, botulinum toxin type A is a reference treatment, alongside rehabilitation ³¹⁴.

💊 Corticosteroids and antivirals: what does the evidence say?

Bell's palsy is the most common acute mononeuropathy, with an annual incidence estimated at between 11.5 and 53.3 cases per 100,000 people depending on the population ². Its spontaneous prognosis is broadly favourable, more than two thirds of patients recover completely without treatment, and this rate reaches 90 % in children and pregnant women ¹. Yet this reassuring natural history must not obscure the fact that there is a narrow therapeutic window, and that the patient's fate, complete recovery or lifelong sequelae, is largely decided in the first 72 hours. The physiotherapist, often placed as an advanced observer along the care pathway, has a decisive role: recognising the urgency and directing the patient to medical prescription without delay.

Early corticosteroids: the demonstrated mainstay

Oral corticosteroid therapy started early is the first-line treatment. The landmark Scottish double-blind randomised trial ¹⁰compared prednisolone, aciclovir, the combination of both and placebo, in patients treated within 72 hours of symptom onset. The result is clear-cut: recovery of normal facial function at 3 months was 83.0 % with prednisolone versus 63.6 % without, and at 9 months 94.4 % versus 81.6 % (p < 0.001). In other words, treating early with corticosteroids shifts nearly one patient in five from residual weakness to complete recovery.

This efficacy is confirmed at the highest level of evidence by the Cochrane meta-analysis ¹¹ : 17 % (79/452) incomplete recovery with corticosteroids versus 28 % (125/447) in the control group, a relative risk of 0.63 (95 % CI 0.50-0.80), with no excess of adverse effects. The number needed to treat to prevent one incomplete recovery is 10.

A crucial point for the rehabilitation practitioner: corticosteroids do not only prevent motor deficit, they also reduce long-term motor synkinesis at follow-up ¹¹. And these involuntary co-contractions, frequent sequelae of aberrant reinnervation, are precisely the target that physiotherapy will subsequently try to limit. Well-conducted corticosteroid therapy upstream therefore means less desynkinesis work downstream. The reference clinical practice guideline ⁹ is explicit: clinicians should prescribe oral corticosteroids within 72 hours of symptom onset.

Antivirals: a debated and overestimated benefit

The pathophysiological rationale, herpes reactivation, long raised hopes of a benefit from antivirals. The evidence has tempered that hope. In Sullivan's trial ¹⁰, aciclovir alone provided no significant benefit : 71.2 % recovery versus 75.7 % without aciclovir (adjusted p = 0.50). Antiviral monotherapy is therefore probably inferior to the corticosteroid alone ¹².

That leaves the question of the combination. The dedicated Cochrane review ¹² shows that adding an antiviral to corticosteroids has little or no effect on the rate of incomplete recovery compared with corticosteroids alone (RR 0.81; 95 % CI 0.38-1.74; 3 trials, N = 766; low-certainty evidence). Here one must remain honest about the uncertainty: the wide confidence interval reflects inconclusive data, not proof of uselessness. There is, however, a signal on another outcome: antivirals plus corticosteroids reduce the proportion of patients developing long-term sequelae (synkinesis, gustatory lacrimation known as "crocodile tears"), compared with placebo plus corticosteroids ¹².

The practical synthesis: the corticosteroid is the mainstay, the antiviral a possible add-on whose interest lies more in preventing sequelae than in motor recovery. The Baugh guideline ⁹ is clear on the point that matters most for triage: clinicians should not prescribe an antiviral alone.

What the physiotherapist needs to know: the 72-hour urgency

The operational message fits in one sentence: medical treatment of Bell's palsy is a relative emergency. The 72-hour window is not an administrative convenience, it is the condition under which corticosteroids have demonstrated efficacy. A patient who consults their physiotherapist for a "drooping mouth" or an eye that no longer closes, which appeared the day before, must not be left waiting: they need a medical opinion the same day so that corticosteroid therapy can be started in time. Losing 48 hours "observing" means risking falling outside the window in which the benefit exists.

Two clinical reflexes frame this referral.

1. The peripheral/central red flag. Bell's palsy affects the whole hemiface, forehead included: the patient can neither raise the eyebrow nor close the eye on the affected side. The frontalis muscle receives bilateral cortical innervation: a patient able to wrinkle the forehead and close the eye tightly on the paralysed side does not have peripheral involvement but a central lesion (stroke, tumour), which calls for urgent management of a completely different nature ⁵¹. This simple test of the spared forehead must be systematic: it radically changes the referral pathway.

2. Eye protection. Any patient unable to close the eye must be taught protective measures without delay: artificial tears or lubricating gel during the day, taping the eyelid closed at night ¹⁹. Loss of orbicularis oculi function leads to lagophthalmos and corneal exposure; severe exposure keratitis may progress to a neurotrophic corneal ulcer through loss of sensation ⁶. This is a preventable complication, and the physiotherapist is often the first professional to see the patient regularly again: checking eye protection at every session is a safety measure, not a detail.

Placing rehabilitation within the care strategy

Understanding the medical treatment also means understanding what physiotherapy does not have to do. Since more than two thirds of typical Bell's palsies resolve spontaneously ¹ and corticosteroids improve this prognosis further, rehabilitation does not aim to "bring about recovery" in cases destined to resolve on their own: it targets severe forms, chronic cases and the prevention of synkinesis. The Cochrane review on physiotherapy ¹³ finds no high-quality evidence, but low-quality evidence that tailored facial exercises improve function, and a notable signal on prevention: significantly fewer patients developed synkinesis after targeted exercises (RR 0.24; 95 % CI 0.08-0.69). The Baugh guideline ⁹ remains cautious: for lack of sufficient evidence, no formal recommendation can be made on physiotherapy. This should be told to the patient honestly, without overpromising.

Medical treatment and rehabilitation are therefore not in opposition: they take over from each other. Corticosteroids early to maximise recovery and reduce synkinesis; then, in forms that leave sequelae, specialised facial rehabilitation focused on symmetry and desynkinesis. Recognising the 72-hour urgency and referring quickly remains, for the physiotherapist, the most evidence-based contribution at this early stage.

🎭 Which facial rehabilitation should be offered?

The question of rehabilitation in peripheral facial palsy must be framed clearly. The spontaneous prognosis is broadly favourable: more than two thirds of patients with typical Bell's palsy recover completely, and the rate reaches up to 90 % in children and pregnant women ¹. In the largest published longitudinal series (2,570 palsies followed over 25 years), normal facial movement is found in 71 % of patients, with recovery beginning within 3 weeks in 85 % of them ⁴. In other words, a large share of palsies resolve on their own, and rehabilitation must never be presented as the driver of this natural recovery.

The fact remains that close to one patient in three retains a sequela: contractures (17 %), synkinesis (16 %), mild to severe residual paresis ⁴. This is precisely where facial physiotherapy has its place: less in the recovery of cases destined to resolve than in moderate to severe forms, chronic cases, and above all the prevention and treatment of sequelae.

What the evidence says: honest and modest

It must be stated plainly to the reader: facial rehabilitation has no high-level evidence behind it. The dedicated Cochrane review finds no high-quality evidence of benefit, or of harm, from physical therapies taken as a whole in idiopathic facial palsy ¹³. The American clinical practice guidelines point the same way: for lack of sufficiently good evidence, no formal recommendation can be made on the place of physiotherapy ⁹.

This caution does not mean "do nothing". The same Cochrane review finds low-quality evidence that tailored facial exercises improve facial function, particularly in moderate palsies and chronic cases, and reduce sequelae in acute cases ¹³. The key word is tailored : these are targeted, guided exercises, the opposite of maximal, forced or global analytical work, which has no basis and could maintain co-contractions.

The most interesting result of this review concerns prevention: significantly fewer patients developed motor synkinesis after a programme of targeted exercises (relative risk 0.24; 95 % CI 0.08–0.69) ¹³. The aim of rehabilitation is therefore as much the prevention of aberrant reinnervation as pure motor recovery.

Facial neuromuscular therapy (mime therapy)

The best-supported specialised facial rehabilitation is Beurskens' mime therapy : a neuromuscular approach combining massage, relaxation, inhibition of synkinesis, and exercises in coordination and emotional expression. In a randomised controlled trial in 50 patients with long-standing paresis (more than 9 months), three months of mime therapy improved facial symmetry by 20.4 points on the Sunnybrook Facial Grading System (95 % CI 10.4–30.4) and reduced severity by 0.6 grade on the House-Brackmann scale (95 % CI 0.1–1.1), independently of age, sex and duration of the paresis ⁸.

This is the evidence base for sequelae-oriented facial rehabilitation, and it concerns long-standing paresis, which usefully reframes the timeline: neuromuscular therapy retains value even long after the acute episode, in patients one might think of as "stable".

This approach fits into a now formalised therapeutic hierarchy. The recommendations of an international ENT scientific group place, for post-paralytic synkinesis, facial rehabilitation (biofeedback) as first line, followed by botulinum toxin, with surgery reserved for first-line failures ¹⁵. Physiotherapy is therefore the foundation, and toxin a targeted add-on, not the other way round.

Biofeedback and fine analytical work

The guiding principle of analytical work in facial rehabilitation is not strength but control. Faced with a muscle recovering in a disorganised way, the aim is to restore selective, symmetrical movements while inhibiting parasitic co-contractions. Biofeedback, visual (mirror) or electromyographic, supports this learning of dissociation: relearning how to move one region of the face without dragging another along with it.

This framework explains why mime therapy rests as much on inhibition of synkinesis and on relaxation as on strengthening: in a patient whose eye closes when smiling, asking for "more smile, harder" worsens the synkinesis. Fine work means breaking down, slowing down, dissociating: a logic consistent with the Cochrane signal favouring tailored exercises over maximal ones ¹³.

Progression: when and how?

Progression is reasoned according to phase and severity, graded with standardised scales. The House-Brackmann scale, the historical reference validated by the American Academy of Otolaryngology, grades facial nerve function from I (normal) to VI (complete palsy) ⁷. The Sunnybrook scale, more sensitive to synkinesis, is widely used in rehabilitation for fine-grained monitoring.

• Acute / flaccid phase : priority to eye protection and education; no forced analytical exercises. The aim is to avoid maintaining abnormal patterns while reinnervation is taking place.
• Recovery phase (moderate forms) : introduction of tailored, targeted facial exercises with feedback, to improve function and prevent synkinesis ¹³.
• Sequelae / chronic phase : neuromuscular therapy of the mime therapy type, centred on symmetry and inhibition of established synkinesis ⁸ ; recourse to botulinum toxin as an add-on if necessary.

Established sequelae: the place of botulinum toxin

Synkinesis that is already established (for example involuntary eye closure on smiling) results from aberrant reinnervation and responds poorly to exercises alone. Botulinum toxin type A is the reference treatment here, alongside rehabilitation. A systematic review with meta-analysis (34 studies, 1,314 patients) finds a significant improvement in synkinesis scores measured with the Synkinesis Assessment Questionnaire (mean difference 11.6; p < 0,01), les auteurs recommandant un traitement individualised rather than a standardised protocol ³. A clinical trial confirms this functional improvement, with a fall in the SAQ synkinesis score from 46.22 to 37.55 after injection (p = 0.001) and an improvement in quality of life ¹⁴.

Electrical stimulation: stating the uncertainty

This is the most disputed point, and it would be dishonest to settle it. The guidelines do not recommend electrical stimulation, and the historical Cochrane review found no benefit from it on incomplete recovery at six months compared with placebo ¹³. The classic fear is that, applied in the acute phase, it promotes synkinesis.

Recent data qualify this picture without overturning it. One meta-analysis finds that, added to usual care within 7 days, electrical stimulation reduces the risk of incomplete recovery (RR 0.65; 95 % CI 0.48–0.88; moderate certainty), while stressing that its efficacy and safety in the acute phase remain controversial and insufficiently studied ¹⁶. Other syntheses point towards a possible benefit on overall function without a significant excess risk of synkinesis, but with marked heterogeneity in the measures used ¹⁷¹⁸.

The honest conclusion for practice: the data are contradictory and the level of evidence insufficient to recommend electrical stimulation in routine practice. Where some teams see an encouraging signal with early, parameterised application, others find no difference in synkinesis. Caution, patient information, and no systematic use.

Never forget: the eye first, and the red flag

Before any exercise programme, one priority dominates: eye protection. Any patient unable to close the eye must be taught protective measures (artificial tears or lubricating gel, taping the eyelid closed at night), because loss of orbicularis function leads to lagophthalmos and corneal exposure, which may be complicated by exposure keratitis and then a neurotrophic corneal ulcer ¹⁶. No rehabilitation session takes precedence over this instruction.

Finally, the physiotherapist must keep the diagnostic red flag in mind. Peripheral facial palsy affects the whole hemiface, forehead included. A patient able to wrinkle the forehead and close the eye tightly on the affected side, forehead spared, has central involvement (stroke, tumour) and not Bell's palsy, because the frontalis muscle receives bilateral cortical innervation ¹⁵. This abnormality calls for an urgent opinion, not rehabilitation.

⚡ Synkinesis: preventing it, and the mistakes not to make

Motor recovery is not the only issue in peripheral facial palsy. In close to one patient in three, healing leaves sequelae, and the most frequent of them has a name: synkinesis. In the largest published longitudinal series (2,570 peripheral facial palsies followed over 25 years), associated movements were found in 16 % of patients and contractures in 17 % ⁴. It is precisely here that rehabilitation plays its subtlest part: not pushing the muscle to work harder, but helping it to reorganise correctly. And it is also here that certain well-intentioned habits do the most damage.

The mechanism: reinnervation that goes to the wrong address

Synkinesis is an involuntary movement that accompanies a voluntary one: the eye that closes when the patient smiles, the corner of the mouth that contracts when they blink, the cheek that tightens on swallowing. It results from aberrant reinnervation of the facial nerve ³. After axonal injury, nerve fibres grow back, but they do not always find their original target: axons destined for orbicularis oculi reinnervate fibres in the perioral region, and vice versa. The result is a genuine wiring error: a single motor intention now activates several muscle territories that should remain independent.

This pathophysiological fact is decisive for rehabilitation. Unlike ordinary muscle weakness, synkinesis is not a deficit of strength: it is a failure of selectivity. A synkinetic muscle is often overactive, not lazy. The whole logic of intensive analytical strengthening is thereby overturned.

Prevention: gentle, selective and early rehabilitation

Good news: physiotherapy can prevent synkinesis, provided it is designed in the right spirit. The Cochrane review devoted to rehabilitation in Bell's palsy reports a major result for prevention: significantly fewer patients developed motor synkinesis after a programme of targeted facial exercises, with a relative risk of 0.24 (95 % CI 0.08–0.69) ¹³. Well-conducted exercise therefore divides the risk of this sequela by four in the cases concerned.

The key word is tailored: made to measure. The same review finds low-quality evidence that tailored facial exercises improve function, particularly in moderate palsies and chronic cases ¹³. It is therefore not a matter of working the face "in general", but of training precise, isolated movements below the threshold that triggers co-contractions. Rehabilitation aims to teach the brain to command each territory separately again.

Beurskens' mime therapy embodies this philosophy. It is a facial neuromuscular rehabilitation combining massage, relaxation, inhibition of synkinesis, and exercises in coordination and emotional expression. In a randomised controlled trial in 50 patients with long-standing facial paresis (more than 9 months), three months of mime therapy improved facial symmetry by 20.4 points on the Sunnybrook Facial Grading System (95 % CI 10.4–30.4) and reduced severity by 0.6 grade on the House-Brackmann scale (95 % CI 0.1–1.1), independently of age, sex and duration ⁸. This protocol explicitly includes a synkinesis-inhibition component: it is not strengthening, it is retraining of control.

Prevention even begins before physiotherapy. Early oral corticosteroids, the mainstay of medical treatment, do not merely improve overall motor recovery: the Cochrane meta-analysis shows that they also reduce long-term motor synkinesis (RR 0.64; 95 % CI 0.45–0.91) ¹¹. Adding an antiviral to the corticosteroid points the same way for late sequelae, synkinesis and "crocodile tears", with an RR of 0.56 (95 % CI 0.36–0.87) versus corticosteroid alone, even though its effect on incomplete recovery remains little or not demonstrated ¹². The prevention window therefore opens in the very first days, and early medical management already does part of the work.

Treating established synkinesis: physiotherapy first, toxin next

When synkinesis is already established, the therapeutic hierarchy is well established. The practice recommendations of an international ENT scientific group place facial rehabilitation (biofeedback) as first line, followed by botulinum toxin, with surgery reserved for first-line failures ¹⁵. Physiotherapy remains the foundation; toxin is a targeted add-on, not a shortcut.

The botulinum toxin type A is the reference treatment for residual synkinesis. A systematic review with meta-analysis (34 studies, 1,314 patients) finds a significant improvement in synkinesis scores on the Synkinesis Assessment Questionnaire (mean difference 11.6; p < 0,01), avec des effets indésirables généralement légers et transitoires ³. A further clinical trial confirms the fall in SAQ score after injection (from 46.22 to 37.55; p = 0.001), with gains in muscle function and quality of life ¹⁴. A crucial point stressed by the authors: treatment must be individualised, muscle by muscle, rather than standardised ³. The toxin selectively weakens the muscle that contracts wrongly; combined with rehabilitation, it restores a balance that strengthening alone could not achieve.

The mistakes not to make

This is the most important side of this section, because rehabilitation errors are not neutral: they can worsen the very sequela they claim to treat.

Mistake no. 1: forced, maximal exercises. The intuitive logic of "the harder I push, the more I recover" is counterproductive here. The synkinetic muscle is overactive, and a maximal contraction massively recruits the aberrantly reinnervated fibres: it therefore reinforces the abnormal co-contraction instead of undoing it. All the available evidence argues for the opposite: the Cochrane review finds benefit only for targeted and tailoredexercises, not for intensive analytical work ¹³, and it is gentle training, below the synkinetic threshold, that divides the risk of sequelae by four (RR 0.24). The aim is never brute strength, it is selectivity.

Mistake no. 2: global, uncontrolled movements. Asking the patient to "pull all their faces" or to move the whole hemiface as a block maintains exactly the pattern we want to dissolve. As synkinesis is a failure of separation between territories, rehabilitation must fractionate: isolate the eye, isolate the mouth, work each muscle independently, with visual feedback (mirror, biofeedback). Global movement reactivates the aberrant common command; segmental movement dismantles it. Mime therapy rests entirely on this fine, coordinated approach, never on massive effort ⁸.

Mistake no. 3: rehabilitating too hard, too early, on a face that is still paralysed. In the flaccid phase there is nothing to "build up": the nerve is not yet conducting. Intensively working a paralysed face mainly exposes the patient to overuse of the healthy contralateral muscles and to asymmetrical compensations. The priority of this phase lies elsewhere: eye protection. Any patient unable to close the eye must be given artificial tears or lubricating gel and night-time eyelid taping, failing which exposure keratitis may develop and progress to a neurotrophic corneal ulcer ¹⁶. Forgetting the eye in order to concentrate on the muscle is the mistake with the gravest consequences.

Mistake no. 4: overlooking the diagnostic red flag. Before any rehabilitation, let us recall that peripheral palsy affects the whole hemiface, forehead included. A patient who can still wrinkle the forehead and close the eye tightly on the affected side has a central lesion (stroke, tumour), not Bell's palsy: this is a referral emergency, not a physiotherapy case ¹⁵.

The electrical stimulation controversy: not settling it too quickly

Should a paralysed face be electrically stimulated? The question divides opinion, and honesty demands that we say so to the reader rather than assert. The historical fear is that electrical stimulation, in the acute phase, promotes synkinesis by maintaining disorganised activation during reinnervation. This caution retains a solid basis: the reference Cochrane review finds no benefit of electrical stimulation on incomplete recovery at six months compared with placebo ¹³, and the clinical practice guidelines do not recommend it: they conclude, moreover, that no firm recommendation can be made on physiotherapy for lack of sufficient evidence ⁹.

Recent data nevertheless qualify the picture without overturning it. One meta-analysis (13 trials, n = 1,191) reports that, added to usual care within 7 days, electrical stimulation would reduce the risk of incomplete recovery (RR 0.65; 95 % CI 0.48–0.88; moderate certainty), while judging its efficacy and safety in the acute phase "controversial" and insufficiently studied ¹⁶. Another systematic review (6 studies, 243 patients) suggests a possible benefit on overall facial function without a significant excess risk of synkinesis, 3 studies out of 5 showing no difference, but stresses the heterogeneity of the measures ¹⁷. A targeted synthesis likewise concludes that results are promising, including a reduction in synkinesis, when stimulation is applied early with suitable parameters (low frequency, daily), while underlining the limited data ¹⁸.

Where does that leave us? On uncertain ground, acknowledged as such. The literature is contradictory: the benefit signal exists, the synkinesis danger signal is neither confirmed nor definitively ruled out, and the quality of evidence remains heterogeneous. The reasonable position is therefore neither to ban nor to generalise electrical stimulation, but to handle it with caution : outside routine recommendation, never as a substitute for targeted motor retraining, which has the most direct evidence for preventing synkinesis ¹³.

The guiding line fits in one sentence: in facial palsy, we do not rehabilitate strength, we rehabilitate control. Gentleness, selectivity and early treatment prevent synkinesis; excessive force and global movement create it.

🗂️ What do concrete clinical cases teach us?

To make the reasoning tangible, here are two illustrative and deliberately fictional cases. They describe no real patient: they are teaching scenarios built solely from the evidence presented above. Their purpose is to show how the evidence translates into clinical decisions, from the first consultation to the management of sequelae: peripheral/central triage, eye protection, choice of rehabilitation and prevention of synkinesis.

Case 1 (fictional): An acute palsy: triage before treating

Let us imagine "Mrs L.", 42 years old, who presents with sudden distortion of the right hemiface that appeared on waking: the corner of the mouth droops, the smile is asymmetrical, the right eye waters and no longer closes completely. This picture is common: Bell's palsy is the most frequent acute mononeuropathy, with an incidence of the order of 20 to 30 cases per 100,000 people per year ¹².

Triage first. Before any treatment, the decisive question is that of the central red flag. Mrs L. is asked to wrinkle the forehead and to close the eyes tightly. On the affected side she can neither raise the eyebrow nor close the eyelid: the involvement concerns the whole hemiface, forehead included. This is the signature of a peripherallesion. If, conversely, she had retained the ability to wrinkle the forehead on the paralysed side, a central lesion (stroke, tumour) would have had to be suspected and an urgent opinion arranged, since the frontalis muscle receives bilateral cortical innervation, it is spared in supranuclear lesions ⁵¹.

The rest of the neurological examination is normal, with no other deficit: the diagnosis made is idiopathic peripheral facial palsy, whose dominant pathophysiological hypothesis is reactivation of herpes simplex virus type 1 around the geniculate ganglion ². Severity is graded with the House-Brackmann scale, the six-grade reference, from grade I (normal) to grade VI (no movement) ⁷ : Mrs L. is graded IV.

Protect the eye, immediately. As orbicularis oculi is no longer working, there is lagophthalmos with corneal exposure. Loss of eyelid closure exposes the patient to exposure keratitis, which may be complicated, if corneal sensation is lost, by a neurotrophic ulcer ⁶. The patient is therefore taught immediately: artificial tears during the day, lubricating gel, and taping the eyelid closed at night ¹. This simple measure takes precedence over everything else and is not up for discussion.

First-line medical treatment. We are less than 72 hours from symptom onset: the window for early oral corticosteroid therapy is open. The doctor prescribes prednisolone. The benefit is solid: in the landmark double-blind randomised trial, complete recovery at 3 months was 83.0 % with prednisolone versus 63.6 % without, and 94.4 % versus 81.6 % at 9 months ¹⁰. The Cochrane meta-analysis confirms this at the highest level of evidence: 17 % incomplete recovery with corticosteroids versus 28 % in the control group (RR 0.63; 95 % CI 0.50–0.80), that is a number needed to treat of 10 ¹¹.

And antivirals? The patient is told that an antiviral alone is not justified: aciclovir on its own provides no benefit ¹⁰, and corticosteroids alone are probably more effective than antivirals alone ¹². Adding an antiviral to the corticosteroid has little or no effect on incomplete recovery ¹². It might, however, reduce long-term sequelae (synkinesis, crocodile tears), compared with the corticosteroid alone ¹². The antiviral can therefore only be considered, if at all, as an add-on, never as monotherapy.

The physiotherapist's role at this stage. The spontaneous prognosis is broadly favourable: more than two thirds of typical Bell's palsies recover completely without treatment, up to 90 % in children and pregnant women ¹. There is therefore no question of over-treating. The aim of rehabilitation in the acute phase is twofold: to support motor recovery and, above all, to prevent synkinesis. The Cochrane review reports low-quality evidence that tailored facial exercises improve function, and an important signal: significantly fewer patients developed motor synkinesis after exercises ¹³. Targeted, gentle exercises guided in front of a mirror are therefore preferred to maximal or global analytical movements. The patient is honestly informed that this level of evidence remains modest and that the official guidelines do not settle the question of physiotherapy for lack of sufficiently good evidence ⁹.

Mrs L.'s recovery begins within three weeks, as in 85 % of patients ⁴, and she regains normal facial movement.

Case 2 (fictional): An established sequela: the time of synkinesis

Let us now imagine "Mr B.", 55 years old, seen eleven months after a left peripheral facial palsy. Movement has broadly returned, but he describes a persistent nuisance: when he smiles, his left eye partly closes; when he closes his eyes, the corner of his mouth contracts. These are synkinesis: involuntary co-contractions linked to aberrant reinnervation of the facial nerve ³. This scenario is by no means anecdotal: in Peitersen's large longitudinal series, associated movements persisted in 16 % of patients and contractures in 17 %, a reminder that close to one patient in three retains a sequela, synkinesis being the most frequent ⁴.

Re-grade with the right scale. The House-Brackmann scale, useful in the acute phase, is not very sensitive to synkinesis. Here the Sunnybrook scaleis preferred, being finer and quantifying separately resting symmetry, voluntary movement and, precisely, synkinesis ⁷⁸. It will serve as the reference measure for demonstrating progress.

Specialised facial rehabilitation as first line. In this chronic patient, the recommended strategy is clear: the first-line treatment of synkinesis is facial rehabilitation (neuromuscular rehabilitation, biofeedback), followed by botulinum toxin; surgery is reserved for failures only ¹⁵. Physiotherapy is the foundation, toxin a targeted add-on.

Mr B. is therefore offered mime therapy : massage, relaxation, inhibition of synkinesis, and exercises in coordination and emotional expression. It is the best-supported intervention in long-standing paresis: in Beurskens' randomised controlled trial (patients with paresis of more than 9 months), three months of mime therapy improved facial symmetry by 20.4 points on the Sunnybrook (95 % CI 10.4–30.4) and reduced severity by 0.6 grade on House-Brackmann, independently of age, sex and duration ⁸.

Botulinum toxin as a targeted add-on. As Mr B.'s synkinesis remains troublesome despite rehabilitation, injection of botulinum toxin type Ais discussed, the reference treatment for established synkinesis. A systematic review with meta-analysis (34 studies, 1,314 patients) shows a significant improvement in synkinesis scores measured with the Synkinesis Assessment Questionnaire (mean difference 11.6; p < 0,01), avec des effets indésirables généralement légers et transitoires; les auteurs insistent sur un individualised treatment rather than a standardised protocol ³. Another study reports a fall in the SAQ score from 46.22 to 37.55 after injection, with improvement in muscle function and quality of life ¹⁴. The toxin does not replace rehabilitation: it complements it.

What these two cases highlight

Comparing the two scenarios illuminates the logic of management: the same disease does not call for the same actions at different stages.

A word on electrical stimulation, to be handled with caution. In neither of these two cases did we offer it as routine, and that is a deliberate choice in the face of a contradictory literature. The reference Cochrane review finds no benefit of electrical stimulation on incomplete recovery at six months compared with placebo ¹³. More recent data are more nuanced: one meta-analysis suggests that, added early to usual care, it might reduce the risk of incomplete recovery ¹⁶, and other work finds no clear excess risk of synkinesis ¹⁷¹⁸. But the measures are heterogeneous, safety in the acute phase is still judged controversial, and the guidelines do not recommend it. As things stand, the matter cannot be settled : electrical stimulation is not a routine measure and the decision remains case by case, with the patient informed of the uncertainty.

Finally, these two stories are a reminder that the level of evidence for facial rehabilitation remains broadly modest : there is no high-quality evidence for (or against) physical therapies taken as a whole ¹³⁹. What emerges is a consistent direction (exercises that are tailored and guided rather than forced, neuromuscular rehabilitation targeting synkinesis in long-standing forms) and not a universal protocol. Stating this uncertainty to the patient is part of the care.

🧭 How to apply this in practice?

Faced with peripheral facial palsy, the first-line clinician's first task is not to rehabilitate, but to triage : rule out a central cause, make the eye safe, ensure early corticosteroid therapy, then calibrate rehabilitation targeted at the patients who genuinely need it. Let us recall the context: Bell's palsy is the most common acute mononeuropathy, with an incidence of 20 to 30 cases per 100,000 people per year ¹², and its spontaneous prognosis is good, more than two thirds of patients recover completely, up to 90 % in children and pregnant women ¹. Physiotherapy is therefore aimed above all at severe forms and sequelae, not at cases destined to resolve on their own.

A five-step management algorithm

The following sequence organises the decision, from first contact to follow-up. It applies to the physiotherapist, the general practitioner and the ENT specialist, each acting at their own level.

1. Confirm that the palsy is peripheral, the involvement affects the whole hemiface, forehead included : inability to raise the eyebrow and to close the eye on the paralysed side ¹. This is the diagnostic pivot.
2. Rule out the central red flag, if the forehead is spared (the patient still wrinkles the forehead and closes the eye on the affected side), the lesion is supranuclear: think stroke or tumour and refer urgently ⁵¹.
3. Protect the eye as soon as it no longer closes completely: this is a functional emergency, not an option (detailed below).
4. Check the corticosteroid window: oral prednisolone must be started within 72 hours of symptom onset; if the patient consults the physiotherapist first, refer back to the doctor without delay ⁹¹⁰.
5. Grade severity and plan follow-up: House-Brackmann from I (normal) to VI (no movement) for the overall grade ⁷, Sunnybrook in rehabilitation as it is more sensitive to synkinesis ⁸. It is this grade that determines the intensity and timetable of physiotherapy follow-up.

Faced with a facial palsy, you do not rehabilitate first: you protect the eye, secure the corticosteroid and rule out a stroke.

Eye protection: the absolute priority

Loss of orbicularis oculi leads to lagophthalmos (an eye that no longer closes) and corneal exposure. Left unmanaged, it exposes the patient to exposure keratitis, which may be complicated by loss of corneal sensation and a neurotrophic ulcer ⁶. Any patient unable to close the eye must be given, and above all understand, the following measures ¹⁹ :

• Artificial tears during the day, to be repeated generously to maintain the tear film.
• Lubricating gel or ointment , more viscous, particularly before periods of risk.
• Night-time eyelid taping with adhesive tape (eyelid closed, without pressing on the globe), because the protective blink disappears during sleep.

The physiotherapist has a watchdog role here: at each session, ask about ocular discomfort (gritty sensation, redness, pain, blurred vision). Any sign of keratitis calls for an ophthalmological opinion ¹⁶. This reflex must never come after the motor work.

The medical foundation: early corticosteroids, antivirals as an adjunct

The physiotherapist does not start the medical treatment, but must know its logic in order to reinforce adherence and spot delays in care. The mainstay is early oral corticosteroid therapy. In Sullivan's landmark double-blind randomised trial ¹⁰, complete recovery at 3 months reached 83.0 % with prednisolone versus 63.6 % without, and 94.4 % versus 81.6 % at 9 months. The Cochrane review confirms the benefit at the highest level of evidence: 17 % incomplete recovery with corticosteroids versus 28 % in the control group (RR 0.63; 95 % CI 0.50-0.80), that is only one patient to treat out of ten to prevent a sequela ¹¹. Notable for the physiotherapist: corticosteroids also reduce synkinesis at long-term follow-up ¹¹: they therefore partly prevent the very sequela that rehabilitation will subsequently seek to limit.

The antivirals alone are ineffective and must not be prescribed as monotherapy ⁹. Aciclovir alone provided no benefit in Sullivan's trial ¹⁰, and corticosteroids alone probably do better than antivirals alone ¹². Added to corticosteroids, antivirals barely improve recovery (RR 0.81; 95 % CI 0.38-1.74) but reduce long-term sequelae, synkinesis and "crocodile tears", versus corticosteroids alone ¹². The antiviral is therefore to be considered as an add-on, never as isolated first-line treatment.

Rehabilitation: targeted, tailored, never forced

Here honesty is required: the level of evidence for facial physiotherapy is modest. The Cochrane review finds no high-quality evidence of benefit or harm from physical therapies taken as a whole, and the AAO-HNS guideline concludes that no formal recommendation can be made for lack of sufficient evidence ⁹¹³. This does not mean "do nothing": it points towards individualised and tailored exercises rather than maximal or global analytical exercises.

What the evidence, even of low quality, supports:

• Some tailored facial exercises improve facial function, particularly in moderate palsies and chronic cases, and reduce sequelae in acute cases ¹³.
• Targeted exercise prevents synkinesis after an acute palsy: significantly fewer patients developed it ¹³. This is a strong argument in the early phase.
• For long-standing paresis (> 9 months), Beurskens' mime therapy (massage, relaxation, inhibition of synkinesis, coordination and expression exercises), improves facial symmetry by 20.4 points on the Sunnybrook (95 % CI 10.4-30.4) and reduces severity by 0.6 grade on House-Brackmann ⁸.

Established sequelae. Synkinesis (involuntary co-contractions, for example eye closure on smiling) affects about 16 % of patients ⁴ and results from aberrant reinnervation. The recommended hierarchy is clear: facial rehabilitation (biofeedback) as first line, then botulinum toxin A, with surgery reserved for failures ¹⁵. Toxin A significantly improves synkinesis ³¹⁴, always in an individualised way, with no standardised protocol.

The case of electrical stimulation. The literature remains contradictory and we shall not settle it. The historical fear is that it promotes synkinesis in the acute phase; some syntheses show no benefit on incomplete recovery at six months ¹³. Others, more recent, are nuanced: one meta-analysis suggests that, added early to usual care, it would reduce incomplete recovery ¹⁶, while one review finds no clear excess risk of synkinesis ¹⁷¹⁸. Honest conclusion: heterogeneous evidence, no routine recommendation.

Key messages to convey to the patient

• "Most cases of Bell's palsy resolve", reassure without trivialising: complete recovery in the majority, often within the first weeks ¹⁴.
• "The most urgent thing is the eye", explain tears, gel and night-time taping, and the corneal warning signs ¹⁶.
• "The corticosteroid is decided within 3 days", insist on starting early and on adherence ¹⁰.
• "We work gently, not forcefully", maximal grimacing or systematic electrical stimulation are not the answer; the aim is coordination and the prevention of synkinesis ¹³.
• "Recovery is counted in months", most cases resolve within 4 to 6 months, with almost complete resolution within a year ² ; giving this timetable prevents anxiety and doctor-shopping.

Common mistakes to avoid

• Confusing peripheral and central. A spared forehead is not Bell's palsy: it is a signal of possible stroke that calls for emergency care ⁵¹.
• Neglecting the eye in favour of motor work: exposure keratitis is the real preventable complication ⁶.
• Delaying the corticosteroid or settling for an antiviral alone, which is ineffective as monotherapy ⁹¹².
• Imposing maximal, global analytical exercises instead of tailored, graded work: a risk of maintaining co-contraction patterns ¹³.
• Presenting electrical stimulation as an established treatment when the evidence is contradictory ¹⁶¹³.
• Over-treating a mild form destined to resolve spontaneously, instead of reserving effort for severe forms and sequelae ¹.