Condition guide · Neurological
Bell's Palsy
Sudden weakness of one side of the face. Most people recover, and the single most important treatment is a drug started early — not an exercise. Facial rehabilitation has a real but narrower role, and it matters most for what happens after the first few weeks.
In plain words. The facial nerve controls the muscles of one side of the face — the eyebrow, the eyelid, the cheek and the corner of the mouth. In Bell's palsy that nerve stops working properly, usually over hours, for reasons that are not fully understood. The face droops on one side, the eye may not close, and speech and eating become awkward.
It is frightening, and it is usually not a stroke. But the two can look similar enough that sudden facial weakness should always be assessed urgently rather than waited out.
The most useful thing on this page is not physiotherapy. The 2023 Japanese clinical practice guideline, which examined nine treatment questions using GRADE methodology, strongly recommends systemic standard-dose corticosteroids for the management of Bell's palsy. Every other treatment it considered — including physical therapy — is only weakly recommended, because the evidence is insufficient. [3] If your face has just become weak, the priority is medical assessment today, not an exercise programme.
Key points
- Corticosteroids started early are the established treatment. Strongly recommended; everything else is weakly recommended. [3]
- Antivirals add little on top of steroids for recovery — but probably do reduce the long-term sequelae. [4]
- Physical therapy probably reduces non-recovery. Pooling four studies and 418 participants, the risk ratio was 0.51 (95% CI 0.31 to 0.83) — on low-quality evidence. [1]
- The benefit is clearest early, and there is also evidence of value in chronic cases. [2]
- A structured programme of motor imagery, manipulation and mirror therapy beat basic counselling in a randomised trial of severe acute palsy followed for a year. [6]
- Eye protection is not optional. If the eye does not close, the cornea is at risk, and that is the one complication that can cause permanent harm.
- Synkinesis — unwanted linked movement — is the problem rehabilitation is most concerned with, and the evidence that therapy prevents it is uncertain. [1]
- Electrotherapy has not been shown to work here. [8]
What is actually happening
The facial nerve leaves the brainstem and travels through a narrow bony canal in the skull before fanning out across the face. In Bell's palsy the nerve becomes inflamed and swollen inside that canal. Because the canal cannot expand, the swelling compresses the nerve against bone, and conduction fails.
That anatomy explains two things. It explains why the weakness can be complete rather than partial, and it explains why treatment aimed at inflammation — corticosteroids — is the intervention with the strongest evidence. [3]
Recovery happens as the inflammation settles and the nerve fibres recover or regrow. Where fibres regrow, they do not always find their original destination. A fibre that once supplied the corner of the mouth may end up supplying the eyelid. The result is synkinesis: the eye narrows when you smile, or the mouth twitches when you blink. This is the specific problem that facial rehabilitation is aimed at, and it is why the treatment is about control and coordination rather than strength.
What it feels like
- Weakness or complete paralysis of one side of the face, usually developing over hours.
- The eyebrow does not lift and the eyelid does not close fully on that side.
- Drooping at the corner of the mouth; drinks escape from that side.
- Speech that is difficult to form, particularly sounds made with the lips.
- Pain behind or below the ear, often before the weakness appears.
- Altered taste, and sounds seeming uncomfortably loud on that side.
- A dry or gritty eye.
The forehead is the key observation. In Bell's palsy the forehead is affected along with the rest of that side of the face. In a stroke affecting the face, the forehead is usually spared. That distinction belongs to a doctor making it in person, not to a reader making it in a mirror.
How it is diagnosed
Clinically, and largely by exclusion. Bell's palsy is the diagnosis when one side of the face is weak, the whole side including the forehead is involved, and no other cause is found. Sudden facial weakness needs same-day medical assessment: stroke, herpes zoster affecting the ear, middle ear disease, Lyme disease and tumours can all present this way, and the treatments differ.
Severity is graded, most often with the House-Brackmann or the Sunnybrook facial grading system. This is not bureaucracy. The guideline is explicit that the absolute risk reduction of each treatment differed according to disease severity, and that physicians and patients should decide on treatment based on it. [3] The grade also tells you which evidence on this page applies to you: the randomised trial of structured facial therapy, for instance, recruited only people with moderately severe to total palsy. [6]
How physiotherapy and rehabilitation help
Three separate jobs, in rough order of importance: protect the eye, support recovery in the early weeks, and manage synkinesis if it develops.
What the meta-analysis found
A 2024 systematic review and meta-analysis pooled randomised trials comparing physical therapy against placebo or no treatment in peripheral facial palsy. Seven trials met criteria. On the primary outcome, non-recovery at the end of follow-up, data from four studies and 418 participants showed that physical therapy might reduce non-recovery (risk ratio 0.51, 95% CI 0.31 to 0.83). Pooling the Sunnybrook composite score from three studies and 166 participants, physical therapy might increase the score (mean difference 12.1, 95% CI 3.11 to 21.0). [1]
Both of those findings were graded low quality. The third outcome is the one worth dwelling on: for reduction of sequelae — synkinesis or hemifacial spasm — data from two articles and 179 participants gave a risk ratio of 0.64 with a confidence interval from 0.07 to 5.95, which the authors describe as very uncertain. [1] A confidence interval that wide contains both large benefit and large harm; it means the question has not been answered.
An earlier and broader review reached a compatible conclusion by a different route. It found seven new randomised trials, nine observational studies and three quasi-experimental or pilot studies, covering 854 participants. Differences in study design prevented pooling. All four of the high-quality trials reported positive impacts, as did the three observational studies rated high or moderate quality. The authors concluded that the newer research strengthens previous conclusions about the benefits of facial exercise therapy early in recovery, and adds to the evidence of value in chronic cases. [2]
The one good trial of a structured programme
Most facial therapy is described vaguely. One randomised controlled trial tested a defined protocol. Forty patients with moderately severe to total palsy, all of whom had received standard medication, were recruited within 14 days of onset and randomised to either the "Mirror Effect Plus Protocol" — motor imagery, manipulations and facial mirror therapy — or to basic counselling, then followed for a year. [6]
The protocol group did better on facial symmetry (House-Brackmann 2.0 score, mean 5.1 with standard deviation 1.44, against 7.40 and 3.15 in the control group), on synkinesis measures (p = 0.008), and on quality of life (98.36% against 83.17%, p = 0.002). There was no group difference in how intelligible listeners found their speech. [6]
Forty people is a small trial, and it is a single trial. But it tested a specified programme, in the population most likely to have persistent problems, with a year of follow-up — which is more than most of this literature offers.
What a programme involves
| Component | What it is for | Evidence |
|---|---|---|
| Eye protection lubricant, taping or a shield at night |
Preventing corneal damage while the eyelid does not close. This is the one complication that can cause permanent injury. | Standard care Not an efficacy question — an unprotected cornea is a known hazard while the blink is absent |
| Facial exercise therapy graded, mirror-guided, specific rather than forceful |
Supporting recovery of controlled, symmetrical movement. | Probably helpful Non-recovery risk ratio 0.51 (0.31 to 0.83) and Sunnybrook composite mean difference 12.1 (3.11 to 21.0), both low quality [1]; benefits early in recovery and in chronic cases, but no pooling possible [2] |
| Motor imagery and mirror therapy as a defined protocol | Retraining the movement pattern in severe palsy, including after the acute phase. | Supported by one trial Better facial symmetry, synkinesis (p = 0.008) and quality of life (p = 0.002) than basic counselling at one year, in 40 patients with moderately severe to total palsy [6] |
| Botulinum toxin for established synkinesis a medical treatment, delivered alongside rehabilitation |
Reducing the unwanted linked movement that can follow recovery. | Helpful, but not standardisable Synkinesis Assessment Questionnaire mean difference 11.599 across three studies and 106 patients, p < 0.01; the reviewers could not establish a relationship with dose or injection site and recommend patient-tailored rather than standardised treatment [7] |
| Forceful exercise and grimacing the intuitive approach |
Offered on the assumption that the muscles are weak and need strengthening. | Not supported by the trials cited here The studied programmes are graded and coordination-focused rather than maximal-effort [2][6]; this page makes no claim that forceful exercise has been tested and failed, only that it is not what was tested |
| Electrotherapy electrical stimulation of the facial muscles |
Widely offered on the theory that stimulation maintains the muscle while the nerve recovers. | Not demonstrated Seven studies, 131 cases and 113 controls; electrotherapy was never given alone, so its own effect was not evaluated, and the reviewers could not prove efficacy [8] |
An amber badge is a statement about the evidence cited on this page, not a verdict on the treatment.
Severity is graded at assessment because it determines what applies to you: the absolute risk reduction of each treatment differed according to disease severity, [3] and the trial of structured facial therapy recruited only moderately severe to total palsy. [6]
What a course of treatment looks like
In the first days, the priorities are medical: assessment, corticosteroids where indicated, [3] and protecting the eye. Rehabilitation at this stage is mostly education — how to look after the eye, how to eat and drink, what to expect, and what not to do.
As movement begins to return, therapy becomes specific. The aim is small, accurate, symmetrical movements practised with a mirror, rather than maximal effort. This is the opposite of the intuition most people bring, and it is the part that benefits most from being taught rather than looked up.
If synkinesis appears, the work changes again: learning to move one region without the other, often combined with botulinum toxin injections given by a specialist. [7] This phase can run for months, and the evidence base for preventing synkinesis in the first place is genuinely uncertain. [1]
Reviews are spaced according to the trajectory. A palsy that is recovering quickly needs little; one that has not begun to recover by three weeks needs medical review, not more exercises.
Where the problem is dizziness or hearing rather than facial movement, see vestibular neuritis. Where facial weakness comes with weakness elsewhere in the body, this is not Bell's palsy — see stroke rehabilitation and seek urgent assessment.
What the evidence supports — and what it does not
Supported
- Systemic standard-dose corticosteroids — the only strongly recommended treatment in the guideline. [3]
- Adding antivirals to corticosteroids to reduce long-term sequelae. Risk ratio 0.56 (95% CI 0.36 to 0.87), moderate certainty. [4]
- Physical therapy reducing non-recovery, on low-quality evidence. [1]
- Facial exercise therapy early in recovery, and in chronic cases. [2]
- A structured motor imagery and mirror protocol in severe acute palsy. [6]
Not supported
- Antivirals alone. No clear benefit over placebo. [4]
- Antivirals plus steroids improving recovery over steroids alone. Little or no effect, and imprecise. [4]
- Routine high-dose corticosteroids. The evidence that they reduce non-recovery is very low quality and the right dose is unknown. [5]
- Confidence that therapy prevents synkinesis. Risk ratio 0.64 with a confidence interval from 0.07 to 5.95. [1]
- Electrotherapy. Efficacy could not be demonstrated. [8]
How certain is this?
Evidence grade: Low.
The grade applies to the physiotherapy, which is what this page is about. The drug evidence is stronger and is cited here mainly so that readers do not mistake the order of priorities.
The meta-analysis supporting physical therapy is explicit about its own weakness: the included studies had high risk of bias, imprecision or inconsistency, so the certainty of evidence was low or very low, and the authors call for further well-designed randomised trials. [1] The larger review could not perform any quantitative pooling at all because of differences in study design, and it rated all six of the new review articles it found as critically low quality. [2]
The guideline is a strong document, but note what it actually says: standard-dose corticosteroids are strongly recommended, and other treatments — physical therapy among them — are weakly recommended due to insufficient evidence. [3] That is a fair summary of everything above.
Two of the drug findings are worth flagging as weaker than they are often presented. The evidence that high-dose corticosteroids reduce non-recovery at six months (odds ratio 0.42, 95% CI 0.22 to 0.80) came from eight non-randomised comparisons and was graded very low quality, with initial doses ranging from 120 mg to 200 mg daily and no established optimum. [5] And the Cochrane review's central comparison — antivirals plus corticosteroids against corticosteroids alone — found little or no effect on incomplete recovery, with a confidence interval running from 0.38 to 1.74. [4]
What to expect
Most people recover. The measure of that in the research is "non-recovery" at the end of follow-up, and it is the minority outcome even in the control groups of these trials. [1]
Recovery usually begins within two to three weeks and continues over months. A palsy that has not started to improve by three weeks should be reviewed medically, because it raises the question of whether the diagnosis is right.
Where recovery is incomplete, the residual problem is usually synkinesis rather than weakness — and that is a different problem needing different treatment, including possibly botulinum toxin. [7]
Expect the psychological effect to be larger than the physical one. The face is how people are recognised, and the trial that measured quality of life found a large difference between groups on that outcome. [6] It is a reasonable thing to raise with your clinician.
Common questions
Is this a stroke?
Usually not — but that judgement has to be made in person, today. Sudden facial weakness needs urgent medical assessment. The classic distinction is that Bell's palsy affects the forehead on the weak side while a stroke usually spares it, but weakness or numbness elsewhere in the body, difficulty speaking or swallowing, or sudden severe headache all point away from Bell's palsy and towards an emergency.
What is the single most important treatment?
Corticosteroids, started early. The 2023 clinical practice guideline strongly recommends systemic standard-dose corticosteroids and only weakly recommends everything else, including physical therapy, because the evidence is insufficient. [3] Physiotherapy is worth having; it is not the first priority in the first days.
Should I take antiviral tablets too?
That is a decision for the doctor prescribing, and the evidence is mixed. Adding antivirals to corticosteroids may have little or no effect on incomplete recovery (risk ratio 0.81, 95% CI 0.38 to 1.74, low certainty). But the same review found that the combination probably reduces the long-term sequelae — synkinesis and crocodile tears — compared with corticosteroids alone (0.56, 0.36 to 0.87, moderate certainty). Antivirals alone showed no clear benefit over placebo. [4]
Should I be doing facial exercises?
Probably yes, and the way they are done matters more than how many. Pooled data suggest physical therapy reduces non-recovery (risk ratio 0.51, 95% CI 0.31 to 0.83) and improves the Sunnybrook composite score, both on low-quality evidence. [1] The programmes that were studied were graded and coordination-focused, done in front of a mirror — not maximal grimacing. Get them taught rather than copied from a video.
My eye will not close. What do I do?
Treat it as the priority. Without a blink the cornea dries and can be damaged permanently, and that is the one lasting harm most likely to come from Bell's palsy. Lubricating drops during the day, ointment and taping or a shield at night, and a medical review of the eye are all standard. This is not an area to improvise in.
My eye twitches when I smile. Is that normal?
It is common, it has a name — synkinesis — and it is treatable. It happens because regrowing nerve fibres reach the wrong muscles. Specific rehabilitation aimed at separating the movements is the mainstay, and botulinum toxin is frequently used alongside it: pooling three studies and 106 patients gave a mean improvement of 11.599 on the Synkinesis Assessment Questionnaire. [7] Note that the reviewers could not identify a standard dose or injection pattern and recommend individually tailored treatment. [7]
Will electrical stimulation speed up recovery?
There is no good evidence that it does. A systematic review of seven studies, with 131 cases and 113 controls, found that electrotherapy was always given combined with other treatments, so its own effect was never evaluated, and concluded that efficacy could not be proven. [8] That is not the same as showing it fails — but it does mean nobody can honestly promise you it works.
It has been six months and I am not fully better. Is it too late?
No. The extended systematic review specifically noted that the newer studies add to the evidence of value in chronic cases, not only early ones. [2] The goals change — symmetry, control and synkinesis rather than recovery of gross movement — but there is still something to work on.
References
- Nakano H, Fujiwara T, Tsujimoto Y, et al. Physical therapy for peripheral facial palsy: A systematic review and meta-analysis. Auris, Nasus, Larynx. 2024 Feb;51(1):154–160. doi:10.1016/j.anl.2023.04.007 PMID 37149416 Systematic review and meta-analysis
- Khan AJ, Szczepura A, Palmer S, et al. Physical therapy for facial nerve paralysis (Bell's palsy): An updated and extended systematic review of the evidence for facial exercise therapy. Clinical Rehabilitation. 2022 Nov;36(11):1424–1449. doi:10.1177/02692155221110727 PMID 35787015 Systematic review
- Fujiwara T, Hato N, Kasahara T, et al. Summary of Japanese clinical practice guidelines for Bell's palsy (idiopathic facial palsy) - 2023 update edited by the Japan Society of Facial Nerve Research. Auris, Nasus, Larynx. 2024 Oct;51(5):840–845. doi:10.1016/j.anl.2024.07.003 PMID 39079445 Clinical practice guideline
- Gagyor I, Madhok VB, Daly F, et al. Antiviral treatment for Bell's palsy (idiopathic facial paralysis). Cochrane Database of Systematic Reviews. 2019 Sep 5;9(9):CD001869. doi:10.1002/14651858.CD001869.pub9 PMID 31486071 Cochrane systematic review
- Fujiwara T, Namekawa M, Kuriyama A, et al. High-dose Corticosteroids for Adult Bell's Palsy: Systematic Review and Meta-analysis. Otology & Neurotology. 2019 Sep;40(8):1101–1108. doi:10.1097/MAO.0000000000002317 PMID 31290805 Systematic review and meta-analysis
- Martineau S, Rahal A, Piette E, et al. The "Mirror Effect Plus Protocol" for acute Bell's palsy: A randomized controlled trial with 1-year follow-up. Clinical Rehabilitation. 2022 Oct;36(10):1292–1304. doi:10.1177/02692155221107090 PMID 35722671 Randomised controlled trial
- de Jongh FW, Schaeffers AWMA, Kooreman ZE, et al. Botulinum toxin A treatment in facial palsy synkinesis: a systematic review and meta-analysis. European Archives of Oto-Rhino-Laryngology. 2023 Apr;280(4):1581–1592. doi:10.1007/s00405-022-07796-8 PMID 36544062 Systematic review and meta-analysis
- Burelo-Peregrino EG, Salas-Magaña M, Arias-Vázquez PI, et al. Efficacy of electrotherapy in Bell's palsy treatment: A systematic review. Journal of Back and Musculoskeletal Rehabilitation. 2020;33(5):865–874. doi:10.3233/BMR-171031 PMID 32144972 Systematic review
About this guide
If you need assessment
This page explains. It does not diagnose.
Sudden facial weakness needs same-day medical assessment to rule out stroke and other causes, and the treatment with the strongest evidence is a prescription started early. An eye that does not close needs protecting immediately. Physiotherapist India publishes information and takes no bookings. Ask a qualified physiotherapist to examine your own case before acting on anything written here.