Guillain-Barré syndrome

The community physiotherapist rarely sees Guillain-Barré syndrome in its acute phase: that is played out in hospital, sometimes in intensive care. What they see is what comes afterwards, and that phase is poorly covered by the literature and by training alike. A patient who comes home after several weeks in hospital, having been told that the disease is monophasic and that recovery would be good, and who discovers that they are exhausted, in pain, and that their former life remains out of reach. This article deals with that phase: what recovery actually does, why residual fatigue affects six patients in ten with no relation to initial severity, what rehabilitation has shown, and the precise moment when the physiotherapist must stop adjusting the programme and write to the neurologist. On the chronic form that is defined against this one, our article on chronic inflammatory demyelinating polyradiculoneuropathy (CIDP) deals with the other side of the same differential.

• Updated 16 August 2026
• Level Clinical summary
• Sources 37 verified references
• ICD-11 8C01.0

What is Guillain-Barré syndrome, and what is its curve?

A brief immune attack on the peripheral nerve, most often triggered by an infection. What defines it is not a sign, it is the shape of a curve: a rapid worsening, a floor, then a climb back.

A nerve attacked after an infection

Guillain-Barré syndrome is the commonest cause of acute flaccid paralysis in the world 3. Most patients report an illness preceding the onset of the deficit, most often an upper respiratory tract infection or diarrhoea. Several micro-organisms have been associated with it, foremost among them Campylobacter jejuni, Zika virus, and in 2020 SARS-CoV-2. In the forms linked to C. jejuni, the mechanism rests on molecular mimicry between structural components of the peripheral nerve and those of the micro-organism, which triggers cross-reacting autoantibodies 3.

The term covers several distinct clinical and electrophysiological variants, including the classic demyelinating form, the axonal forms and Miller Fisher syndrome 4. About 100,000 people develop the disease each year worldwide, and the severe generalised form with respiratory failure concerns 20 to 30 % of cases 4. Ten to thirty per cent of affected adults require mechanical ventilation during the acute phase 11.

What the Brighton series measured, and which serves as the reference for everything else

A series of 494 adult patients, assembled to validate the Brighton Collaboration diagnostic criteria, provides the reference figures on presentation and chronology 6. Median age 53, a slight male predominance (56 %). All developed a bilateral limb deficit, generally in all four; it remained limited to the legs in 6 % and to the arms in 1 %. Reflexes were initially diminished in the paretic limbs in 91 % of patients, and in all of them during follow-up; ten patients (2 %) nevertheless kept normal reflexes in paretic arms.

Two details of this series deserve to be known, because they upset common expectations. First, CSF protein was raised in only 64 % of patients, and that proportion depends heavily on when the lumbar puncture is done: 49 % on the first day of the deficit, 88 % after two weeks. Second, electrophysiology was compatible with a neuropathy in 99 % of patients, but only 59 % met the criteria for a precise subtype 6. A report that does not conclude to a precise form is therefore not a doubtful report: it is the case in four patients in ten.

Who is affected

The reference meta-analysis on population incidence brought together sixteen North American and European studies, that is 1,643 cases and 152.7 million person-years of follow-up. It establishes two regularities: incidence increases by about 20 % for every ten years of age, and the risk is higher in men than in women 5. A patient of 70 therefore has a markedly higher risk than a patient of 30, which the common image of a disease of young people does not reflect.

When should you doubt the Guillain-Barré label?

A patient who has come through Guillain-Barré and is referred for rehabilitation carries a diagnosis made in hospital. In about one case in twenty, that diagnosis will change, and it is the professional who sees the patient every week who will give the signal.

Two forks in the road, separated by the calendar

Guillain-Barré is monophasic in 95 % of patients. But “monophasic” does not mean “without relapse”: among those patients with a monophasic course, 10 % had a treatment-related fluctuation 6. A treatment-related fluctuation is defined as clinical deterioration occurring within two months of symptom onset, after stabilisation or improvement obtained under treatment 25. An Argentinian series of 124 cases finds seven, that is 5.6 % of all patients 25. The two figures do not count the same thing, one referred to monophasic patients and the other to the whole; they agree that this situation is not rare.

The second fork is elsewhere in time, and it changes the treatment for years. In the Brighton series, 23 patients out of 494, that is 5 %, deteriorated beyond the eighth week 6. The European guideline sets out the rule explicitly: the diagnosis should be reconsidered as acute-onset chronic inflammatory demyelinating polyradiculoneuropathy if progression continues beyond eight weeks from onset, which happens in about 5 % of patients initially diagnosed with Guillain-Barré 1.

The issue is not nosological. Guillain-Barré is treated with a short course; CIDP requires long-term immunosuppression. Getting the side wrong means letting a patient deteriorate for want of maintenance treatment. The literature devoted to this distinction, applied to recurrent forms, in fact concludes that it remains difficult, that the electrophysiological, ultrasound and immunological criteria proposed have not settled it, and that the very concept of recurrent Guillain-Barré is debated 24.

A patient labelled Guillain-Barré who is still deteriorating in the third month does not need a programme adjustment. They need a letter.

The last row is deliberately identical in all three columns. The physiotherapist does not have to decide between these three courses, and does not have the means to. What they hold, and nobody else does, is the date: the day the patient stopped progressing, the day they started declining again, and by how much. Our article on CIDP deals with the other side of this boundary, with the proximal and distal deficit that marks it and the European diagnostic criteria.

What do the treatments do, and what is left afterwards?

The physiotherapist does not administer these treatments, but does receive the patients who have had them. Knowing what they have achieved, and what they do not achieve, changes how objectives are set.

What the European guideline retains

The joint guideline of the European Academy of Neurology and the Peripheral Nerve Society, published in 2023 following GRADE methodology, sets out clear recommendations. It recommends intravenous immunoglobulin at 0.4 g/kg for 5 days in patients unable to walk unaided, within two weeks of the onset of the deficit, or a course of plasma exchange of 12 to 15 litres in four to five exchanges over one to two weeks, within four weeks of onset. It recommends against a second course of immunoglobulin in patients with a poor prognosis, against oral corticosteroids, and weakly against intravenous corticosteroids. It does not recommend following plasma exchange with immunoglobulin. For pain, it weakly recommends gabapentinoids, tricyclic antidepressants or carbamazepine. And for fatigue, it recommends no specific treatment 1.

That last sentence deserves rereading. The symptom most frequently reported at a distance from the illness is the one for which no treatment is recommended. That is precisely where the physiotherapist's work lies, and it is the subject of the next chapter.

The two treatments are equivalent, and one is completed more often than the other

The Cochrane review devoted to immunoglobulin retained twelve trials. Seven of them, covering 623 severely affected participants, compared immunoglobulin with plasma exchange; across the five trials where the outcome was available, in 536 participants, the mean difference in improvement on a seven-grade disability scale after four weeks was 0.02 grade in favour of immunoglobulin, with a 95 % confidence interval running from -0.20 to 0.25, that is, no difference. Adverse effects were no more frequent with one or the other, but immunoglobulin is far more often completed. Finally, giving it after plasma exchange brings no significant additional benefit 8.

The Cochrane review devoted to plasma exchange, which rests on six trials and 649 participants, gives the figures most telling for a rehabilitation clinician, because they concern walking. In a trial of 220 severely affected patients, the median time to walking again with aid was 30 days with plasma exchange against 44 days without. At four weeks, the proportion of patients who had recovered assisted walking was higher (relative risk 1.60; 95 % CI 1.19 to 2.15), and improvement of at least one disability grade was 1.64 times more frequent (95 % CI 1.37 to 1.96). The need for artificial ventilation was reduced (relative risk 0.53; 95 % CI 0.39 to 0.74). At one year, full recovery of muscle strength was more likely (relative risk 1.24; 95 % CI 1.07 to 1.45) and severe motor sequelae less frequent (relative risk 0.65; 95 % CI 0.44 to 0.96). Against that, relapses were more frequent by the end of follow-up (relative risk 2.89; 95 % CI 1.05 to 7.93) 9.

The prognostic scores, and what they announce

Two families of tools are used in hospital, and it is useful to know what they predict when reading a report. The modified respiratory insufficiency risk score, built on the 1,133 eligible patients of the first 1,500 in the international outcome study, predicts the need for mechanical ventilation from three factors: a short interval between the onset of the deficit and admission, the presence of bulbar involvement, and weakness of the neck and hip flexors. Its area under the curve is 0.84 (95 % CI 0.80 to 0.88) 10. Neck flexor weakness, testable in three seconds at the edge of a bed, therefore features in a validated prognostic score.

The 2023 neuroprognostication guidelines retain bulbar weakness and the respiratory risk score as moderately reliable predictors of the need for mechanical ventilation, and the outcome scores as moderately reliable predictors of independent walking at three months and beyond. They insist on a methodological point that also holds in the clinic: the complete clinical picture must be considered, rather than an isolated variable 11.

What does the community physiotherapist see when the patient comes home?

A patient who has been told that the disease is monophasic and that recovery would be good, and who discovers the gap between that sentence and their daily life. That gap is documented, and naming it is already part of the care.

A real recovery, and an incomplete one in many

The systematic review devoted to long-term outcomes, which retained studies reporting results at six months and beyond, concludes that most patients show substantial neurological recovery in the first year, but that a clinically significant proportion continue to experience persistent symptoms beyond one to three years. Fatigue and neuropathic pain are among the most frequently reported sequelae and contribute most to impaired quality of life, despite apparent motor recovery. Residual functional limitations, participation restrictions and psychosocial impact are also described. The factors associated with a poorer outcome are older age, severe disease at presentation, the use of mechanical ventilation, dysautonomia, and an axonal electrophysiological profile 23.

The formulation those authors use deserves to be taken up as it stands with the patient: recovery is often incomplete despite improvement in neurological function. That is not bad news to hide, it is a frame that stops the patient concluding that they are failing.

A Japanese review also points out that the prognosis of Guillain-Barré is not as good as one might expect: 15 to 20 % of patients do not walk independently six months after onset 33. One patient in six or seven, then, for a disease described as having a good prognosis.

What the patient does not say spontaneously

A scoping review devoted to patients' experience, which mapped the biological, psychological, social and spiritual repercussions in adults, stresses that the clinical aspects have been widely studied while those dimensions remain poorly documented 34. A British qualitative study conducted with people who had returned to work after Guillain-Barré brings out three recurring themes: the perceived value of work, the loss and then the reconquest of a familiar identity, and the factors that ease or hinder the return 31. The authors note that residual problems affect a substantial minority and complicate that return.

In practice, three questions often open a conversation that the analytical assessment does not trigger: how long the patient lasts in a day before having to stop; what they have given up doing without telling anyone; and what an acceptable week would look like for them, rather than a normal one.

Why is residual fatigue so frequent and so little treated?

It is the symptom patients mention first, the one that decides the return to work, and the one for which no treatment is recommended. It is also the one the examination explains worst.

The figure, and what it does not predict

A series of 100 patients measured the occurrence of severe fatigue after Guillain-Barré, defined by a mean score of at least 5.0 on the Fatigue Severity Scale. It was present in 60 % of patients, more frequently in women and in patients over 50 (p < 0.01). And above all, no significant relationship was found between the intensity of the fatigue and the level of disability at the nadir, the events or infections preceding the illness, the clinical variables, or the time elapsed since the episode 13.

That absence of a relationship is the most important clinical fact in this chapter. A patient who had a moderate form, never ventilated, discharged in three weeks, can be as profoundly fatigued as a patient who spent two months in intensive care. The intuitive reasoning, which consists of calibrating your expectations on the initial severity, is at fault here.

The nerve has repaired, and the fatigue has stayed

A standardised nerve conduction study was conducted in sixteen fatigued patients, on average 6.5 years after diagnosis, thirteen of them relatively well recovered from Guillain-Barré. Unlike what was observed in CIDP, most conduction values were remarkably restored and lay within normal limits in the patients who had had Guillain-Barré. No correlation was found between the electrophysiological parameters and the fatigue 16. Residual subclinical peripheral dysfunction therefore does not explain this fatigue.

A further study broke down physiological fatigue during a sustained two-minute maximal voluntary contraction, in ten neurologically well recovered patients and twelve controls matched for age and sex, distinguishing the peripheral fatigue component from the central activation failure 15. This work draws a symptom that is neither purely muscular nor purely psychological, and that the analytical examination does not capture.

What has been tried other than exercise

Amantadine was the subject of a randomised, double-blind, placebo-controlled crossover trial in patients whose fatigue was severe, with a primary outcome of improvement of at least one point on the Fatigue Severity Scale 17. The Cochrane review devoted to interventions for fatigue in peripheral neuropathy, which covers drug treatments as well as physical, psychological and behavioural interventions, was conducted to assess precisely that question 19. At the end of that road, the 2023 European guideline recommends no specific treatment for fatigue 1. That therapeutic void is exactly the space in which the physiotherapist works.

What rehabilitation, and at what level of evidence?

The literature exists, it is consistent, and it rests on tiny samples. It has to be presented as such: it guides practice, it does not demonstrate it.

The first signal, in 2004

Twenty patients with severe fatigue, sixteen of them relatively well recovered from Guillain-Barré and four with stable CIDP, followed twelve weeks of bicycle training. The training appeared well tolerated, self-reported fatigue scores fell by 20 % (p = 0.001), and physical fitness, functional outcome and quality of life improved 14. A later analysis of the same patients sought to untangle the relationships between fitness, fatigue, objectively measured mobility, perceived physical functioning and perceived mental functioning 18. This is a case series, with no control group.

The only randomised trial of the chronic phase

Sixteen adults with stable residual disability at least six months after the onset of Guillain-Barré were randomised between two twelve-week programmes. The experimental group received 60-minute sessions supervised by a physiotherapist, combining strengthening, endurance, breathing exercises, gait retraining and pain management, two to three times a week. The control group received a home programme of 30 minutes of maintenance exercises and self-management education, at the same frequency.

At six months, the median between-group difference was 5 points on the Barthel Index (95 % CI 0 to 20) for functional independence, the primary outcome, 8 points on the 60-point MRC scale for strength (95 % CI 4 to 18), -13 points on the Fatigue Severity Scale scored from 0 to 63 (95 % CI -28 to -1), and 12 points on the environment domain of quality of life (95 % CI 3 to 13). At twelve months, the estimated effects were of comparable magnitude, but most confidence intervals were wider 20.

This result has to be read with its lower bound: for the primary outcome, the confidence interval touches zero. What the trial shows most solidly is not the gain in independence, it is the gain in strength and the reduction in fatigue. And it covers sixteen people.

What the syntheses retain

The most recent systematic review analysed thirteen papers totalling 173 participants, of very heterogeneous methodologies: case studies, randomised and non-randomised clinical trials, a prospective study, case series, a replicated single-case design. Five of those studies included patients with CIDP, for 37 participants. Its synthesis is that multicomponent programmes combining strengthening, aerobic work, balance and breathing exercises are associated with improvement in fatigue and functional capacity. It notes a recurring pattern in the literature analysed: the best results seem linked to supervised, individualised sessions of 45 to 60 minutes, conducted three to four times a week for at least twelve weeks, while itself specifying that this observation is drawn from heterogeneous designs 21.

A scoping review published in 2025 screened 1,021 papers to retain only sixteen, and concludes that physical exercise can improve strength, reduce fatigue and support functional independence, calling for larger, better quality studies and for research into the mechanisms of chronic fatigue 22. A 2026 literature review devoted to the general aspects of rehabilitation stresses for its part the early, coordinated and multidisciplinary character of care 32.

Modalities and levels of evidence

Should you fear overworking a nerve that is repairing?

The question arises exactly as in motor neurone disease, and it calls for a different answer. What changes is not the quality of the data, which is weak on both sides: it is the direction of the disease.

The fear, and its origin

The notion of overwork weakness, overwork weakness, was formulated in 1958 about partially denervated skeletal muscle, from the observation of patients with polio 29. It has been transposed to peripheral nerve damage as a whole, and it continues to inspire a caution that has never been tested in Guillain-Barré syndrome: no trial has compared a high load with a low load there.

One must nevertheless name what distinguishes this situation from a degenerative disease. Guillain-Barré is monophasic and recovering. The nerve remyelinates, the axons regrow, and the programme accompanies a repair in progress instead of fighting a loss that continues. That difference of direction authorises no particular load, but it changes the reasoning: a progression of load that follows recovery is consistent with what the disease is doing, whereas in amyotrophic lateral sclerosis it runs against it.

What the published programmes actually dosed

The documented doses are, once again, those of the protocols. Twelve weeks of cycling in severely fatigued, well recovered patients 14. Supervised 60-minute sessions, two to three times a week for twelve weeks, combining strengthening, endurance, breathing exercises, walking and pain management, in patients whose residual disability had been stable for at least six months 20. Multicomponent programmes of 45 to 60 minutes, three to four times a week, over at least twelve weeks 21.

Two characteristics of these protocols count as much as their volume. The first is that all were conducted in the chronic or recovery phase, never during the worsening. The second is that the comparator in the only randomised trial was not the absence of exercise but a less intense home programme: what has been demonstrated is therefore the superiority of supervision, not that of exercise in itself.

An approach that takes account of the uncertainty

1. Progress with the recovery, not against the fatigue. The load follows what the patient regains; it does not try to force what they have not yet recovered.
2. Write the stopping rule before the session. Fatigue that persists beyond thirty minutes after the effort, incomplete recovery the next day, a clear and dated loss of strength: the first two lower the load, the third means writing to the neurologist.
3. Distinguish the fatigue that accompanies effort from the one that structures the day. The first is managed by dosing; the second is a symptom of the disease, and it will not yield to exercise volume alone.
4. Keep the calendar. The date of symptom onset, that of the nadir, that of treatment and that of any later deterioration. It is the item most often missing from the notes, and the one the physiotherapist is best placed to keep.

What to do about pain, which lasts longer than people think?

Pain is frequent, often intense, and regularly absent from the discharge summary. It also concerns patients in whom the examination finds no sensory disturbance.

Its frequency over time

A prospective cohort of 156 patients, including 18 with Miller Fisher syndrome, assessed the location, type and intensity of pain by questionnaire at standardised times over one year. Pain was reported by 36 % of patients in the two weeks preceding the onset of the deficit, by 66 % in the acute phase, defined as the first three weeks after inclusion, and still by 38 % one year later. In the majority, its intensity was moderate to severe, and the longitudinal analysis finds high mean scores throughout the follow-up 12.

Three details of that study change how it should be asked about in the clinic. Pain occurs across the whole spectrum of the disease, including in patients with Miller Fisher syndrome, in mildly affected patients and in purely motor patients. Its mean intensity is highest in patients with sensory disturbance and in severely affected patients. And the severity of the deficit and of the disability correlates with the intensity of the pain only at the late stages of the disease 12.

Why a purely motor patient can be in pain

A prospective study in 32 patients measured intraepidermal nerve fibre density at the distal leg and in the lumbar region, and related it to pain, dysautonomia and outcome. In the acute phase, that density was reduced in 60 % of patients at the distal leg and in 61.9 % in the lumbar region, and 43.7 % complained of neuropathic pain. In the latter, density at the distal leg was significantly lower than in patients without pain (p < 0.001) and correlated with the intensity of the pain. A notable point raised by the authors: patients with a purely motor variant who had pain also showed low density 27.

That result has a direct consequence for the clinic: a normal sensory examination does not excuse you from asking about pain, and pain reported by a patient labelled “pure motor form” is not suspect. It has a measurable substrate.

What is recommended, and what falls to the physiotherapist

On the drug side, the European guideline weakly retains gabapentinoids, tricyclic antidepressants or carbamazepine 1, and the 2005 multidisciplinary consensus on supportive care already noted that pain management there is difficult 26. On the rehabilitation side, pain management was explicitly among the components of the supervised programme in the 2022 randomised trial 20, alongside strengthening, endurance, breathing exercises and gait retraining. It is not a separate treatment: it is one of the five strands of the programme.

How do you measure recovery, and spot what is not recovery?

Three measures are enough, provided they are dated and repeated at the same interval. Their main value is not to document progress: it is to make an inflection visible.

The three measures to keep

The only randomised trial conducted in the chronic phase gives a direct answer to the question of instruments, since it had to choose them: the 100-point Barthel Index for independence in activities of daily living, retained as the primary outcome; the 60-point MRC scale for segmental strength; the Fatigue Severity Scale scored from 0 to 63; a visual analogue scale for pain intensity, and a quality of life measure 20. That set is reproducible in the clinic, with no equipment.

The fatigue scale deserves a particular place, for a reason that runs through this whole chapter: it is the only one of the three that measures the dominant symptom at a distance, and it is also the one the patient will never see mentioned in their discharge summary. Measuring it already signals to the patient that it is regarded as a clinical fact and not as a complaint.

A three-year follow-up study, conducted in 24 adult patients assessed on admission, at discharge, then at one year and at three years, used a similar set: the Functional Independence Measure, the Functional Ambulation Classification, the Hughes disability scale, the six-minute walk test and the Fatigue Severity Scale 36. The six-minute walk test is a useful addition once the patient is walking again: it captures endurance where analytical testing sees only preserved strength.

What should still be monitored, even at a distance

About a quarter of patients develop respiratory failure during the illness, and many show signs of autonomic dysfunction 7. A 2025 review is a reminder that respiratory failure concerns up to 20 % of cases, and that new treatments, notably antibodies blocking the C1q component of complement, are under study 37. These figures concern the acute phase, but they justify keeping in mind, during the recovery phase, two simple markers: the patient's ability to finish a sentence without pausing for breath, and their tolerance of lying flat.

The neuroprognostication guidelines insist on a principle that holds here: prognostication must take account both of the acute phase and of the recovery phase, and consider the complete clinical picture rather than an isolated variable 11. In other words, a strength score that climbs while fatigue worsens and participation falls back is not good news: it is a whole to be looked at together.

What do published cases teach us?

Two recent reports describe a structured programme and how it unfolded. They demonstrate nothing, and they illustrate well what the recovery phase demands.

How do you apply this in the clinic?

Five stages. The first and the last are the ones most often skipped, and they are the ones that count most.

Stage 1: reconstruct the calendar before doing anything else

Date of the first symptoms, date of the nadir, nature and date of the treatment received, date of discharge. Those four dates make it possible to place the patient on the curve and to know, at every later session, whether a deterioration falls before or after the eighth week. It is the information most often missing, and it is collected in three minutes at the first assessment. The international ten-step management guide in fact structures the whole pathway around that chronology, from early recognition to the management of complications and sequelae 2.

Stage 2: set a baseline on the three symptoms that count

Segmental strength, fatigue, independence in activities of daily living. Add an assessment of pain, including if the sensory examination is normal 27. Repeat at the same interval, noting the date each time.

Stage 3: build a supervised, multicomponent programme

Strengthening, endurance, balance and breathing work, in a session of 45 to 60 minutes, three to four times a week, over at least twelve weeks: that is the format the literature associates with the best results 21, and it matches the protocol of the only randomised trial available 20. The point that trial documents best is supervision: at comparable volume, it made the difference.

Stage 4: treat fatigue as an objective, not as an obstacle

No drug treatment is recommended for it 1, it correlates with neither the initial severity nor electrophysiology 1316, and training is the only intervention that has moved it in the literature 1420. It therefore deserves a written objective and a measure, in the same way as strength.

Stage 5: know when to stop adjusting and write

A new deterioration, whatever its intensity, is noted, dated and passed on the same day. Before two months, it suggests a treatment-related fluctuation; beyond eight weeks, it calls the label itself into question 1. In both cases, what the neurologist expects is not an impression but two figures and two dates: the previous score, today's score, and the days between them.

Frequently asked questions