Amyotrophic lateral sclerosis (ALS, motor neurone disease)

Amyotrophic lateral sclerosis is one of the rare conditions in which the physiotherapist works knowing that function is going to be lost. That does not make their role a minor one: it moves it. The question that governs everything is not whether to rehabilitate, but at what dose, and on what evidence. Two answers have been in conflict for sixty years: a fear, that of overworking an already sick motor neurone, formulated in 1958 from patients with polio; and a series of randomised trials conducted since 2001, all small, which have never found that overwork at the intensities they tested. This article sets out what those trials actually measured, what they do not allow us to say, and what the literature establishes about respiratory support and airway clearance, which is probably where the physiotherapist counts most. One clarification from the outset: on the optimal dose, there is no data. We shall say so rather than compose a reassuring answer.

• Updated 16 August 2026
• Level Clinical summary
• Sources 39 verified references
• ICD-11 8B60.0

What is ALS, and what does the physiotherapist see first?

A degeneration that affects both levels of motor control. What makes it recognisable in the clinic is not an isolated sign, it is the coexistence of signs that elsewhere do not go together.

Two motor neurones, one disease

Amyotrophic lateral sclerosis is a neurodegenerative disease that affects both the upper motor neurones, in the motor cortex, and the lower motor neurones, in the brainstem and the anterior horn of the cord 2. It is that dual involvement that makes its clinical signature: loss of the upper motor neurone gives stiffness and spasticity, loss of the lower motor neurone gives weakness, fasciculations, wasting and flaccidity. Finding both in the same territory is unusual, and that is precisely what should draw attention.

The disease almost always begins focally, then spreads to the other regions. Onset is in a limb in about 65 % of cases, with weakness of the hand or a foot drop, in the cranial muscles in 20 to 25 % of cases, with difficulty speaking or swallowing, and in the axial musculature in 5 to 10 % of cases, with a forward-leaning posture 2. About 85 % of cases are sporadic and 15 % familial; more than sixty genes have been associated with the disease, including C9orf72expansions, found in 40 % of familial forms 2.

The disease is not purely motor. Up to half of patients have extra-motor manifestations, behavioural change, executive dysfunction, language disturbance, and 10 to 15 % meet the clinical criteria for frontotemporal dementia 3. That point has a direct practical consequence: a home programme assumes a planning capacity that the disease may have eroded, and an instruction poorly followed is not always a failure of motivation.

How many people are affected?

ALS is rare, and the published figures vary enormously between countries and with the quality of data collection. The largest systematic review, which retained 140 papers published between 2010 and 2021, gives an incidence ranging from 0.26 per 100,000 person-years in Ecuador to 23.46 in Japan, and a point prevalence ranging from 1.57 per 100,000 in Iran to 11.80 in the United States 4. That spread is as much down to methods of case-finding as to real differences, which the authors themselves stress.

The most solid figure available in France comes from the Limousin register, whose completeness has been estimated at 98.4 % by capture-recapture: 279 incident cases between 2000 and 2011, that is a crude incidence of 3.19 per 100,000 person-years and an incidence standardised to the European population of 2.58 5. The standardised male-to-female ratio there is 1.3 overall, but it rises with age: 1.1 before 65, 1.7 between 65 and 75, 1.9 beyond.

How is the diagnosis made, and why does it take so long?

The diagnosis remains clinical, supported by electromyography. The criteria were simplified in 2020, and that simplification has a direct consequence for what the physiotherapist can see before the neurologist.

What the Gold Coast criteria changed

The historical El Escorial criteria, then Awaji, classified patients into categories of certainty (definite, probable, possible) according to the number of regions involved. Complex and error-prone, they excluded patients who genuinely had the disease. The Gold Coast criteria, published in 2020, use a simpler formulation: documented progressive impairment, upper and lower motor neurone signs in at least one region or lower motor neurone signs in at least two regions, and the exclusion of other causes by investigations 8.

Two independent series have measured what that simplification brings. In 506 patients reviewed retrospectively, the sensitivity of the Gold Coast criteria is 92 % (95 % CI 88.7 to 94.6), against 90.3 % for Awaji and 88.6 % for revised El Escorial, with that sensitivity maintained whatever the site of onset, the duration of illness and the level of disability, and a clear gain in atypical phenotypes 6. In a European multicentre series of 404 patients referred for electromyography on suspicion of ALS, of whom 272 had the diagnosis confirmed by the course and 94 had an ALS-mimicking condition, the sensitivity of the Gold Coast criteria is 88.2 % (95 % CI 83.8 to 91.7) 7.

One figure from that last series deserves to be remembered by the physiotherapist: of 404 patients referred with a suspicion, 94 had something else. A patient referred for investigation does not have ALS, they have a suspicion of ALS, and those two situations do not call for the same way of speaking.

What the physiotherapist sees and the consultation does not

The diagnosis rests on demonstrating progression. But progression is seen over time, and it is the professional who sees the patient every week who sees it first. Three observations are worth dating and passing on:

• wasting that sets in over a few weeks in a precise territory, in particular the first dorsal interosseous or the thenar eminence;
• diffuse fasciculations, visible at rest, together with weakness and not in isolation;
• a brisk reflex in a territory that is otherwise wasted and weak, which is the clinical expression of the dual involvement.

A motor deficit that progresses, with no sensory disturbance, with brisk reflexes in a muscle that is melting away: that combination has no benign equivalent. It is not managed by adjusting the programme, it is written in a letter.

What the treatments do, and what they do not

There is no curative treatment. Riluzole and edaravone slow progression, at best by two to four months, and tofersen is for carriers of a pathogenic SOD1 2variant. The most important fact for a physiotherapist lies elsewhere: management by a multidisciplinary team, including neurologists, nurses, physiotherapists, occupational therapists, speech and language therapists, dietitians and social workers, is associated with a survival gain of four to seven months and with better quality of life 2. In other words, the organisation of care counts for more than the molecules available, and the community physiotherapist is part of it.

The 2024 European guideline, built following GRADE methodology from 26 research questions, covers disease-modifying treatments, multidisciplinary care, nutritional and respiratory support, communication aids, psychological support, common symptoms and end of life 1. Its authors themselves note that the evidence available was scarce for many of those questions, which the rest of this article keeps on finding.

Should you fear overworking a sick motor neurone?

This is the question that governs everything else. It has a history, a date of birth, and a partial answer. The part of the answer that is missing is the one that matters most in practice.

Where that fear comes from

The notion of overwork weakness, overwork weakness, was formulated in 1958 by Bennett and Knowlton, about partially denervated skeletal muscle 9. The idea is simple and intuitive: when one motor unit in three has gone, the remaining units take on the whole of the task, and asking more of them would hasten their exhaustion. Formulated in the era of polio sequelae, it was transposed without discussion to motor neurone diseases, and it lastingly led to effort being discouraged in these patients.

It is worth naming precisely what makes that transposition fragile. Polio is a residual, stable condition, where the motor neurones lost were lost once. ALS is an ongoing degeneration, where the mechanism of the loss is not mechanical loading but the disease itself. Nothing guarantees that reasoning valid in one is valid in the other, in either direction. It is a hypothesis that has never been directly tested, and people have long acted as though it had been demonstrated.

What the trials actually measured

Four randomised trials structure the question. They are all small, none was designed to detect a rare adverse effect, and it is with that limitation in mind that they must be read.

Drory 2001. Twenty-five patients randomised between a daily moderate exercise programme (n = 14) and no activity beyond usual needs (n = 11), followed for twelve months. At three months, the exercise group deteriorates less on the functional scale and on the Ashworth scale; at six months, the difference is no longer significant, with a persisting trend in favour of exercise; at nine and twelve months, the numbers remaining no longer allow statistical analysis 10. That last sentence alone says what the prognosis costs clinical research in this disease.

Bello-Haas 2007. Twenty-seven patients randomised between a home programme combining daily stretching and strengthening three times a week (n = 13) and daily stretching alone (n = 14). Inclusion required a forced vital capacity of at least 90 % of predicted and an ALSFRS score of at least 30, that is, patients at an early stage. Eighteen completed. At six months, the strengthening group has a significantly higher total ALSFRS and SF-36 physical subscale, no adverse effect related to the intervention occurred, maximum voluntary isometric contraction and vital capacity show no negative effect, and the loss of strength in the lower limbs is smaller in the strengthening group 11.

Clawson 2018. Three arms: resistance, endurance, or stretching and range of motion, the last being the regimen then prescribed to most patients. All the exercises are done at home, with an individualised programme designed by a physiotherapist trained in ALS. The primary outcome was tolerability at twenty-four weeks, defined as the ability of half the participants to complete at least half of the prescribed programme: it is met in all three arms, at twelve as at twenty-four weeks. No difference in rate of progression between the groups, no difference in the secondary outcomes, adherence highest in resistance and stretching, and a post-hoc trend towards fewer falls in resistance and endurance 12.

Kalron 2021. Thirty-two ambulatory patients randomised between a combined programme (recumbent cycling, functional strengthening, stretching and passive mobilisation) and stretching alone at home, two sessions a week for twelve weeks. Twenty-eight completed. Group-by-time analysis finds a difference in favour of the combined group in respiratory function, mobility and ALSFRS-R: the patients in the combined group maintain their capacities, while the stretching group declines. The SF-36 categories “physical functioning”, “energy and fatigue” and “well-being” are also higher after the intervention 13.

What the syntheses make of it

Three recent syntheses pool those trials, and they do not say quite the same thing, which is instructive. They follow the Cochrane review devoted to therapeutic exercise in ALS and motor neurone disease, whose last update dates from 2013 and whose stated premise was already that the effects of exercise in this population remained poorly understood 18. A more recent network meta-analysis has also compared five families of non-drug interventions with each other, exercise, nutrition, respiration, psychological support and integrated physical intervention, taking the ALSFRS-R as its primary outcome 19.

The Meng 2020 meta-analysis, on seven trials and 322 patients, finds higher long-term functional scores with exercise (standardised mean difference 0.47; 95 % CI 0.08 to 0.86; p = 0.02) and a higher predicted percentage of forced vital capacity (mean difference 1.71; 95 % CI 0.10 to 3.31; p = 0.04), with no difference in muscle strength or in quality of life. Endurance or aerobic exercise improves functional scores (standardised mean difference 0.36; 95 % CI 0.04 to 0.68). Above all, exercise worsened neither fatigue nor the rate of adverse events 14.

The Souza 2025 systematic review, focused on non-respiratory strengthening, includes eight studies and 296 individuals. Seven of the eight show a superiority of the experimental intervention, but that superiority is not confirmed by the meta-analyses: the small sample sizes and the strong heterogeneity of the primary studies bring the quality of the evidence down to between low and very low. Five of the eight studies assessed adverse events, reporting none that were serious 15.

The Silva 2024 systematic review, published in BMJ Open, is the strictest. Of 39,415 references screened, only three studies met its criteria, totalling 62 participants. Physiotherapy might improve overall function in the short term, but the clinical interpretation remains inconclusive; on quality of life and fatigue, it is not superior to control in the short term; and it does not increase adverse events 16. Three studies out of nearly forty thousand references: that ratio is the real state of the question.

A more concrete figure, from a systematic review of ten clinical trials, helps picture the order of magnitude of the possible benefit: at six months, the mean loss was 5.8 ALSFRS-R points in the exercise groups against 7.6 points in the control groups 17. That is not a halt to the disease, and it must not be presented as one: it is a slightly less steep slope.

What dose of exercise, and how far can you go?

For want of a dose trial, the only defensible position consists of staying within the envelope tested, working below the fatigue threshold, and writing down in advance the rule that will make you back off.

What the data allow you to say

The documented doses are those of the protocols, and they are modest. Home strengthening three times a week with daily stretching, in patients whose forced vital capacity exceeded 90 % of predicted 11. Two weekly sessions combining recumbent cycling, functional strengthening and stretching, for twelve weeks, in patients still walking 13. An individualised programme designed by a physiotherapist trained in the disease, and judged on its tolerability before its effect 12. That is the whole of what is known.

Two features common to those protocols deserve noting, because they are easier to reproduce than the load itself. The first is that all included patients at an early or intermediate stage, never patients in respiratory failure. The second is that all included stretching and range-of-motion work, including in the control arms: that is the base on which strengthening is added, not what it replaces.

What the data do not allow you to say

No trial has compared two intensities with each other. There is therefore no published threshold above which exercise would become harmful, nor any demonstration that such a threshold does not exist. No trial has tested heavy-load strengthening or work close to muscular failure. None has included patients at an advanced stage, which is probably the population where the question of dose arises most acutely, and it is also the one that fits the demands of a randomised trial least well.

Three statements that often get confused must therefore be distinguished:

• “Moderate exercise has not shown harm in the trials”: true, with the reservations about sample size noted above.
• “Exercise slows ALS”: not established. The most rigorous syntheses conclude to low to very low quality evidence and to an inconclusive clinical interpretation 1516.
• “Loading heavily is safe”: outside the scope of the data. Nobody has tested it.

In a disease where time is short, uncertainty must not produce inaction. It must produce a written dose, a dated measurement, and a stopping rule decided before the session rather than during it.

Modalities and levels of evidence

The cards below summarise what the literature supports, modality by modality. The level of evidence concerns the effect, not safety: several modalities rated as low are nevertheless reasonable to offer, because their cost is nil and their functional benefit plausible.

An approach that takes account of the uncertainty

For want of a published threshold, it is rules of conduct, not figures, that protect the patient. Four hold up against the literature:

1. Stay within the envelope tested. Two to three sessions a week, an intensity that allows the session to be finished without exhaustion, stretching as a base. Leaving that envelope means leaving the data.
2. Set the stopping rule before the session. Fatigue that persists more than thirty minutes after the effort, new breathing difficulty, incomplete recovery the next day: those three signals lower the load at the following session. Writing the rule down in advance avoids negotiating it in the moment, with a patient who wants to do more and a therapist who wants to believe in it.
3. Adapt to the region, not only to the patient. A territory already badly wasted is not strengthened: it is protected from contractures and from excessive loading. Strengthening applies to what still responds.
4. Reassess whether it is still warranted at each stage. The programme useful at the walking stage is not the one for the wheelchair stage. A programme that does not change while the disease changes is no longer a programme, it is a habit.

How do you support breathing?

Respiratory failure is the commonest cause of death. It is also the area where the best-demonstrated intervention exists, and where the physiotherapist has a lookout role as much as a technical one.

What is at stake

Failure of the respiratory muscles is the leading cause of death in ALS, and it usually occurs within two to five years of the onset of the disease 21. Respiratory management has two distinct objectives: to support ventilation when it becomes insufficient, and to maintain the capacity to clear the airway. The two are not the same thing, and the second is often neglected.

Two clinical markers should alert the physiotherapist before any investigation: orthopnoea, that is difficulty breathing when lying flat, which reflects diaphragmatic involvement, and the signs of nocturnal hypoventilation, unrefreshing sleep, morning headaches, daytime sleepiness. Those are the criteria on which the benchmark trial started non-invasive ventilation, together with a maximum inspiratory pressure below 60 % of predicted or symptomatic hypercapnia 20.

Non-invasive ventilation, and what bulbar function changes

Bourke's 2006 trial remains the reference data. Forty-one patients randomised between non-invasive ventilation (n = 22) and standard care (n = 19) at the point when they developed one of the above criteria. In patients whose bulbar function was preserved, ventilation brings a median survival benefit of 205 days (p = 0.006), with quality of life maintained for most of that period. In patients with severe bulbar involvement, it improves certain quality of life indices, notably the symptoms domain of the sleep apnoea quality of life index (p = 0.018), but brings no survival benefit 20.

The 2017 Cochrane review rests on that single trial, which it acknowledges explicitly: across all participants, median survival is 48 days longer with ventilation (219 days against 171; 95 % CI 12 to 91; p = 0.0062), moderate quality evidence. Its authors note that producing new randomised data will be difficult, since it is no longer ethically justifiable to deprive a control group of ventilation, and they call for study of adding cough augmentation techniques to ventilation 21. Seven years later, that question remains open.

Tracheostomy prolongs survival by about two further years, but up to 95 % of American patients choose not to have it 28. That figure is a reminder that the decision is not only a medical one, and that the physiotherapist's role is to support a decision made elsewhere, not to weigh in on it.

Respiratory muscle training

Two sham-controlled trials, conducted by the same team, frame what this modality brings.

The first, published in 2019, covers 48 patients allocated between active and sham expiratory training for eight weeks. Ninety-six per cent completed the protocol, and the training significantly improves maximum expiratory pressure and the swallowing safety score on dynamic imaging (p < 0.02), with no effect on the other secondary outcomes 23.

The second, published in 2023, is multicentre and double-blind: 45 patients at an early stage, twelve weeks of inspiratory and expiratory training, load at 30 % against 0 %. The completion rate is 91 % and no adverse effect related to the training occurred. Maximum expiratory pressure increases on average by 20.8 cm H₂O (95 % CI 7.6 to 33.9) with active training and falls by 1.0 cm H₂O with sham training (p = 0.004). Maximum inspiratory pressure is not significantly improved (p = 0.33), nor is forced vital capacity (p = 0.60). Peak cough inspiratory flow improves (p = 0.02), cough expiratory flow remains borderline (p = 0.06). At twelve months, the slope of the bulbar subscale of the ALSFRS-R declines twice as fast in the sham group as in the active group (-0.29 against -0.12 points a month; p = 0.02), with no difference in the total slope, in feeding status or in time to non-invasive ventilation 22.

The 2025 meta-analysis, on five trials and 170 participants, confirms the gain in respiratory strength, in maximum expiratory and inspiratory pressures, the absence of effect on forced vital capacity and on quality of life, and the absence of reported adverse effects 24. The Cochrane review devoted to respiratory training across neuromuscular diseases already concluded that its effects remained uncertain 25.

How do you clear the airway when coughing is no longer enough?

Respiratory infections are a major cause of morbidity and mortality in neuromuscular disease. The airway clearance strategy is built before it becomes urgent, and setting it up late is a missed opportunity.

A cough that deteriorates in three stages

An effective cough requires three things in sequence: a deep inspiration, glottic closure, an explosive expiration. ALS can affect all three. Inspiration deteriorates with inspiratory weakness, glottic closure with bulbar involvement, expiration with abdominal and expiratory weakness. Insufficient cough typically shows as a prolonged, slow and weak cough effort, which no longer clears secretions and no longer protects the airway 26.

Two points from the 2024 review deserve to be applied as they stand. First, respiratory strength measurements should be taken and followed every three to six months, preferably in a multidisciplinary setting. Second, cough augmentation, whether manual or mechanical, should be put in place as early as possible in the progression of the disease, to control proximal airway secretions and prevent respiratory infections 26. The Sheers 2024 review stresses in the same direction the role of peak cough flow as an assessment instrument, alongside the other options available 27.

The point that defies intuition

Mechanical insufflation-exsufflation is the reference tool for cough augmentation. But its effectiveness can be compromised by severe bulbar involvement, precisely in the patients who would need it most: collapse of the upper airway during the exsufflation phase prevents the flow from being effective 26. That is not a reason not to try it, it is a reason to assess its real effect rather than assume it.

Symmetrically, peripheral strategies such as high-frequency chest wall compression have the advantage of being less affected by bulbar involvement, but the recommendation is explicit: they are used in combination with proximal strategies, never in their place 26. Mobilising distal secretions in a patient unable to clear them proximally makes the situation worse instead of resolving it.

Salivary secretions, which are not the same job

Salivary pooling is not treated by bronchial clearance. Normal saliva production is 0.5 to 1.5 litres a day; when swallowing deteriorates, that saliva stagnates and becomes a major nuisance. The Cochrane review devoted to the subject assesses drug treatments, radiotherapy and surgery 33. The physiotherapist does not act on it directly, but is often the professional who notices the problem and can pass it on: a British national survey of practice in supporting cough and secretions in ALS shows that physiotherapists made up more than half the respondents, and that 71 % of them reported a role in managing saliva and secretions 39.

What do you do about the symptoms that weigh most in daily life?

Fatigue, pain, cramps, spasticity. These are the complaints that come back every week, and they are not treated as in a patient without motor neurone disease.

Fatigue, almost universal, and rarely simple deconditioning

Two sources give apparently contradictory figures, and the contradiction is instructive. The Hamad 2024 meta-analysis, which retained only studies measuring fatigue with a validated instrument and a defined threshold, brings together eleven studies and 1,072 patients and finds a pooled frequency of 48 % (95 % CI 40 to 57). That figure rises to 62 % (95 % CI 43 to 79) when the ALSFRS-R score is below 30, against 43 % (95 % CI 37 to 49) when it is 30 or above, and meta-regression confirms the negative association between fatigue and ALSFRS-R (p = 0.02) 29.

The Young 2026 cohort study, on 1,058 patients, measures differently: it asks about the presence of the symptom rather than the crossing of a threshold, and then finds fatigue reported by 97.8 % of patients on the disease-specific motor neurone disease scale. Grading it, 27.3 % have mild fatigue, 36.1 % moderate and 36.2 % severe. Over a mean follow-up of 11.6 months, 60.5 % have stable fatigue, 23.8 % increasing fatigue and 15.8 % decreasing fatigue. The structural equation model ranks the determinants in decreasing order of effect: function, cognition, spasticity, breathlessness and pain 30.

That list of determinants is directly usable in the clinic. Increasing fatigue is not treated by increasing exercise: it is treated by looking for which of those five levers has moved. Worsening spasticity, breathlessness appearing, poorly controlled pain are more likely explanations than deconditioning, and two of them call for a medical reassessment rather than an adjustment of the programme.

Pain, frequent and undertreated

Pain is a widely neglected symptom in ALS, although the majority of patients report it. It occurs at every stage and may even precede motor dysfunction; it correlates with a deterioration in quality of life and with an increased prevalence of depression, and at late stages it is among the events that predict clinical deterioration and death 32. A cross-sectional series comparing 46 patients with 46 matched controls and with 23 controls with myotonic dystrophy type 2 finds pain in 78 % of ALS patients against 54 % of controls, with moderate to severe pain in 42 % against 20 %, and significantly greater interference with daily activities 31.

The useful distinction in practice is one of mechanisms: cramp pain, nociceptive musculoskeletal pain related to immobility, sustained positions and joint loading, and neuropathic pain 32. The second is the one on which the physiotherapist acts most directly: positioning, range-of-motion work, prevention of contractures, adaptation of support surfaces and transfers.

Cramps, spasticity, sleep

Cramps, pain, spasticity, immobilisation, restless legs syndrome, sleep-disordered breathing and difficulty clearing secretions all contribute to disrupted sleep in ALS, to which are added existential anxiety and depression 38. A patient who sleeps badly is a patient in whom exercise will be less well tolerated, and asking about sleep is therefore part of the assessment.

The 2024 European guideline makes specific recommendations on muscle cramps, spasticity, pseudobulbar affect, thick secretions, sialorrhoea and pain 1. The Cochrane review of symptomatic treatments in motor neurone disease, which synthesises all the systematic reviews available, stresses for its part the weakness of the evidence base for those treatments 34. Here again, uncertainty is the rule rather than the exception.

How do you measure, and at what interval?

In a progressive disease, measurement does not serve to prove progress. It serves to document a slope, to see an inflection, and to give the neurologist information they do not have.

The two measures to keep

The revised ALS functional rating scale, the ALSFRS-R, is the instrument used as the primary outcome by almost all the trials and reviews cited in this article 17. It is not a physiotherapy instrument, but it is the common language of everyone following the patient: a dated score, repeated at the same interval, turns an impression into a slope. That is exactly the form in which the trials use it, in points per month.

Respiratory strength measurements, taken every three to six months, form the second series 26. They serve two distinct decisions: when to introduce non-invasive ventilation, which belongs to the doctor, and when to put cough augmentation in place, which is prepared much earlier.

Those two series are only of value in relation to the stage. A review devoted to the place of rehabilitation in ALS puts it simply: the clinician must know the expected trajectory of the disease and apply, at each stage, the intervention that corresponds to it 37.

A third measure, less formal, deserves to be noted at every session: what the patient has stopped doing since last time. Climbing stairs, getting up from a low chair, holding a fork, turning over in bed. It is the most useful information in the notes, and it is the one nobody else collects.

What do published cases teach us?

The physiotherapy case literature in ALS is thin, and what it describes best is not a protocol: it is the role of the initial assessment and the limits of what a technique can bring.

How do you apply this in the clinic?

Five stages, in this order. They do not remove the uncertainty: they make it manageable, and they leave a record of what was decided and why.

Stage 1: establish a quantified baseline, from the first session

An ALSFRS-R, analytical testing of the territories concerned, a sit-to-stand time and a walking distance if it is still measurable. The first record is only of value through the ones that follow: it is the slope that informs, not the isolated value. Also note the date of symptom onset as the patient reports it, which serves as the reference for the whole later trajectory.

Stage 2: set the programme within the documented envelope

Two to three sessions a week, moderate intensity, daily stretching as a base, and strengthening applied to the territories that still respond. That is the format of the trials that showed functional maintenance superior to stretching alone 1113. The programme is partly done at home, as in the three trials that tested it, and that means checking that it is genuinely feasible: in the three-arm trial, the primary outcome was tolerability, not effect 12.

Stage 3: write the stopping rule before you need it

Three signals, noted in the record and explained to the patient from the first session: fatigue that persists beyond thirty minutes after the effort, incomplete recovery the next day, new breathing difficulty. Each lowers the load at the following session; the third also means calling the doctor. That rule protects against drift in both directions, the one that loads too much because the patient asks for it, and the one that no longer dares anything because the disease is frightening.

Stage 4: prepare the breathing before the incident

Take respiratory strength measurements every three to six months, systematically ask about orthopnoea, sleep and morning headaches, and have an airway clearance strategy put in place early in the course rather than in the aftermath of the first infection 26. When mechanical cough augmentation is in place, check its real effectiveness, particularly if bulbar involvement is progressing.

Stage 5: write to the reference centre, with figures

The physiotherapist is often the only professional who sees the patient every week, and multidisciplinary care is associated with a survival gain of four to seven months 2. A useful letter comes down to a few lines: the ALSFRS-R slope over the last few months, what the patient has stopped doing since the last consultation, the respiratory signs asked about and their answers, and the precise question that prompts the letter. An impression cannot be passed on; a slope can.

Frequently asked questions