Charcot-Marie-Tooth Disease

What Works

Dark cinematic study of the foot and ankle under clinical lighting

Tier 1 — Strong evidence

1. Recognise it and refer for diagnosis HIGH

The highest-value action on this page, and it costs nothing. Both clinical practice guidelines state that the diagnosis is clinical, made by examination rather than by scanning the sore part. Confirmation is genetic, and it carries implications for the whole family.

Sivera Mascaró 2025 (PMID 38431252); Yiu 2022 (PMID 35140138)

2. Screen the hips and the spine, not just the foot MODERATE

Hip dysplasia was present in 9.6% to 28.1% of 178 children with this condition, and 9.6% needed surgery. Scoliosis appeared in 45 of 298 patients, where 24 of 34 followed curves progressed and bracing succeeded in only 3 of 16. Both are invisible to a foot-focused assessment, and both are the single most consequential omission in this condition.

Tan 2025 (PMID 39887493); Karol 2007 (PMID 17606789). Both are retrospective single-centre series, which is why this is MODERATE rather than HIGH.

Tier 2 — Moderate evidence

3. Progressive resistance training of the foot dorsiflexors MODERATE

The dose, quoted from the trial rather than paraphrased: 50% rising to 70% of the most recent one-rep max, 3 sessions per week on non-consecutive days, for 6 months (72 sessions), both feet, using an adjustable ankle cuff, with the load re-anchored to a re-tested max.

Read the timeline before you promise anything. This is the field's only double-blind sham-controlled exercise trial, and it found a strength difference of zero at its 6-month primary endpoint (95% CI −0.37 to 0.42; p=0.91). The significant result appeared at 24 months (0.6; 95% CI 0.03 to 1.12; p=0.041), eighteen months after the training stopped. The honest promise is a slower decline, not strength gained. A patient told otherwise quits at month six, exactly when it looks like failure and is not. The trial was in children aged 6 to 17 with CMT1A, and applying it to adults is extrapolation.

Burns 2017 (PMID 30169201), graded moderate certainty by Conde 2023 (PMID 37060329)

4. Offer exercise, without an overwork-weakness warning MODERATE

The fear that exercise "overworks" already-weak muscles kept this population out of the gym for decades. It has now been tested three independent ways and not confirmed: MRI muscle and intramuscular fat volume were unchanged against sham at 6 and 24 months; grip strength does not differ between the dominant and non-dominant hand across 166 people, which is a lifetime of asymmetric use; and treadmill training produced no overwork weakness and no deterioration.

Burns 2017 (PMID 30169201); Roberts-Clarke 2016 (PMID 27177353); Mori 2020 (PMID 31444929). None is a dedicated harm trial, which is why this is MODERATE.

5. Ankle-foot orthosis, customised to the person LOW

78% of 306 brace users say their brace improves their balance. Pooled across 15 studies, the objective gait effects were small-to-moderate and non-significant, and balance could not be pooled at all. That is a measurement gap rather than a refutation: the studies had samples as small as one and wildly different brace designs. Keep prescribing them, customise them, and be honest that the literature cannot currently show the effect people report feeling.

Kim 2024 (PMID 39276325); Anderson 2025 (PMID 40264638)

6. Assess falls and balance confidence explicitly MODERATE

86% of 107 people reported falls or near falls, most of them at home. Among 306 brace users, 14% had fallen in the previous 24 hours and 38% in the previous week. Confidence holds above 60% for level walking, drops to 40-55% on stairs and slopes, and falls below 40% on icy ground, uneven surfaces, or when bumped. Ask about those specific situations rather than asking "do you fall?"

Ramdharry 2018 (PMID 29282812); Anderson 2025 (PMID 40264638). No trial in this literature has ever used falls as an outcome, so there is no evidence that any intervention reduces them. NO EVIDENCE

Tier 3 — Emerging, and one of them is partly negative

7. Home-based balance training with proximal strengthening

A 12-week home programme with three physical therapist visits produced strong effect sizes on functional balance and walking, and was feasible and safe. It was a proof-of-concept study with 14 randomised and 13 analysed. That is a feasibility signal, not an efficacy result. (Dudziec 2024, PMID 38156498)

8. Proximal hip flexor strengthening

Sixteen weeks of home resistance training with 93% adherence produced a significant hip flexor strength gain on the left side only, and no change in walking speed or endurance. The authors concluded their own protocol "may not be optimal". The rationale is sound and the dose is not established. (Ramdharry 2014, PMID 25582960)

9. Telecoaching and remote delivery

Improvements across strength, cardiovascular capacity, function, gait and fatigue, from 7 of 382 screened studies, moderate quality and small samples. Promising as an answer to access, not established as an answer to efficacy. (Leale 2024, PMID 38784904)

What doesn't work

Night splinting to gain ankle range. Tested properly: +1 degree (95% CI −3 to 4; p=0.72). A Cochrane review found no benefit from any intervention for this endpoint. This is a negative result, not an unstudied one. Stretch if it feels good; do not expect it to change the ankle. (PMID 16942454; PMID 20166090)
Episodic intensive rehabilitation as a maintenance plan. Three weeks of inpatient rehab, 2-4 hours a day, improved every measured outcome and had lost all of it by 12 months. Useful for a crisis. Not a plan. (PMID 37552411)
Withholding exercise for fear of overwork weakness. Tested three ways, not confirmed, and the belief itself is a documented barrier to activity in this population.
Bracing a curve here with the expectations of an ordinary adolescent curve. It succeeded in 3 of 16. (PMID 17606789)
Treating the foot and never examining the hip or the spine.

Exercise Prescription

Only the first exercise below has been tested in this condition, and it was tested in children. The rest are borrowed from general foot, ankle and balance rehabilitation and are here because they are safe and sensible, not because research has proved them for this. The well-supported parts of the plan are the diagnosis, the hip and spine screen, footwear and bracing, falls-proofing, and staying active without fear.

Exercise	How	Sets × Reps	Frequency	Pain guide
Seated ankle lifts with a weighted cuff (the tested one)	Leg out in front, weighted cuff around the foot. Pull the foot up towards the shin slowly, lower slowly. Start near half your one-rep max and build to about two thirds over months, re-checking the max as you improve.	3 × 10-12 each foot	3 days/week, never two days in a row	Effort in the front of the shin expected. Sharp pain is not
Toe raises holding a support	Hold a worktop. Lift toes and forefoot off the floor with heels down, lower slowly.	3 × 10	3 days/week	Work, not pain
Seated heel raises	Sitting, press the balls of the feet down and lift the heels, lower slowly.	3 × 12	3 days/week	Mild calf effort
Toe spreading and towel scrunches	Bare feet. Spread the toes, then scrunch a towel towards you.	3 × 10	Most days	No pain expected
Supported balance holds	Beside a worktop with both hands ready to grab it. Feet together, then narrow, then one foot in front of the other. Hold 20-30 seconds.	3 holds per position	Most days	No pain. Always beside something solid
Sit-to-stand from a chair	Stand without using the hands if possible, sit down under control.	3 × 8-10	3 days/week	Leg effort expected
Walking on varied but safe ground	Grass, gentle slopes, firm paths. Build up gradually.	20-30 minutes	Most days	Tiredness expected

Two standard exercises are deliberately left out, and here is why

Unsupported single-leg balancing, and wobble-board or foam-pad work. This is the most standard progression in all of ankle rehabilitation and it is withheld here on purpose. The defining problem in this condition is a foot that does not clear the ground and an ankle that gives way, and the falls numbers are not theoretical: 86% report falling or nearly falling, and 38% of brace users had fallen within the past week. Taking your hands off the support is not a small step in that context. Progress to it with a therapist watching, not alone at home.

Heel walking. The obvious exercise for a foot that will not lift, and it puts you on the least stable part of a foot that already trips you.

Two different diseases share this surname. This page is about Charcot-Marie-Tooth disease, an inherited neuropathy. It is not about Charcot neuroarthropathy, the diabetic destructive foot. Different cause, different population, different urgency.

Red Flags

Most of this condition is slow and manageable. These are the things that are not, and they are worth reading before anything else on this page.

Dark cinematic anatomical study of the lower leg and peripheral nerve

Get seen — do not wait

Weakness getting worse quickly, or on one side only. This condition is slow and even on both sides. Anything else is a different problem, and some of those are treatable.
New numbness spreading upwards, or any change in bladder or bowel control. Same day.
A sore, blister or hard skin breaking down on the foot, or a foot that goes red, hot and swollen.
New back pain, or an uneven back or shoulders, especially in a child or teenager. Around one in seven develop a curve, and in this condition bracing usually does not hold it.
Hip pain or a limp, especially in a child. Hip dysplasia affects up to 28% of children with this condition and a hip that is never imaged is never found.
Falls that are getting more frequent, or a fall that caused an injury.
If you are ever offered chemotherapy, tell that team you have this condition. Two specific drugs can affect the nerves more in people who have it. This is not a reason to refuse cancer treatment, and the researchers who found it say so plainly. It is a reason for them to know and to watch.

Who to see: a neurologist for the diagnosis and for anything asymmetric or fast-moving. A foot and ankle surgeon early, not at crisis point. A spine specialist for any curve. A paediatric orthopaedic team for hip screening in a child. Emergency care for a sensory level, bladder or bowel change, or new breathlessness.

Return to Training

This condition does not resolve, so these are fitness-to-progress criteria rather than discharge criteria. The plan that survives a decade beats the one that is optimal for eight weeks.

Diagnosis confirmed, or a neurology referral made
Hips (in a child) and spine examined, and imaged where indicated
Footwear and orthotic needs assessed, and a brace trialled if there is toe catch or tripping
Home falls-proofing completed, since most falls happen at home
Balance confidence recorded at baseline, including stairs, slopes and uneven ground
Dorsiflexion strength recorded as an age and sex normed Z score
The person can state the honest expectation in their own words: slowed decline judged at 24 months, not strength gained in 6 weeks
The person can name their own red flags, especially asymmetric or rapidly progressive weakness
No unsupported balance or unstable-surface work until cleared under supervision

For lifters specifically: the bar path is unaffected. The problem is the foot and the floor. Flat stable shoes, no barefoot or minimalist lifting, work from a rack with safeties, and treat a rolled ankle as a serious event, because it will not be the last one.

Right now, take your shoes and socks off and put both feet side by side. Are both arches high? Do both ankles give way? A sprain happens to one ankle on one day. This happens to both, for years. If both feet match, book an appointment and ask specifically for your ankle reflexes to be tested.

Weak ankles for years, and both sides? Get the nerves checked, not just the ankle.

Think of the nerves running to your feet as electrical cables, and the ones reaching furthest down the body are the ones that fray first. As the cable to the muscle that lifts your foot goes bad, that muscle fades, while the muscle pulling your foot down keeps its supply and keeps pulling. Nothing is wrong with the bones or the ligaments. The foot is being pulled into a high arch and a curled shape by a tug-of-war it is losing on one side, and that is also why the ankle keeps rolling: the muscle that should catch it never gets the message in time.

What this actually is: an inherited condition of the nerves, not an injury, and it is the most common inherited neurological disorder there is.
What most people get wrong: it gets treated as a run of bad luck with sprains, because nobody compares the two feet or tests the reflexes.
Start here: get the diagnosis confirmed, then get your hips and your back checked too, because those are the parts everyone misses while staring at the feet.

Best for

Anyone with years of ankle sprains on both sides, high arches, curled toes, or a foot that catches when they walk. And anyone with a family member whose feet look the same.

Skip if

Your problem is one ankle with a date it started and a normal-looking other foot. Or your weakness is getting worse quickly or on one side only, in which case you need seen rather than read.

Want the full evidence, the anatomy and the debate? Keep scrolling. ⌄

Conviction

MODERATE MODERATE — and it splits hard by claim, so the split is worth more than the headline.

HIGH The musculoskeletal picture (bilateral, high arches, clawed toes, foot drop, near-universal instability). That the diagnosis is clinical. That falls are frequent.

MODERATE Dorsiflexor training slowing decline in children with CMT1A. That exercise does not cause overwork weakness. That hips and spine are part of the picture. That night splinting does not work.

LOW That braces measurably improve gait or balance on objective testing. Note this means the evidence cannot show it, not that braces do not help.

NO EVIDENCE Any specific adult exercise protocol. That balance training reduces falls. Any adult rate of progression. Surgery versus conservative care. The sensitivity or specificity of any clinical sign.

NO EVIDENCE and LOW are different claims. LOW means somebody looked and the result was weak. NO EVIDENCE means nobody has done the study. Collapsing them into one grade is how a field's blank spaces get mistaken for its weak findings.

What would change my mind on the exercise recommendation

A multicentre randomised trial in adults with genetically confirmed CMT1A, at least 200 participants, comparing continuous home-based progressive resistance training with quarterly review, against annual three-week intensive blocks, against usual care, over 36 months, with falls per person-year as the primary endpoint and MRI muscle and fat volume as the safety endpoint. If continuous home training showed no advantage over usual care on falls at 36 months, the recommendation here should be withdrawn for adults and kept only for the children it was earned in.

What would change my mind on the assessment section

A diagnostic accuracy study of the bedside triad (bilateral symmetry, absent ankle reflexes, high arch with clawing) against genetic confirmation, in consecutive patients presenting to musculoskeletal services with recurrent ankle sprains. No sensitivity or specificity exists for any clinical sign in this condition anywhere, which is why this page has no test table and refuses to invent one.

Tired of guessing whether it's the joint or the nerve?

The Verdict sends one evidence-checked protocol a week, written the way this page is: what the studies actually found, what they didn't, and what to do on Monday.

Join free

The Full Picture — Anatomy, Diagnosis & Evidence

What's Actually Going On

Dark cinematic anatomical study of peripheral nerve and lower leg musculature

Charcot-Marie-Tooth disease is a family of inherited peripheral neuropathies. The demyelinating forms, of which CMT1A from a duplication of the PMP22 gene is much the commonest, damage the insulating sheath and slow conduction. The axonal forms reduce the number of working fibres. Either way, the longest nerves fail first, which is why the picture is length-dependent and, crucially, bilateral. The feet go before the hands, and the hands follow years later.

The deformity is a muscle balance problem, not a bone problem. Tibialis anterior and peroneus brevis weaken while peroneus longus and tibialis posterior are relatively spared. Peroneus longus, now unopposed, drives the first ray down. Tibialis posterior, now unopposed, pulls the hindfoot into varus. The small muscles inside the foot fail, so the toes claw. That is the whole triad: high rigid arch, clawed toes, hindfoot varus.

The pressure signature has been measured and it is consistent. Across 6 studies and 146 patients using two different technologies: marked midfoot off-loading with overload at the forefoot and rearfoot, a lateralised centre-of-pressure path, and prolonged pressure-time exposure. The foot is a tripod that has lost its middle and is loading its edges, which is exactly where the callus builds.

Then come the compensations. Foot drop forces the hip to do the lifting, and hip flexor fatigue becomes the thing that limits walking distance. By the time someone reaches a clinic complaining of hip or calf fatigue, the primary problem is two joints away. And the sensory side matters as much as the motor side: proprioceptive loss is what turns a weak ankle into an unstable one, which is why strengthening alone never fixes the giving way.

Prevalence: pooled at 17.69 per 100,000 (95% CI 12.32-24.33) across 31 studies, with CMT1 at 10.61 per 100,000. That makes it the most common inherited neurological disorder.

How to Identify It

Dark cinematic study of a foot and ankle examination under directional light

There is no validated special test for this condition, and that is a finding rather than a gap in this page. Across a 1,183-record literature search including two clinical practice guidelines, no published sensitivity, specificity or likelihood ratio exists for any clinical sign. The 2025 guideline's position is that the diagnosis is clinical, and that routine follow-up needs no ancillary testing at all.

One number is very easy to misuse here, so it is worth naming. The validated instability questionnaire's paper says the tool "demonstrated excellent sensitivity". That is sensitivity to discriminate between groups, not diagnostic sensitivity, and there is no specificity to pair with it. Its real numbers are group comparisons χ² = 15.10 | p < .001 against children who had genuinely sprained their ankles, and χ² = 33.69 | p < .001 against controls, with test-retest reliability ICC 0.73. Useful for measuring. Not a diagnostic accuracy statistic.

What actually discriminates, all of it free:

Is it both sides? The highest-yield question on the page. 29 of 30 children and adolescents tested reported moderate to severe bilateral instability, correlated with the high-arched foot shape at r = 0.69, p < .001.
Are the ankle reflexes absent? Absent ankle jerks with a mechanical story should change the whole consultation.
"When did your feet start looking like that?" The revealing answer is "they've always been like that". There is no date of onset because the person assumed everybody was like this.
Any trouble with the hands? Buttons, keys, jar lids. Feet-then-hands is a polyneuropathy pattern and is not a feature of a mechanical foot problem.
Family history of the same foot shape. Ask about high arches and foot surgery, not about a named diagnosis. Families carry this unnamed across generations.
Watch them walk barefoot. Foot slap, high-stepping, and a weight path that runs down the outside of the foot.
Look at the spine, and in a child, image the hips.

CMT is also named alongside juvenile idiopathic arthritis, osteochondritis dissecans of the talus, tarsal coalition and transitional ankle fractures as one of five conditions commonly misdiagnosed as a paediatric ankle sprain, specifically in the child who presents repeatedly. That paper is a narrative review with no accuracy data, so it establishes that the miss is recognised, not how often it happens.

The Debate

Two clinical practice guidelines exist and both are recent, so this is not the usual case of a stale guideline versus a newer trial. What is worth flagging is what the guideline admits about itself: the paediatric guideline produced 3 evidence-based and 31 consensus-based recommendations across 10 clinical questions, and could not reach consensus in some areas at all. It sits at the top of the evidence hierarchy and is honest that most of it is expert opinion.

Does strengthening make you stronger, or just slow the decline?

What everyone quotes (Burns 2017, 24-month result)

Dorsiflexion strength difference 0.6 (95% CI 0.03 to 1.12; p=0.041) in favour of training, cited as evidence that strengthening works.

What the same trial found first (6-month primary endpoint)

Difference of exactly zero (95% CI −0.37 to 0.42; p=0.91) at the end of the training block, and still nothing at 12 months (p=0.27).

Not a contradiction, a mechanism. In a progressive disease the control group keeps declining, so the intervention separates later. The endpoint is slowed progression, not acquired strength, and the systematic review words it exactly that way. Reading the 24-month number alone imports a promise the 6-month number specifically refuses.

Does exercise cause overwork weakness?

The field's founding orthodoxy

Exercise overworks already-denervated muscle and accelerates its loss. This withheld activity from this population for decades and still shapes what patients believe.

Three independent tests

MRI muscle and intramuscular fat volume unchanged against sham at 6 and 24 months. Dominant versus non-dominant grip strength not different across 5 studies and 166 people, which is a lifetime of asymmetric loading. Treadmill training produced no overwork weakness and no deterioration.

The fear has been tested three ways and not confirmed. It has not been refuted at every dose in every subtype, so this is MODERATE rather than HIGH, but the burden of proof has moved and the default should move with it.

An older review sounds more confident than a newer one. Which wins?

Corrado 2016, 11 studies

"Strength or endurance trainings improve functionality and ADLs of affected patients."

Conde 2023, RCTs only, with GRADE

Could "only recommend exercise to improve muscle strength in children", with 6 of 6 trials at high risk of bias and every non-strength outcome graded very low to low.

The more confident-sounding statement rests on the weaker evidence, because it admitted non-randomised designs and applied no formal appraisal. Newer is not automatically more permissive, and here it is considerably less.

Braces: users say yes, the meta-analysis can't say anything

Anderson 2025, 306 brace users

78% report their brace improves their balance.

Kim 2024, 15 studies pooled

Every pooled gait effect small-to-moderate and non-significant; balance outcomes could not be pooled at all.

A measurement gap, not a refutation. Samples as small as one, wildly different brace designs, and a laboratory gait parameter is not the same construct as feeling steady on a real pavement. Keep prescribing and customising; be honest about what the literature can and cannot show.

Honest Limitations

1. The evidence is paediatric. The patient usually is not.

The research: the only moderate-certainty recommendation in the entire field applies to children aged 6 to 17 with CMT1A, from one trial of 60 children.

The gap: the adult evidence is a pilot with a one-sided result, a 13-participant proof of concept, a single-blind trial reporting responder counts rather than between-group effects, a 2004 study of 20 people, and an uncontrolled retrospective cohort. The 2025 guideline instructs "individualised exercise guidelines" started early, with no adult trial base underneath it.

The adjustment: use the paediatric dose as the best-specified starting template for adults, say out loud that it is extrapolated, and measure the individual rather than trusting the protocol.

2. Everything durable points away from how services are actually bought.

The research: three weeks of intensive inpatient rehabilitation improved every measured outcome and lost all of it by 12 months. The sham-controlled trial's benefit appeared eighteen months after training stopped.

The gap: health systems purchase episodic blocks. The evidence argues for continuous low-dose self-management with periodic review, which is the model with the least funding and the thinnest evidence base. The telecoaching review is interesting for exactly this reason and rests on 7 of 382 screened studies.

The adjustment: design for the next twenty years, not the next six weeks. A small daily thing somebody actually keeps doing beats a block they finish and stop.

3. Adherence here is governed by belief, not by protocol design.

The research: in four focus groups with 10 patients and 12 physical therapists, prior experience determined whether people engaged with activity at all, and reassurance that occasional pain and fatigue are normal was a facilitator. A shortage of clinicians who know this condition restricted opportunity.

The gap: a decades-old overwork-weakness warning is still suppressing activity long after the trials stopped finding harm.

The adjustment: spend the first appointment dismantling the fear, explicitly. It is probably worth more than the exercise selection.

The Nuance

Dark cinematic anatomical study contrasting nerve and joint structures of the lower limb

This is not a foot disease, and that is the part clinicians most reliably miss.

Hips. In 178 children, 64 (35.8%) had hip dysplasia on their first pelvic radiograph, 20 of those normalised over time, 50 (28.1%) still had it at last follow-up, and 17 (9.6%) required surgery. Rates varied sharply by subtype. And when it is found here it is worse than ordinary developmental hip dysplasia on every measured axis, including significantly more arthritic hips (p=0.006), in a matched comparison. A hip that is never imaged is never found.

Spine. Of 298 patients reviewed, 45 had scoliosis, detected at a mean age of 12.9 years with a mean curve of 27.6 degrees. One-third were left thoracic and 49% had increased thoracic kyphosis, both unlike ordinary adolescent curves. 24 of 34 followed curves progressed, and bracing succeeded in only 3 of 16. One more detail matters for surgical planning: intraoperative nerve monitoring was possible in only 3 of 12 attempts, because the peripheral neuropathy degrades the signal.

Surgery versus conservative care: there is no comparison to report. No trial anywhere compares them for the CMT foot, and no success rate exists for either path. What does exist is corrective surgery demonstrably redistributing plantar load, with residual lateral overload usually persisting. The governing document is a consensus statement graded Level V, expert opinion, and its most useful recommendation is a process one: get a foot and ankle surgeon into the care team early, not when the foot is already rigid. Anyone quoting a success rate for either path is quoting something nobody has measured.

One drug caution, stated carefully. A systematic review found evidence that vincristine, and possibly paclitaxel, can occasionally cause a more severe course of nerve toxicity in people with this condition, and found no convincing evidence for any other drug. The same review is explicit that patients with CMT must not be denied effective treatment that may prolong life. The action is to tell the treating team and monitor, never to refuse.

Sources

Sivera Mascaró R, et al. (2025). Clinical practice guidelines for the diagnosis and management of Charcot-Marie-Tooth disease. Neurologia. PMID 38431252. Clinical practice guideline.
Yiu EM, et al. (2022). Clinical practice guideline for the management of paediatric Charcot-Marie-Tooth disease. J Neurol Neurosurg Psychiatry. PMID 35140138. World-first paediatric guideline; 3 evidence-based and 31 consensus-based recommendations.
Burns J, et al. (2017). Safety and efficacy of progressive resistance exercise for Charcot-Marie-Tooth disease in children. Lancet Child Adolesc Health. PMID 30169201. N=60, double-blind sham-controlled; null at 6 months (p=0.91), significant at 24 months (p=0.041).
Conde RM, et al. (2023). Effectiveness of exercise therapy for individuals diagnosed with Charcot-Marie-Tooth disease. J Peripher Nerv Syst. PMID 37060329. Systematic review of RCTs, pooled N=214, 6 of 6 at high risk of bias.
Refshauge KM, et al. (2006). Night splinting does not increase ankle range of motion. Aust J Physiother. PMID 16942454. N=14 crossover; +1° (95% CI −3 to 4; p=0.72).
Rose KJ, et al. (2010). Interventions for increasing ankle range of motion in neuromuscular disease. Cochrane Database Syst Rev. PMID 20166090. 4 studies, 149 participants; no benefit found for this endpoint.
Kim A, et al. (2024). The effect of ankle-foot orthoses on gait characteristics. J Foot Ankle Res. PMID 39276325. 15 studies; pooled effects small-to-moderate and non-significant.
Rose KJ, et al. (2015). Correlates of functional ankle instability in children and adolescents. J Foot Ankle Res. PMID 26543504. 29 of 30 with bilateral instability; r=0.69, p<.001 with cavus structure.
Mandarakas M, et al. (2013). Measuring ankle instability in pediatric Charcot-Marie-Tooth disease. J Child Neurol. PMID 23696628. 104 children; ICC 0.73.
Ramdharry GM, et al. (2018). Frequency and circumstances of falls. Physiother Res Int. PMID 29282812. 86% of 107 reported falls or near falls.
Anderson KM, et al. (2025). Balance confidence and falls. Health Sci Rep. PMID 40264638. 306 brace users; 38% fell within the preceding week; 78% report improved balance.
Tan KA, et al. (2025). Hip dysplasia in Charcot-Marie-Tooth disease. J Peripher Nerv Syst. PMID 39887493. 178 children; 9.6% to 28.1% prevalence; 9.6% required surgery.
Novais EN, et al. (2014). Hip dysplasia is more severe in CMT than in developmental dysplasia of the hip. Clin Orthop Relat Res. PMID 23943527. Matched comparison; more arthritic hips (p=0.006).
Karol LA, Elerson E (2007). Scoliosis in patients with Charcot-Marie-Tooth disease. J Bone Joint Surg Am. PMID 17606789. 45 of 298; bracing succeeded in 3 of 16.
Ferraro F, et al. (2024). Effects of intensive rehabilitation on functioning. Neurol Sci. PMID 37552411. N=37; all gains lost at 12 months.
Cavaletti G, et al. (2023). Toxic medications in Charcot-Marie-Tooth patients. J Peripher Nerv Syst. PMID 37249082. Vincristine and possibly paclitaxel; no convincing evidence for other drugs.
Roberts-Clarke D, et al. (2016). Examining hand dominance as evidence for overwork weakness. Int J Rehabil Res. PMID 27177353. 5 studies, 166 participants; no dominant/non-dominant difference.
Ma M, et al. (2023). A meta-analysis on the prevalence of Charcot-Marie-Tooth disease. J Neurol. PMID 36631678. 31 studies; 17.69/100,000 (95% CI 12.32-24.33).
Mori L, et al. (2020). Treadmill training in patients affected by Charcot-Marie-Tooth neuropathy. Eur J Neurol. PMID 31444929. N=53; no overwork weakness.
Arceri A, et al. (2025). Plantar pressure distribution in Charcot-Marie-Tooth disease. Sensors. PMID 40732443. 6 studies, 146 patients.
Granberg AL, et al. (2026). Exploring the experiences of physical activity in people with Charcot-Marie-Tooth disease. Disabil Rehabil. PMID 42057752. 10 patients, 12 physical therapists.

Educational self-management guidance, not personalized treatment. If you have this condition or think you might, the first step is a proper diagnosis, not a protocol from the internet.