A 62-year-old retired bank clerk has become unsteady over three years. He holds the wall in the bathroom at night. His feet feel numb, and his wife says he has had a dry cough for as long as she can remember. His brain MRI shows mild cerebellar atrophy and nothing else. This is an illustrative case, not a real patient.
The Ataxia Diagnostic Pathway walks you through his examination, his tempo and the tests to send. It no longer carries the anatomy lesson, the Indian series or the trial figures behind each step. Those used to sit in its Examination and Evidence tabs, and they have moved here so that the tool stays quick at the bedside. By the end of this article you will know why the examination comes before the scan, why the order of genetic tests is fixed, and why “cerebellar atrophy” is never the end of the work-up.
Use this at the bedside
Open the tool: Ataxia Diagnostic Pathway
Get the CDSS app: install the MEDiscuss CDSS on your phone
1. Three systems, one sign
Ataxia is incoordination with preserved power. It comes from the cerebellum, from the sensory pathways that tell the cerebellum where the limbs are, or from the vestibular system that tells it where the head is. Within the cerebellum the midline vermis governs the trunk and gait, each hemisphere governs the limbs on its own side, and the flocculonodular lobe governs eye movements and balance. The double crossing of the cerebellar outflow is why a hemispheric lesion produces signs on the same side.
“Then why does my patient get worse in the dark?” Because standing needs two of three inputs: vision, joint position sense and the vestibular system. Take away vision, and a man whose feet no longer report where they are has only one input left. The table below is the one the tool keeps in two halves. Here it is whole.
| Finding | Cerebellar ataxia | Sensory ataxia | Vestibular ataxia |
|---|---|---|---|
| Effect of closing the eyes | Little change; unsteady with eyes open too | Much worse: Romberg positive; worse in the dark and in the bathroom | Somewhat worse; sways toward the affected side |
| Vibration and joint position sense | Normal | Reduced or lost at the toes; pseudoathetosis of the outstretched fingers | Normal |
| Ankle reflexes | Present, may be pendular | Absent in neuropathy and neuronopathy; present in a posterior column lesion of the cord | Present |
| Nystagmus | Gaze-evoked, direction-changing; downbeat in midline lesions | None | Horizontal-torsional, beats one way in every gaze, suppressed by fixation |
| Speech | Scanning, slurred, explosive | Normal | Normal |
| Limb coordination | Dysmetria, intention tremor, dysdiadochokinesia; rebound | Clumsy only when vision is removed; heel-shin worse with eyes closed | Normal; past-pointing toward the lesion at most |
| Gait | Wide-based, irregular, lurching; cannot tandem | High-stepping, stamping, watches the feet | Veers to one side; vertigo, nausea |
| Company | Hypotonia, pendular reflexes, titubation | Distal sensory loss, Lhermitte sign in a cord lesion, anaemia in B12 deficiency | Vertigo, oscillopsia, hearing loss, tinnitus |
Our patient is worse in the dark and has numb feet: a sensory ataxia. He also has cerebellar atrophy. He may have both, and that combination is a clue in itself.
2. The treatable ataxias
Every one of these is missed by an assessment that stops at “cerebellar atrophy on MRI”. The tool keeps this table in two halves, the clue and the test in one and the treatment in the other. Here it is whole.
| Cause | Clue | Test | Treatment |
|---|---|---|---|
| Wernicke encephalopathy | Alcohol, vomiting, malnutrition, bariatric surgery; ophthalmoplegia, confusion | Clinical: 2 of the 4 Caine criteria. MRI supports, never delays | Thiamine intravenously before glucose (dose in the acute lane) |
| Hypothyroidism | Cold intolerance, weight gain, slow reflexes; often no clue | TSH, free T4 | Levothyroxine; the ataxia usually improves over months |
| Vitamin B12 deficiency | Sensory ataxia, macrocytosis, vegetarian diet, metformin, proton pump inhibitor | B12, methylmalonic acid or homocysteine; look for cord signal change | Parenteral B12 replacement |
| Copper deficiency | Sensory ataxia like B12 deficiency with a normal B12; zinc excess, bariatric surgery, denture cream | Serum copper, ceruloplasmin, zinc | Copper replacement, stop the zinc; deterioration halts, recovery is partial |
| Vitamin E deficiency | Friedreich-like picture with retinopathy; fat malabsorption; the recessive AVED | Vitamin E, adjusted for lipids | High-dose vitamin E, up to 1500 mg a day in AVED |
| Gluten ataxia | Gait-predominant sporadic ataxia; gut symptoms in only a minority | Anti-gliadin antibodies, tissue transglutaminase, HLA-DQ2 and DQ8; duodenal biopsy | Strict gluten-free diet with a dietitian; judge at 1 year |
| Anti-GAD and primary autoimmune cerebellar ataxia | Type 1 diabetes, thyroiditis, stiffness; subacute, vermis-predominant, CSF oligoclonal bands | Anti-GAD65 titre, CSF, the autoimmune screen | Immunotherapy: corticosteroids, immunoglobulin, then a steroid-sparing agent |
| Paraneoplastic cerebellar degeneration | Rapid, severe, bilateral; smoker, breast or ovarian risk; opsoclonus | Onconeural antibodies; CT chest, abdomen and pelvis, mammography, pelvic ultrasound, PET-CT | Find and treat the tumour; immunotherapy; stop a checkpoint inhibitor |
| Drug and alcohol | Phenytoin, lithium, carbamazepine, metronidazole, fluorouracil, amiodarone; alcohol | The chart and a level | Stop or reduce; thiamine and abstinence for alcohol |
| Cerebrotendinous xanthomatosis | Tendon xanthomas, juvenile cataracts, childhood diarrhoea, cognitive decline | Plasma cholestanol; CYP27A1 | Chenodeoxycholic acid |
| Niemann-Pick type C | Vertical supranuclear gaze palsy, splenomegaly, psychiatric onset, dystonia | Plasma oxysterols; NPC1 and NPC2 | Miglustat, in a specialist centre |
| Refsum disease | Retinitis pigmentosa, anosmia, deafness, ichthyosis, neuropathy | Plasma phytanic acid | Dietary phytanic acid restriction; plasma exchange in crisis |
| Abetalipoproteinaemia | Fat malabsorption from infancy, acanthocytes, retinitis, very low cholesterol | Lipid profile, blood film, vitamin E | Vitamin E and A, low-fat diet |
| Coenzyme Q10 deficiency | Childhood-onset ataxia with seizures, myopathy or nephropathy | Muscle CoQ10; genetic testing | High-dose coenzyme Q10 |
| Wilson disease | Onset below 40 with tremor, dystonia, liver disease or psychiatric change | Ceruloplasmin, 24-hour urinary copper, slit-lamp examination | Chelation and zinc, lifelong |
| Episodic ataxia type 2 | Attacks of minutes to hours, nystagmus between attacks, onset in youth | CACNA1A | Acetazolamide; 4-aminopyridine 5 mg three times a day |
| Friedreich ataxia | Areflexia with extensor plantars, sensory loss, scoliosis, cardiomyopathy, diabetes | FXN GAA repeat | Omaveloxolone from age 16 slows progression; not yet registered in India |
Gluten ataxia: the figure behind it
“Why test for gluten sensitivity in a man with no gut symptoms?” Because in the Sheffield series anti-gliadin antibodies were found in 41 per cent of patients with sporadic idiopathic ataxia against 12 per cent of controls, with an enteropathy on biopsy in about one in four of the positives (Hadjivassiliou 2003). The gut is usually silent.
3. The genetic ataxias, and the Indian numbers
A repeat-expansion test is cheap and fast, and exome sequencing does not see a repeat expansion at all. So the order is fixed: the family tree, then the repeat panel that fits it, then sequencing for what remains. The Indian frequencies below come from three regional series, and they differ from each other, which is itself the lesson: order by the local pattern, and send all four common dominant expansions together.
| Pattern | Send first | Phenotype clues | Indian data |
|---|---|---|---|
| Dominant | SCA1, SCA2, SCA3, SCA12 together; then SCA6, SCA7, SCA17, DRPLA, SCA8, SCA10; then FGF14 (SCA27B) in late onset | Slow saccades: SCA2. Retinopathy: SCA7. Action tremor at onset: SCA12. Chorea or dementia: SCA17, DRPLA, Huntington. Myoclonic epilepsy: DRPLA. Bulging eyes, spasticity, dystonia: SCA3. Pure late cerebellar: SCA6, SCA27B | Eastern India: SCA2 in 57 per cent of 28 dominant families, SCA1 in 14 per cent, no SCA3 (Sinha 2004). Southern India, 318 probands: SCA1 commoner than SCA2, then SCA3 (Krishna 2007). Delhi, 95 genetically proven: SCA12 the commonest at 39 per cent, four in five from the Agrawal community (Radhakrishnan 2018) |
| Recessive, or sporadic with onset below 25 | FXN (Friedreich) first; alpha-fetoprotein, albumin, cholesterol, immunoglobulins, vitamin E, cholestanol, phytanic acid, oxysterols, ceruloplasmin; then a recessive ataxia panel or exome (ATM, APTX, SETX, SACS, SPG7, SYNE1, POLG, ANO10) | Telangiectasia, infections, oculomotor apraxia, raised alpha-fetoprotein: ataxia-telangiectasia. Oculomotor apraxia with low albumin: AOA1; with raised alpha-fetoprotein at 10 to 20: AOA2. Spasticity: SPG7, ARSACS. Ophthalmoplegia, neuropathy, epilepsy: POLG | Friedreich ataxia is less frequent in India than in Europe, but is present and is the first recessive test. Consanguinity raises the yield of the whole recessive list |
| Sporadic, onset after 35 | RFC1 (biallelic AAGGG); the dominant panel anyway, because a parent may have died young or carried a reduced-penetrance allele; FMR1 CGG in men with tremor; FGF14 GAA; and the full acquired screen from the subacute lane | Sensory neuronopathy, bilateral vestibular failure, chronic cough: RFC1 (22 per cent of late sporadic ataxia; 63 per cent with neuronopathy; 92 per cent with the full triad). Autonomic failure and parkinsonism within 3 years: multiple system atrophy. Tremor, parkinsonism, MCP sign, grandchildren with fragile X: FXTAS | FGF14 GAA expansion in 10 per cent of 31 Indian late-onset index patients (Pellerin 2023). FXTAS in 3.3 per cent of an Indian late-onset ataxia and tremor cohort (Faruq 2014) |
| X-linked | FMR1 CGG repeat (55 to 200 is the premutation); very long chain fatty acids if spastic | Men over 50 with intention tremor and ataxia: FXTAS. Spastic paraparesis with adrenal failure: adrenomyeloneuropathy | The daughters of a FXTAS patient carry the premutation and are at risk of early menopause; the grandsons are at risk of fragile X syndrome |
The three Indian series
| Series | What it found |
|---|---|
| Eastern India, 28 dominant families | SCA2 in 16 and SCA1 in 4, no SCA3 at all, with a founder haplotype behind the SCA2 |
| Southern India, 318 probands at a national institute | SCA1 commoner than SCA2, then SCA3. One in five of the positives had no family history |
| Delhi, 95 genetically proven patients | SCA12 commonest at 39 per cent, four in five from the Agrawal community, an action tremor of the arms years before the gait went. The expanded allele carries 55 to 69 repeats against a normal range of 7 to 31 |
“Which one do I send first, then?” All four common dominant expansions together, SCA1, SCA2, SCA3 and SCA12, because the commonest one depends on where the family comes from (Sinha 2004; Krishna 2007; Radhakrishnan 2018). The tool keeps that instruction; the series behind it are here.
Our patient has no family history and an onset after 35. His triad is a sensory neuronopathy, a vestibular problem that makes the dark so difficult, and a cough that began years before the unsteadiness. That is the picture of RFC1 disease, found in 92 per cent of those with the full triad (Cortese 2019).
4. The SARA: where the scale came from
The Scale for the Assessment and Rating of Ataxia was built in 2006 in 167 and then 119 patients with spinocerebellar ataxia, takes about 14 minutes, and has an inter-rater agreement of 0.98 (Schmitz-Hubsch 2006). It has no bands for mild or severe. Its job is to describe one patient over time, which is why the tool asks for it at every visit.
5. Drugs that slow the disease, and drugs that did not arrive
Omaveloxolone is the first drug shown to slow Friedreich ataxia (MOXIe). The propensity-matched extension suggested about 3.6 points less progression over 3 years. The tool keeps its dose, its age limit and its status in India.
Two others did not arrive. Troriluzole for the spinocerebellar ataxias received a complete response letter from the United States regulator in November 2025 and was not approved when this was written; vatiquinone for Friedreich ataxia received one in August 2025. So when a family asks about a new drug they have read about, check its status before you raise their hopes.
6. The acutely ataxic child: what the series showed
In a referral centre series of 120 children, post-infectious cerebellar ataxia was the commonest cause (Thakkar 2016). 85 per cent of those children were aged 1 to 6, most after a viral illness and classically after varicella. CSF showed a pleocytosis in 40 per cent of those tested, MRI was normal, and 91 per cent recovered within 30 days.
An earlier 40-case series: post-infectious ataxia, ingestion and Guillain-Barre syndrome made up 80 per cent, a suggestive drug history was present in about six in ten ingestions, and imaging was positive in only 2 of 26 (Gieron-Korthals 1994). That is why the tool asks for a toxicology screen in every child.
Back to the bank clerk
His examination gives the address: a sensory ataxia with cerebellar signs, worse in the dark. His tempo, three years, puts him in the chronic lane. The treatable list is run in full anyway: B12, copper, thyroid, vitamin E and the gluten antibodies, because “cerebellar atrophy on MRI” is not a diagnosis. With no family history and an onset after 35, the tool suggests RFC1 first, and his cough makes that the likeliest answer. If it is negative, the dominant panel is still sent, because a parent may have died young. Each of those steps is a line in the tool, and now you know what each line stands on.
Try this
A 4-year-old girl comes in a week after chickenpox. She is alert and afebrile, and walks with a wide, unsteady gait. The rest of her examination is normal. Do you scan her tonight? Illustrative case.
Answer. Not on these findings. This is the picture of post-infectious cerebellar ataxia, and in the larger series 91 per cent of such children recovered within 30 days. Observe her, and send a toxicology screen, because ingestion was among the commonest causes in both series. Image first if she becomes drowsy or febrile, develops headache with vomiting, papilloedema, focal signs, weakness, absent reflexes or opsoclonus, has a history of trauma, or is still worsening after 3 weeks.
Take-home points
- Examine before you scan. The examination gives the address: cerebellar, sensory or vestibular.
- Ask about the tempo. Hours, weeks, years and attacks each have their own short list.
- Run the treatable list in every chronic sporadic ataxia, even when the MRI shows atrophy.
- Send the four common dominant expansions together. Exome sequencing does not see a repeat expansion.
- In a late-onset sporadic ataxia with numb feet, trouble in the dark and a chronic cough, think of RFC1.
Use this at the bedside
Open the tool: Ataxia Diagnostic Pathway
Get the CDSS app: install the MEDiscuss CDSS on your phone
Declaration: the author built and maintains the tool. It is free to use. The figures in this article are those the tool carried in its Examination and Evidence tabs on 3 November 2026, each from the source named beside it.
References
- Schmitz-Hubsch T, du Montcel ST, Baliko L, et al. Scale for the assessment and rating of ataxia: development of a new clinical scale. Neurology. 2006;66(11):1717-1720.
- Lynch DR, Chin MP, Delatycki MB, et al. Safety and efficacy of omaveloxolone in Friedreich ataxia (MOXIe study). Ann Neurol. 2021;89(2):212-225.
- de Silva RN, Vallortigara J, Greenfield J, Hunt B, Giunti P, Hadjivassiliou M. Diagnosis and management of progressive ataxia in adults. Pract Neurol. 2019;19(3):196-207.
- Cortese A, Simone R, Sullivan R, et al. Biallelic expansion of an intronic repeat in RFC1 is a common cause of late-onset ataxia. Nat Genet. 2019;51(4):649-658.
- Pellerin D, Danzi MC, Wilke C, et al. Deep intronic FGF14 GAA repeat expansion in late-onset cerebellar ataxia. N Engl J Med. 2023;388(2):128-141.
- Hagerman RJ, Hagerman PJ. Fragile X-associated tremor/ataxia syndrome: features, mechanisms and management. Lancet Neurol. 2013;12(8):786-798.
- Caine D, Halliday GM, Kril JJ, Harper CG. Operational criteria for the classification of chronic alcoholics: identification of Wernicke’s encephalopathy. J Neurol Neurosurg Psychiatry. 1997;62(1):51-60.
- Hadjivassiliou M, Graus F, Honnorat J, et al. Diagnostic criteria for primary autoimmune cerebellar ataxia: guidelines from an international task force on immune-mediated cerebellar ataxias. Cerebellum. 2020;19(4):605-610. The Sheffield antigliadin figures are from Hadjivassiliou M et al, Brain. 2003;126(3):685-691.
- Strupp M, Kalla R, Claassen J, et al. A randomized trial of 4-aminopyridine in EA2 and related familial episodic ataxias. Neurology. 2011;77(3):269-275.
- Thakkar K, Maricich SM, Alper G. Acute ataxia in childhood: 11-year experience at a major pediatric neurology referral center. J Child Neurol. 2016;31(9):1156-1160. The 40-case series is Gieron-Korthals MA et al, J Child Neurol. 1994;9(4):381-384.
- Sinha KK, Worth PF, Jha DK, et al. Autosomal dominant cerebellar ataxia: SCA2 is the most frequent mutation in eastern India. J Neurol Neurosurg Psychiatry. 2004;75(3):448-452.
- Krishna N, Mohan S, Yashavantha BS, et al. SCA 1, SCA 2 & SCA 3/MJD mutations in ataxia syndromes in southern India. Indian J Med Res. 2007;126(5):465-470.
- Radhakrishnan DM, Goyal V, Srivastava AK, Shukla G, Behari M. Evaluation of various movement disorders in patients of genetically proven spinocerebellar ataxia: a study from a tertiary care center in northern India. Ann Indian Acad Neurol. 2018;21(1):24-28. The repeat sizes are from Srivastava AK et al, Ann Neurol. 2001;50(6):796-800.
- Faruq M, Srivastava AK, Suroliya V, et al. Identification of FXTAS presenting with SCA 12 like phenotype in India. Parkinsonism Relat Disord. 2014;20(10):1089-1093.
