Cerebellar Ataxia

Cerebellar ataxia is a neurological sign characterized by incoordination of gait, limb movements, speech, and eye movements resulting from dysfunction of the cerebellum or its afferent/efferent pathways. It is not a single disease but a final common pathway for a diverse group of disorders that differ fundamentally in etiology, prognosis, and treatment.

Also Called / Synonyms

• Spinocerebellar ataxia (SCA) - dominantly inherited forms (e.g., SCA1, SCA2, SCA3, SCA6, SCA7) ^[1]A1c
• Multiple system atrophy of cerebellar type (MSA-C) - sporadic, rapidly progressive ^[1]A1c
• Idiopathic late-onset cerebellar ataxia (ILOCA) - sporadic, slowly progressive, no clear cause ^[19]B3b
• Sporadic adult-onset ataxia of unknown etiology (SAOA) - synonym for ILOCA ^[12]D5
• Autoimmune cerebellar ataxia (ACA) - includes primary autoimmune (PACA) and paraneoplastic forms ^{[14]B2b[18]B3b}
• Paraneoplastic cerebellar degeneration (PCD) - immune-mediated, cancer-associated ^[20]C4
• Cerebellar ataxia, neuropathy, and vestibular areflexia syndrome (CANVAS) - recessive RFC1 expansion ^[12]D5
• Cerebellar ataxia, areflexia, pes cavus , optic atrophy, sensorineural hearing loss (CAPOS) - ATP1A3-related ^[6]C4
• Gerstmann-Sträussler-Scheinker syndrome (GSS) - inherited prion disease (P102L) ^[10]D5
• Spastic ataxia - ataxia with spasticity grade ≥2 on Modified Ashworth Scale ^[13]B2b
• Hereditary cerebellar ataxia (HCA) - genetic, includes dominant, recessive, X-linked ^[13]B2b

Classification by Etiology

Cerebellar ataxia is best classified by underlying cause, as this determines workup and management. The table below organizes the major categories supported by current evidence.

Clinical Significance

Cerebellar ataxia is a common neurological sign with a broad differential. Among sporadic adult-onset cases, up to 28.9% of ILOCA patients transition to MSA-C over a mean of 5.7 years ^[19]B3b. Autoimmune causes account for a substantial proportion: in a Japanese nationwide survey, 51.1% of clinically diagnosed ACA patients had detectable autoantibodies, and 66.2% responded to immunotherapy ^[18]B3b. Early recognition of treatable etiologies, autoimmune, paraneoplastic, or genetic, is critical because delayed treatment worsens outcomes ^{[14]B2b[18]B3b}.

Pearl: When evaluating adult-onset cerebellar ataxia, the combination of autonomic failure (orthostatic hypotension ≥30 mm Hg systolic or urinary incontinence) and rapid progression (wheelchair dependence within 5 years) strongly favors MSA-C over ILOCA ^{[1]A1c[19]B3b}; conversely, subacute onset with CSF inflammation should prompt an autoimmune workup, as immunotherapy response is excellent if started early ^{[14]B2b[18]B3b}.

Building on the classification of cerebellar ataxia into hereditary, acquired, and degenerative forms, the pathophysiology of ataxia converges on disruption of cerebellar circuitry at specific neuroanatomic nodes. The cerebellum integrates motor and sensory input through a precise loop: mossy fibers and climbing fibers synapse onto Purkinje cells via granule cells, and Purkinje cells provide the sole output from the cerebellar cortex to the deep cerebellar nuclei. Lesions at any point, Purkinje cell loss, afferent pathway degeneration, or synaptic dysfunction, produce the characteristic incoordination.

Neuroanatomic Substrate

Postmortem studies in cerebellar ataxia with neuropathy and vestibular areflexia syndrome (CANVAS) reveal marked loss of Purkinje cells, predominantly in the vermis, combined with a severe dorsal root ganglionopathy and secondary tract degeneration ^[30]D5. This dual pathology, cerebellar cortical atrophy and sensory neuronopathy, explains the triad of cerebellar, vestibular, and sensory deficits. In spinocerebellar ataxia type 36 (SCA36), neuropathology shows significant cerebellar Purkinje cell degeneration with obvious loss of lower motor neurons, linking cerebellar ataxia to motor neuron disease ^[33]D5. Similarly, multiple system atrophy (MSA) involves α-synuclein accumulation in oligodendrocytes, leading to striatonigral and olivopontocerebellar degeneration ^[9]D5.

Molecular Pathways

Genetic Mechanisms

Repeat expansions are a frequent cause. In CANVAS, biallelic (AAGGG)n expansions in RFC1 (800-1,000 repeats) cause a loss-of-function mechanism: truncating variants in trans with the expansion lead to nonsense-mediated mRNA decay and reduced RFC1 transcript and protein ^[21]C4[28]D5. In SCA36, a (GGCCTG)n expansion in NOP56 causes Purkinje cell and motor neuron degeneration without intranuclear inclusions ^[33]D5. For SCA10, the (ATTCT)n expansion in ATXN10 does not alter transcription or splicing, and a simple gain or loss of function is unlikely ^[32]D5.

Ion channel and transporter defects impair neuronal excitability. Heterozygous mutations in ATP1A1, ATP1A2, and ATP1A3, encoding Na+/K+-ATPase α subunits, cause severe loss of pump function, leading to cerebellar ataxia, hemiplegic migraine, and epileptic encephalopathy ^[29]D5[37]D5. RAB3A variants (e.g., p.Arg83Trp) reduce GTP hydrolysis and fail to rescue synaptic phenotypes in Drosophila, indicating a partial loss-of-function mechanism that disrupts neurotransmitter release ^[36]D5.

Metabolic and lysosomal pathways also converge on cerebellar vulnerability. Functionally impaired COQ2 variants impair coenzyme Q10 biosynthesis; plasma CoQ10 levels are lower in MSA patients (mean 0.51 vs 0.72 µg/mL), supporting a role for mitochondrial dysfunction ^[23]D5[34]D5. Niemann-Pick disease type C (NPC1/NPC2 mutations) causes lysosomal cholesterol trafficking defects, with cerebellar ataxia in 76% of adult-onset cases ^[26]C4. Vanishing white matter disease (eIF2B mutations) triggers undue activation of the unfolded-protein response, selectively damaging oligodendrocytes and astrocytes ^[24]D5.

Motor protein and epigenetic defects add further layers. KIF1C mutations impair microtubule-based motor function, causing hereditary spastic paraplegia with cerebellar ataxia ^[31]D5. DNMT1 mutations lead to cytoplasmic translocation and aggresome formation, disrupting heterochromatin binding and causing selective neurodegeneration ^[27]C4. Bi-allelic ACBD6 variants impair N-myristoylation of proteins, leading to a neurodevelopmental syndrome with cerebellar ataxia in 85% of patients ^[38]D5.

Autoimmune Mechanisms

Autoantibodies target synaptic and cell-surface proteins, producing reversible circuit dysfunction. DPPX antibodies bind the regulatory subunit of Kv4.2 potassium channels, causing hyperekplexia, cerebellar ataxia, and stiffness ^[22]C4. Other antibodies, against NMDAR, GAD, glycine receptor, and IgLON5, induce ataxia through receptor blockade, complement activation, or T-cell-mediated cytotoxicity ^[39]D5. The parallel between genetic and autoimmune forms (e.g., glycine receptor mutations in hereditary hyperekplexia vs. glycine receptor antibodies in acquired stiff-person spectrum) underscores shared pathophysiological nodes ^[39]D5.

Pearl: The final common pathway of cerebellar ataxia is disruption of Purkinje cell output or its afferent/efferent connections; identifying the specific node, repeat expansion, ion channel defect, autoantibody, or metabolic lesion, guides targeted therapy and genetic counseling.

The anatomic and molecular pathways described above give rise to a broad differential diagnosis, and the burden of cerebellar ataxia in the population reflects this heterogeneity. Hereditary cerebellar ataxias (HCA) have a population prevalence of 8.9 per 100,000 (dominant 5.6, recessive 3.3) in the largest population-based survey, with Machado-Joseph disease (SCA3) the most common at 3.1 per 100,000 and Friedreich ataxia at 1.0 per 100,000 ^[53]D5. Acquired causes, autoimmune, paraneoplastic, toxic, infectious, and vascular, add substantially to this burden, though their precise incidence is not captured in prevalence surveys of hereditary disease.

Etiologic Buckets

• Genetic: Repeat expansions (e.g., RFC1, FGF14, SCA1/2/3, Friedreich ataxia), point mutations (e.g., SPG7, SACS, SETX, SYNE1, CACNA1A) ^[44]C4, and de novo mutations account for 35% of childhood-onset cerebellar atrophy ^[54]D5. RFC1 biallelic AAGGG expansions cause 14% of unselected late-onset ataxia and 67% of patients with ≥2 features of CANVAS or ataxia with chronic cough ^[46]B3b. SCA27B (FGF14) is a common late-onset ataxia with episodic onset in 51% ^[47]B3b.
• Autoimmune/Paraneoplastic: Anti-CRMP5 neuropathy includes cerebellar ataxia in 21% of 105 patients, with 88% smokers and 69% cancer (75% small-cell lung cancer) ^[51]D5. Anti-MAP1B (PCA-2) associates with cerebellar ataxia in 38% and cancer in 79% (mostly SCLC) ^[55]D5. Anti-CASPR2 encephalitis includes ataxia in 33% ^[45]C4.
• Post-Infectious: Acute cerebellar ataxia can follow respiratory viral illness, as in a toddler with acute-onset ataxia after a viral prodrome ^[48]C4.
• Toxic/Metabolic: Alcohol, chemotherapy, and metabolic disorders (e.g., acute intermittent porphyria-related leukoencephalopathy with cerebellar atrophy ^[43]C4) are established causes.

Risk Factors

Special Considerations

• Paraneoplastic ataxia often precedes cancer diagnosis by a median of 185 days (CRMP5) ^[51]D5; urgent cancer screening (CT chest, PET) is warranted.
• RFC1 disease presents with chronic cough a median of 16 years before gait ataxia ^[46]B3b; testing should be considered in any late-onset ataxia with cough or sensory neuropathy.
• De novo mutations are a major cause of childhood-onset cerebellar atrophy, emphasizing the role of exome sequencing even without family history ^[54]D5.

Pearl: In any adult with late-onset ataxia and chronic cough, test for RFC1 repeat expansions; in any smoker with painful asymmetric neuropathy and ataxia, test for CRMP5-IgG and screen for lung cancer.

The clinical presentation of cerebellar ataxia spans a wide spectrum, from acute post-infectious cerebellitis in toddlers to slowly progressive degeneration in the sixth decade. The tempo of onset, acute, subacute, episodic, or chronic, and the pattern of associated neurologic and systemic features provide the first clues to etiology.

Presenting Symptoms

Onset and progression vary by cause. In episodic ataxia type 1 (EA1), the first episode occurs before age 20 in all but one patient, with an average age of onset of 7.9 years; attacks last minutes and are triggered by physical exertion, emotional stress, or environmental temperature ^[58]C4. In contrast, GAD65-antibody-associated ataxia presents subacutely over weeks in 38% of patients, with a median age of 58 years (range 33-80); 26% report preceding episodes of brainstem/cerebellar dysfunction or persistent vertigo ^[57]C4. Adult Niemann-Pick type C (NPC) has a mean onset of 25 ± 9.7 years, with a diagnostic delay of 6.2 ± 6.4 years ^[26]C4. Autosomal recessive cerebellar ataxia type 1 (ARCA-1) begins in middle age (mean 31.6 years, range 17-46) and progresses slowly ^[68]D5. Hereditary sensory and autonomic neuropathy type 1E (HSAN1E) starts at a mean of 37.7 years ^[27]C4. Multiple system atrophy of cerebellar type (MSA-C) has an average onset of 58.4 years, and median disease duration from symptom onset to death is 7.51 years (95% CI 7.18-7.78) ^[60]B3b. CANVAS (cerebellar ataxia, neuropathy, vestibular areflexia syndrome) typically begins in the sixth decade with progressive unsteadiness; a dry spasmodic cough often precedes walking difficulty by decades ^[67]D5. SCA27B, caused by an FGF14 repeat expansion, presents in the 40s to 70s with slowly progressive pancerebellar syndrome; up to half of patients report episodic symptoms, imbalance, vertigo, visual disturbances, or dysarthria, that may occur without interictal cerebellar signs ^[61]C4.

Associated symptoms commonly include:

• Dysarthria (63% in adult NPC ^[26]C4; prominent in ARCA-1 ^[68]D5)
• Dysphagia (37% in adult NPC ^[26]C4)
• Cognitive impairment (61% in adult NPC ^[26]C4; 89% by age 45 in HSAN1E ^[27]C4)
• Psychiatric disorders (45% in adult NPC ^[26]C4; 15.3% in Twinkle-related disorders ^[42]B3b)
• Sensory symptoms (paresthesias, numbness) in CANVAS ^[67]D5
• Oscillopsia (CANVAS) ^[67]D5
• Autonomic dysfunction (orthostatic intolerance, bladder symptoms) in MSA ^[60]B3b
• Hearing loss (17.5% in Twinkle-related disorders ^[42]B3b; sensorineural in XP ^[59]B2b)
• Seizures (epilepsy in adult NPC ^[26]C4; myoclonic seizures in HSAN1E ^[27]C4)

Neurological Examination Findings

The examination should systematically assess cerebellar function and associated neurologic signs.

Gait and stance: Gait ataxia is the most common presenting complaint. In CANVAS, half of patients need walking aids after 10 years and a quarter are wheelchair-dependent after 15 years ^[67]D5. In MSA, falls within 3 years of onset are an independent predictor of shorter survival (HR 2.31, P < 0.0001) ^[60]B3b.

Limb coordination: Finger-to-nose and heel-to-shin testing reveal dysmetria, intention tremor, and dysdiadochokinesia. The Scale for the Assessment and Rating of Ataxia (SARA) quantifies severity: in EA1, average SARA score was 3.15 (range 0-14) overall, but 7.7 in those with progressive cerebellar ataxia versus 2 in isolated episodic ataxia ^[58]C4. In XP, SARA scores increase over time in XPD (0.91 points/year, 95% CI 0.61-1.21) and XPA (0.63 points/year, 95% CI 0.38-0.89) ^[59]B2b.

Oculomotor examination: Vertical supranuclear ophthalmoplegia (VSO) is a hallmark of adult NPC, present in 75% of patients ^[26]C4. Cerebellar oculomotor abnormalities include gaze-evoked nystagmus, saccadic dysmetria, and impaired smooth pursuit. In SCA27B, cerebellar oculomotor abnormalities are prominent ^[61]C4. In HSAN1E, vertical gaze and convergence palsies with nystagmus are described ^[43]C4.

Speech: Dysarthria is common, 63% in adult NPC ^[26]C4, significant in ARCA-1 ^[68]D5.

Reflexes and tone: In ARCA-1, occasional brisk reflexes in the lower extremities are noted ^[68]D5. In CANVAS, sensory neuropathy leads to hyporeflexia or areflexia ^[67]D5. In XP, hyporeflexia and hypopallesthesia are frequent ^[59]B2b. Spasticity may be present in leukoencephalopathies such as CLCN2 mutations ^[65]D5.

Sensory examination: Sensory ataxia from large-fiber neuropathy is a core feature of CANVAS; sensory neuropathy is present in all genetically confirmed cases ^[67]D5. In HSAN1E, sensory neuropathy is part of the classic triad with hearing loss and cognitive decline ^[27]C4.

Other signs: Movement disorders (chorea, dystonia) occur in 58% of adult NPC ^[26]C4 and in XP ^[59]B2b. Palatal myoclonus may be seen in caudal paramedian midbrain syndrome ^[62]C4. Fasciculations and elevated creatine kinase have been reported in CANVAS with a novel (ACAGG)exp repeat motif ^[66]C4.

Phenotypic Variants

Red Flags

• Falls within 3 years of onset in MSA: HR 2.31 for shorter survival ^[60]B3b
• Bladder symptoms or urinary catheterization within 3 years in MSA: HR 1.96 and 1.67, respectively ^[60]B3b
• Orthostatic intolerance within 1 year in MSA: HR 1.28 ^[60]B3b
• Dry spasmodic cough preceding ataxia by decades: suggests CANVAS ^[67]D5
• Subacute onset with systemic autoimmunity (type 1 diabetes, thyroid disease): test for GAD65 antibodies ^[57]C4
• Classic paraneoplastic syndrome (limbic encephalitis, cerebellar degeneration) with older age, male sex, and coexisting neuronal cell-surface antibodies: high risk of underlying cancer (OR 10.5 for classic PNS) ^[64]C4
• Rapidly progressive dementia with ataxia not meeting CJD criteria: test CSF for neuronal surface antibodies (1.7% of suspected CJD have treatable autoimmune cause) ^[5]C4
• Episodic symptoms in older adults without interictal signs: consider SCA27B ^[61]C4

Atypical Presentations

• Acute-onset ataxia in toddlers after viral illness: may be due to genetic etiology (e.g., biotinidase deficiency , mitochondrial disorder) with metabolic acidosis and MRI diffusion restriction in middle cerebellar peduncles ^[48]C4
• Acute cerebellitis with Homer-3 antibodies: presents with headache, nausea, vomiting, confusion, and pancerebellar syndrome within a week; CSF pleocytosis (60 WBC/μL) but normal MRI ^[63]C4
• Isolated sensory neuropathy without cerebellar or vestibular signs: 15% of CANVAS patients have only sensory neuropathy at presentation ^[67]D5
• Episodic ataxia with normal interictal exam: SCA27B may present with episodic imbalance, vertigo, or visual disturbances before developing persistent cerebellar signs ^[61]C4
• Leukoencephalopathy with white matter oedema (CLCN2 mutations): cerebellar ataxia, spasticity, chorioretinopathy, optic neuropathy, cognitive defects; MRI shows restricted diffusion in posterior limbs of internal capsules, cerebral peduncles, and middle cerebellar peduncles ^[65]D5

Pearl: In any patient with late-onset ataxia and sensory neuropathy, ask about a dry spasmodic cough, if present, test for the RFC1 repeat expansion; a positive finding explains up to 88% of complete CANVAS and 6% of unspecified late-onset ataxias ^[21]C4[67]D5.

The clinical features described above guide the initial diagnostic workup, which aims to distinguish among the many causes of cerebellar ataxia. The approach is hierarchical: confirm the ataxia is cerebellar, then use the tempo, associated signs, and family history to select targeted tests.

History and Physical: Red Flags That Direct Testing

• Acute/subacute onset (days to weeks): consider stroke, autoimmune encephalitis , paraneoplastic syndromes, or infection. Subacute onset predicts immunotherapy response in GAD65-antibody ataxia (OR 0.50, 95% CI 0.25-0.99) ^[57]C4.
• Chronic cough preceding ataxia by decades: strongly suggests RFC1 disease (positive predictive value >90%) ^[46]B3b.
• Autonomic failure (orthostatic hypotension, urinary incontinence): points to multiple system atrophy (MSA). Probable MSA requires a systolic blood pressure drop ≥30 mm Hg or diastolic ≥15 mm Hg within 3 minutes of standing ^[1]A1c.
• Family history: autosomal dominant suggests spinocerebellar ataxia (SCA); recessive or sporadic with consanguinity suggests RFC1, Friedreich ataxia , ataxia-telangiectasia (A-T), or ataxia with oculomotor apraxia type 1 (AOA1) ^{[44]C4[46]B3b[72]C4[74]B3b}.
• Oculomotor apraxia: seen in AOA1 and A-T ^{[72]C4[74]B3b}.
• Telangiectasias: pathognomonic for A-T ^[72]C4.
• Cancer history: prompt paraneoplastic antibody testing (anti-Yo, anti-Hu, anti-Tr, anti-CASPR2, anti-GFAP) ^{[5]C4[8]C4[45]C4}.

Gold-Standard Test

• Definite MSA requires autopsy demonstration of widespread CNS α-synuclein-positive glial cytoplasmic inclusions with striatonigral or olivopontocerebellar degeneration ^[1]A1c.
• Probable MSA is diagnosed clinically using consensus criteria: sporadic, progressive, adult-onset (>30 years) disease with autonomic failure (urinary incontinence or orthostatic hypotension as defined above) plus poorly levodopa-responsive parkinsonism or a cerebellar syndrome ^[1]A1c.
• Genetic ataxias: molecular confirmation by targeted repeat expansion testing or next-generation sequencing (exome or panel) ^{[44]C4[71]B3b}.
• Autoimmune ataxias: detection of specific antibodies in serum or CSF (GAD65, CASPR2, GFAP, NMDAR, LGI1, glycine receptor) ^{[5]C4[8]C4[45]C4[57]C4}.

Laboratory Studies

• Serum: α-fetoprotein (AFP) - elevated in A-T (virtually all patients) and in 41% of AOA1 (median 6.0 ng/mL vs 3.4 ng/mL in controls) ^{[72]C4[74]B3b}; vitamin E, thyroid function, celiac serology; anti-GAD65, anti-CASPR2, anti-GFAP, paraneoplastic panel.
• CSF: cell count, protein, oligoclonal bands; 14-3-3 protein (CJD); neuronal surface antibodies - found in 1.7% of patients with suspected CJD, none in definite CJD ^[5]C4; GFAP IgG ^[8]C4.
• Genetic testing: first-line for chronic progressive ataxia with family history or specific clues. Exome-targeted capture yields a very probable/definite diagnosis in 22.6% of undiagnosed patients (highest in spastic ataxia, 35.0%) ^[44]C4. RFC1 repeat expansions account for 14% of unselected late-onset ataxia and 67% of those with ≥2 features of CANVAS/ACC ^[46]B3b. FGF14 GAA repeat expansions (≥250 repeats) are found in 61% of French Canadian, 18% of German, 15% of Australian, and 10% of Indian index patients with late-onset cerebellar ataxia ^[71]B3b.

Imaging

• MRI brain (1.5 T or higher): T2-weighted sequences to detect putaminal hypointensity, hyperintense lateral putaminal rim, "hot cross bun" sign (pontine cruciform hyperintensity), and middle cerebellar peduncle (MCP) hyperintensities - all supportive of MSA ^[1]A1c. Cerebellar vermian atrophy is the most common finding in RFC1 disease (87%) but is not universal ^[46]B3b. POLR3A-related spastic ataxia shows hyperintensities along the superior cerebellar peduncles rather than hypomyelination ^[77]B3b.
• FDG-PET: striatal or brainstem hypometabolism helps diagnose MSA; cerebellar hypometabolism in a patient with parkinsonism suggests MSA-P rather than Parkinson disease ^[1]A1c.
• [¹²³I]-FP-CIT SPECT (DAT scan): demonstrates nigrostriatal dopaminergic denervation. In MSA-C, baseline striatal/occipital ratio is lower (2.3 vs 2.97) and declines faster than in idiopathic late-onset cerebellar ataxia (ILOCA) ^[7]B3b.

Neurophysiology

• NCS/EMG: axonal sensory neuropathy is characteristic of RFC1 disease (present in 100% of patients; sensory amplitudes absent in 26%, length-dependent in 30%, non-length-dependent in 44%) ^{[46]B3b[75]C4}. Motor neuropathy with predominant proximal lower limb weakness occurs in GM2 gangliosidosis ^[76]C4.
• Vestibular testing: impaired visually enhanced vestibulo-ocular reflex (VVOR) is the pivotal sign of CANVAS/RFC1 disease ^[79]C4.
• EEG: periodic sharp-wave complexes suggest Creutzfeldt-Jakob disease .

Autonomic Testing

• Tilt-table test: document orthostatic hypotension per MSA criteria ^[1]A1c.
• Bladder ultrasound: post-void residual ≥100 mL indicates incomplete emptying, common in MSA ^[1]A1c.

Diagnostic Algorithm

Step 1: Classify onset. Acute/subacute presentations require urgent CSF analysis and MRI to exclude treatable autoimmune or prion disease. Step 2: For chronic progressive ataxia, take a detailed family history. Dominant pedigrees point to SCA or FGF14 testing; recessive or sporadic cases warrant RFC1 testing (especially if cough or sensory neuropathy is present) followed by exome sequencing. Step 3: In the absence of family history, assess for autonomic failure. If present, apply MSA consensus criteria and consider DAT-SPECT and MRI for supportive features. Step 4: If cough is a clue, RFC1 testing is first-line. Step 5: For all others, a broad serum screen (AFP, vitamin E, thyroid, celiac, anti-GAD65) and brain MRI are initial steps; if unrevealing, proceed to next-generation sequencing.

Diagnostic Test Performance

Pearl: In a patient with chronic progressive ataxia and chronic cough or sensory neuropathy, test for RFC1 repeat expansions first - it accounts for up to 67% of such cases and can be diagnosed by a simple PCR-based assay ^[46]B3b.

Once the diagnosis of cerebellar ataxia is established, the next step is to quantify its severity, stage the disease, and identify prognostic factors that guide management and trial eligibility. Several validated clinical scales provide objective measures of ataxia severity, functional impairment, and disease progression.

Clinical Rating Scales

The Scale for the Assessment and Rating of Ataxia (SARA) is the most widely used outcome measure. Scores range from 0 to 40 (higher = worse), with a minimal clinically important difference of approximately 1.0 point ^[83]A1b. SARA has high sensitivity (0.99) and specificity (0.96), with excellent interrater and test-retest reliability ^[83]A1b. It is the primary endpoint in most recent trials ^{[81]A1b[82]A1b[83]A1b[85]A1b}. The International Cooperative Ataxia Rating Scale (ICARS) is an alternative, with a 5-point drop considered clinically meaningful ^[4]C4.

For multiple system atrophy (MSA), the Unified MSA Rating Scale (UMSARS) is disease-specific. It comprises part I (activities of daily living, 0-48), part II (motor examination, 0-56), and a total score (0-104) ^[69]B2b. UMSARS progression is rapid: first-year rates average 6.5 points (32.3%) for part I and 8.2 points (44.9%) for part II, with attenuation in the second year ^[69]B2b. The median annual UMSARS total progression rate is 10.27 points/year (IQR 5.31-14.30) ^[91]B2b.

Other scales include the Spinocerebellar Ataxia Functional Index (SCAFI), the modified Disability Rating Scale (mDRS) ^[89]C4, the Fatigue Severity Scale (FSS) ^{[83]A1b[86]B2b}, and quality-of-life instruments such as the SF-36 and EQ-5D ^{[58]C4[69]B2b}.

Staging and Clinical Milestones

Disease staging in MSA uses the Hoehn and Yahr (1-5) and Schwab and England (0-100, lower = worse) scales, along with a three-point global disability rating (mild, moderate, severe) ^[69]B2b. More importantly, specific clinical milestones mark advanced disease and portend short survival:

• Falls at least once a day - probability increases during the second year of follow-up ^[69]B2b.
• Wheelchair dependency - median survival from indoor wheelchair use is <1.5 years ^[91]B2b.
• Gastrostomy placement (for severe dysphagia) - median survival <1.5 years ^[91]B2b.
• Unintelligible speech - median survival <1.5 years ^[91]B2b.

These milestones are reliable late-stage markers; their occurrence should trigger discussions about advanced care planning ^[91]B2b.

Prognostic Factors

In MSA, the strongest predictors of shorter survival are:

• Parkinsonian phenotype (MSA-P) - hazard ratio (HR) 2.08 (95% CI 1.09-3.97) compared to MSA-C ^[69]B2b.
• Incomplete bladder emptying - HR 2.10 (95% CI 1.02-4.30) ^[69]B2b.
• Severe symptomatic autonomic failure at diagnosis (symptomatic orthostatic hypotension, urinary incontinence, or both) - median survival 8.0 years vs 10.3 years without severe autonomic failure ^[70]B2b.
• Later disease onset - HR 1.04 per year (P<0.001) ^[91]B2b.
• Absent levodopa response - odds ratio 3.38 for rapid progression (95% CI 1.12-10.22) ^[69]B2b.
• Higher baseline UMSARS and faster UMSARS progression - both independently associated with shorter survival from baseline (HR 1.03 and 1.07, respectively; both P<0.001) ^[91]B2b.

Using landmark analysis, negative prognostic factors at 3 years from onset include recurrent falls, unintelligible speech, catheter use, and medication for orthostatic hypotension (HR 1.57-3.29; all P<0.05). At 5 years, mobility milestones (walking aid, wheelchair use) become significant, while dysautonomia milestones lose significance ^[91]B2b. The commonest cause of death is respiratory infection (67%) ^[91]B2b.

Genotype-Phenotype Correlations and Risk Stratification

In hereditary ataxias, genotype strongly influences severity and progression:

• SPG7 (spastic paraplegia gene 7): Mean annual SARA increase is 1.0 ± 1.4 points. Patients with loss-of-function variants present more often with pyramidal signs and optic atrophy; those with at least one Ala510Val variant have later onset and more frequent ataxia at onset ^[88]C4.
• Ataxia with oculomotor apraxia type 1 (AOA1): Oculomotor apraxia correlates with more severe disease and is more frequent with truncating mutations (83%) than missense (17%). Missense mutations are associated with later onset (mean 17.7 vs 5.2 years) ^[74]B3b. Elevated α-fetoprotein (AFP) is a biomarker; median AFP 6.0 ng/mL vs 3.4 ng/mL in controls ^[74]B3b.
• Ataxia-telangiectasia (A-T): Elevated AFP and chromosome 7/14 rearrangements are diagnostic. Variant A-T presents with extrapyramidal symptoms before ataxia; malignancy risk is increased ^[72]C4.
• Episodic ataxia type 1 (EA1): Average SARA is 3.15 (range 0-14); 26% develop permanent cerebellar signs, related to disease duration. Attack frequency ranges from daily to monthly ^[58]C4.
• Autosomal dominant Alzheimer disease (ADAD): Cerebellar ataxia occurs in 15% of symptomatic carriers in the DIAN-OBS cohort, more than in literature reports (3.1%) ^[40]B2a. Younger age at onset is associated with higher risk of ataxia (OR 0.82 per year) ^[40]B2a.

Pearl: In MSA, the combination of parkinsonian phenotype, early severe autonomic failure, and absent levodopa response identifies patients with the most rapid progression (median survival ~8 years); clinical milestones such as recurrent falls, wheelchair dependency, and gastrostomy each signal a median survival of less than 1.5 years, making them actionable triggers for palliative care referral.

Once the severity of acute cerebellar ataxia is classified and life-threatening causes excluded (see Severity, Staging & Risk Stratification), management proceeds along a stepwise pathway that targets the underlying inflammatory or infectious trigger while providing symptomatic support. The goal is to abort the acute episode, prevent neurologic deterioration, and transition to long-term disease-modifying therapy when indicated.

Step 1: Initial Assessment and Severity Classification

Triage begins with identification of red flags for a clinically urgent neurologic pathology (CUNP). In a multicenter study of 509 children presenting with acute ataxia, meningeal signs (OR 7.7), focal neurologic signs (OR 3.0), hyporeflexia (OR 3.5), and ophthalmoplegia (OR 3.0) were significantly associated with CUNP ^[106]D5. Each day from symptom onset increased the odds of CUNP by 51% (OR 1.5, 95% CI 1.1-1.2) ^[106]D5. Conversely, a history of varicella-zoster virus infection (OR 0.1) and vertigo (OR 0.5) lowered the risk ^[106]D5.

Severity is graded using a composite score of cerebellar signs (ataxia, dysmetria, balance disturbance, nystagmus, dysarthria) and extracerebellar features (headache, vomiting, hypotonia, behavioral changes, sensory changes). In a retrospective study of 124 children with acute cerebellitis/acute cerebellar ataxia (AC/ACA), 56% were classified as low severity (score 1-3), 39% as moderate (score 4-6), and 5.6% as severe (score ≥7) ^[94]A1b. Severe cases had a higher probability of pathological CT findings (P=0.009) and were more likely to receive intravenous steroids (P=0.006) ^[94]A1b.

Disposition:

• ICU: Severe cases with altered mental status, seizures, respiratory compromise, or cerebellar edema on imaging.
• Ward: Moderate cases requiring IV therapy and monitoring.
• ED observation or outpatient: Mild cases with no red flags and reliable follow-up.

Step 2: First-Line Intervention

For suspected autoimmune or post-infectious cerebellitis:

• High-dose corticosteroids are the mainstay. In a cohort of 38 patients with acute CNS inflammation following COVID-19 vaccination, 57.9% received acute treatment: IV methylprednisolone (n=16), high-dose pulse oral prednisone (n=5), or oral dexamethasone (n=1) ^[92]C4. For children, IV methylprednisolone 30 mg/kg/day (max 1 g) for 3-5 days is typical ^[94]A1b.
• Intravenous immunoglobulin (IVIG) is an alternative first-line agent, particularly when corticosteroids are contraindicated or in antibody-mediated syndromes. In a case of post-EBV acute cerebellar ataxia, IVIG 0.4 g/kg/day for 5 days led to rapid clinical improvement after steroid failure ^[108]C4.

For suspected viral cerebellitis (VZV, HSV, EBV):

• Acyclovir 10 mg/kg IV every 8 hours for 5-7 days is often added, though evidence for benefit is limited. In the pediatric AC/ACA study, 65% received IV acyclovir, but outcome was not affected by antiviral therapy (P>0.05) ^[94]A1b. For VZV-related cases, antivirals are widely used despite no proven efficacy ^[94]A1b.

For paraneoplastic or immune checkpoint inhibitor-related ataxia:

• High-dose corticosteroids are first-line, with escalation to IVIG or plasma exchange if no response ^[98]C4.

For anti-mGluR1 encephalitis:

• First-line therapy includes glucocorticoids, IVIG, and plasma exchange ^[97]B2a. In a pooled analysis of 53 patients, 65.9% clinically improved with immunotherapy ^[99]C4.

Step 3: Second-Line and Escalation

Inadequate response to first-line (no improvement after 3-5 days):

• Plasma exchange (PLEX): In the post-vaccine cohort, 13.2% of patients who failed steroids received PLEX ^[92]C4. Typical regimen: 5-7 sessions over 10-14 days.
• Repeat IV methylprednisolone or IVIG if not already used ^[92]C4.

For anti-mGluR1 encephalitis:

• Rituximab is the most commonly used second-line agent ^[97]B2a. Early initiation (within weeks of onset) may improve outcomes. In a case report, rituximab 375 mg/m² weekly for 4 weeks led to significant improvement, including ability to walk unaided, with no relapse at >6 months ^[109]C4.

For refractory cases (experimental):

• Efgartigimod, a neonatal Fc receptor inhibitor, was used off-label in an 11-year-old with seronegative autoimmune cerebellar ataxia. A single dose of 10 mg/kg IV resulted in 67% symptom resolution (SARA score from 18 to 6) within 3 days ^[102]C4. Further studies are needed.

Step 4: Monitoring and Titration

• Clinical monitoring: Use the Scale for the Assessment and Rating of Ataxia (SARA) or Expanded Disability Status Scale (EDSS) at baseline and after each treatment cycle. In the post-vaccine cohort, median EDSS improved from 2.5 at nadir to 2.0 at last follow-up (median 141 days) ^[92]C4.
• Laboratory monitoring: For patients on corticosteroids, monitor glucose, electrolytes, and blood pressure. For IVIG, monitor renal function and infusion reactions. For rituximab, check B-cell counts and IgG levels; screen for hepatitis B reactivation.
• Imaging: Repeat MRI brain at 3-6 months if initial abnormal. In the pediatric AC/ACA study, 35.7% of repeat MRIs showed improvement, 32.1% were stable, and 32.1% worsened ^[94]A1b. Pathological MRI at admission was associated with higher risk of neurological sequelae (P=0.024) ^[94]A1b.

Step 5: Resolution and Transition to Long-Term Management

• Outcome: Most patients achieve partial or complete recovery. In the post-vaccine cohort, 78.9% had partial recovery, 18.4% returned to baseline, and 2.6% worsened ^[92]C4. For anti-mGluR1 encephalitis, complete remission was achieved in only 22.2%, and 61.8% remained disabled ^[97]B2a.
• Transition: Patients with a confirmed autoimmune or demyelinating diagnosis (e.g., multiple sclerosis, MOGAD, NMOSD, anti-mGluR1 encephalitis) should be started on appropriate disease-modifying therapy (see Long-term & Definitive Management). In the post-vaccine cohort, 55.3% started DMT, including ocrelizumab, dimethyl fumarate, ofatumumab, and glatiramer acetate ^[92]C4. For post-infectious cases without recurrence, no long-term immunosuppression is needed.

Drug / Modality Comparison Table

Dosing Table

Treatment Failure Protocol

• Day 3-5: If no improvement after first-line steroids, add IVIG (if not already used) or plasma exchange.
• Day 7-10: If still no response, consider rituximab (for antibody-mediated syndromes) or efgartigimod (experimental).
• Re-evaluate diagnosis: Repeat MRI brain with contrast, CSF analysis for oligoclonal bands and autoantibodies (including anti-mGluR1, MOG, AQP4), and consider paraneoplastic panel.

What NOT to Do

• Do not delay corticosteroids in suspected autoimmune cerebellitis while awaiting antibody results; early treatment improves outcomes ^[97]B2a.
• Do not use acyclovir routinely unless VZV or HSV infection is confirmed or strongly suspected; it does not improve outcomes in non-viral cases ^[94]A1b.
• Do not use benzodiazepines or other sedatives for ataxia; they may worsen gait instability and mask neurologic deterioration.
• Do not administer anticoagulation for acute cerebellar stroke without confirmatory imaging (CT or MRI) ^[107]D5.

Controversies and Guideline Disagreement

Pearl: In acute cerebellar ataxia, initiate high-dose corticosteroids promptly when autoimmune or post-infectious etiology is suspected; escalate to IVIG or plasma exchange within 3-5 days if no response, and consider rituximab early for anti-mGluR1 encephalitis to improve functional outcomes ^{[97]B2a[109]C4}.

After stabilization of acute episodes and exclusion of reversible causes, long-term management of cerebellar ataxia follows an evidence ladder that prioritizes rehabilitation, then subtype-specific pharmacotherapy, then neuromodulation. The ladder is built on a foundation of physiotherapy, with pharmacological and neuromodulatory options reserved for specific subtypes or refractory symptoms.

Step 1: Rehabilitation as First-Line Therapy

Physiotherapy is the cornerstone of chronic management. A meta-analysis of 18 RCTs (398 participants) demonstrated a significant reduction in the Scale for the Assessment and Rating of Ataxia (SARA) score with physiotherapy (mean difference [MD] -1.41, 95% CI -2.16 to -0.66) ^[126]A1a. Multi-aspect training programs combining muscle strengthening, coordination, gait, and activities of daily living (ADL) training were most effective (5 studies, MD -1.59, 95% CI -5.15 to -0.03). Balance training alone (3 studies, MD -1.58, 95% CI -2.55 to -0.62) and aerobic training (3 studies, MD -1.65, 95% CI -2.53 to -0.77) also showed significant benefit. Vibration and dual-task training did not produce significant effects ^[126]A1a.

Tai Chi training was evaluated in one assessor-blinded RCT (n=24). After 12 weeks, the Tai Chi group showed a trend toward improved dynamic balance on the Berg Balance Scale (BBS) (MD 4, 95% CI -1.06 to 8.71) but the difference was not statistically significant, and benefits were not sustained at 24-week follow-up ^[120]A1b. Home-based balance training with optokinetic stimuli was feasible in a small feasibility study (n=12) in spinocerebellar ataxia type 6, with strong test-retest reliability (ICC >0.7) and trends toward improvement ^[125]C4.

Step 2: Pharmacotherapy for Specific Subtypes

Pharmacotherapy is subtype-dependent. The table below summarizes agents with the strongest evidence.

4-aminopyridine is first-line for EA2 and DBN. One RCT showed a significant effect on attack frequency and quality of life in EA2 ^{[115]A1b[116]A1b}. For DBN, two RCTs demonstrated improvement in nystagmus intensity, visual acuity, and postural imbalance ^{[115]A1b[116]A1b}. For cerebellar gait ataxia, evidence comes from two observational studies ^{[115]A1b[116]A1b}.

Riluzole provides Class I evidence for symptomatic benefit in diverse chronic ataxias. In a double-blind, placebo-controlled pilot trial (n=40), 13 of 19 riluzole-treated patients achieved a ≥5-point drop on the International Cooperative Ataxia Rating Scale (ICARS) at 8 weeks compared with 1 of 19 placebo patients (OR 39.0, 95% CI 4.2-364.2; risk difference 63.2%, 95% CI 33.5%-; NNT=2) ^[4]C4. Mean ICARS total score improved by -7.05 (SD 4.96) vs 0.16 (SD 2.65) (p<0.001). Adverse events were sporadic and mild ^[4]C4.

Acetyl-DL-leucine was not superior to placebo in the ALCAT trial (n=105) for SARA total score (mean treatment difference 0.23, 95% CI -0.40 to 0.85) ^[122]A1b. It is not recommended for general cerebellar ataxia outside of clinical trials.

N-acetyl-L-leucine is under investigation for NPC. A multinational phase III crossover trial (IB1001-301) is ongoing, with dosing of 4 g/day for adults and weight-based dosing for children ^[127]A1b. Prior open-label studies suggested symptomatic and possible disease-modifying effects ^[127]A1b.

In pediatric ataxia-telangiectasia , a systematic review of 13 studies (314 participants) found that betamethasone produced transient improvements in SARA and ICARS but with dose-dependent systemic toxicity; erythrocyte-encapsulated dexamethasone was ineffective in phase 3 trials; nicotinamide riboside improved SARA and AT-NEST scores in open-label studies; and amantadine benefited patients with extrapyramidal symptoms ^[113]A1a. No pharmacological intervention has sufficient evidence for routine use in this population ^[113]A1a.

Step 3: Neuromodulation

Non-invasive brain stimulation is a promising option for patients who do not respond adequately to rehabilitation and pharmacotherapy.

Cerebellar transcranial direct current stimulation (tDCS): A randomized, double-blind, sham-controlled trial (n=61) with an open-label extension showed significant improvement in all motor scores (SARA, ICARS), cognition (cerebellar cognitive affective syndrome scale), and quality of life after 2 weeks of daily anodal cerebellar tDCS combined with cathodal spinal tDCS ^[84]A1b. Improvements were sustained at 52-week follow-up, and an add-on effect was observed after two repeated treatments ^[84]A1b. A separate double-blind RCT (n=20) confirmed improvement in SARA, ICARS, 9-hole peg test, 8-m walking time, and cerebellar brain inhibition after 2 weeks of anodal cerebellar tDCS ^[121]A1b.

Repetitive transcranial magnetic stimulation (rTMS): A meta-analysis of 7 RCTs found that cerebellar rTMS significantly improved SARA (standardized mean difference [SMD] -0.87, 95% CI -1.41 to -0.34), ICARS (SMD -1.06, 95% CI -1.47 to -0.64), and BBS (SMD 0.76, 95% CI 0.33 to 1.19) ^[112]A1a. High-frequency rTMS was more effective than low-frequency (SMD -1.28, 95% CI -1.82 to -0.74). Adverse events did not differ significantly from sham (OR 1.73, 95% CI 0.55 to 5.46) ^[112]A1a.

Non-invasive brain stimulation for dysarthria: A systematic review of 10 trials (268 adults) found inconclusive evidence due to high risk of bias and heterogeneity; no conclusions could be drawn ^[111]A1a.

Step 4: Emerging and Investigational Therapies

Mesenchymal stem cells (MSCs): A meta-analysis of 3 single-arm studies (47 patients) found no statistically significant improvement in BBS/SARA (SMD -0.20, 95% CI -0.78 to 0.38) or ICARS (SMD 0.36, 95% CI -0.08 to 0.81) after MSC therapy ^[114]A1a. No significant side effects were reported, but the evidence is insufficient to support clinical use ^[114]A1a.

Acetyl-DL-leucine for NPC: Open-label studies suggested benefit, but the phase III trial with N-acetyl-L-leucine is ongoing ^[127]A1b.

What NOT to Do

• Do not use acetyl-DL-leucine for general cerebellar ataxia outside of clinical trials; the ALCAT trial showed no benefit ^[122]A1b.
• Do not offer stem cell therapy outside of research settings; no RCT evidence supports efficacy ^[114]A1a.
• Do not rely on vibration or dual-task training as primary physiotherapy; meta-analysis showed no significant effect ^[126]A1a.

Controversies and Guideline Disagreement

No major guideline disagreements were identified in the reviewed evidence. The primary challenge is the paucity of high-quality RCTs for most interventions, leading to variable clinical practice. The therapeutic misestimation phenomenon, where patients expect large benefits (e.g., 31.5% average expected reduction in ataxia severity after 2 weeks of tDCS), complicates trial interpretation and may inflate placebo responses ^[119]A1b.

Pearl: Start all patients with multi-aspect physiotherapy (strength, coordination, gait, ADL training); add 4-aminopyridine for episodic ataxia type 2 or downbeat nystagmus, and consider riluzole for diverse chronic ataxias (NNT=2 for ≥5-point ICARS improvement); reserve cerebellar tDCS or rTMS for refractory cases based on emerging RCT evidence ^{[4]C4[84]A1b[112]A1a[126]A1a}.

The evidence base for treating cerebellar ataxia has evolved from anecdotal reports to randomized controlled trials, yet no disease-modifying therapy has been approved. The therapeutic timeline reveals a pattern of early promise, disease-specific heterogeneity, and a shift toward non-pharmacologic interventions.

The Riluzole Story

Riluzole, a glutamate modulator, first showed benefit in a 2010 pilot trial of 40 patients with mixed cerebellar ataxias: 13 of 19 patients achieved a ≥5-point drop on the International Cooperative Ataxia Rating Scale (ICARS) at 8 weeks versus 1 of 19 (5%) on placebo (OR 39.0, 95% CI 4.2-364.2) ^[4]C4. This Class I evidence prompted a larger trial in spinocerebellar ataxia (SCA) and Friedreich's ataxia (FRDA). In 2015, Romano et al. randomized 55 patients to riluzole 50 mg twice daily or placebo for 12 months. The proportion with improved Scale for the Assessment and Rating of Ataxia (SARA) score (≥1 point drop) was 50% (14/28) in the riluzole group versus 11% (3/27) in the placebo group (OR 8.00, 95% CI 1.95-32.83; p=0.002); NNT = 3 to achieve one additional patient with SARA improvement ^[82]A1b. However, the 2022 ATRIL trial in 45 patients with SCA2 found no benefit: 32% (7/22) of riluzole-treated patients improved versus 39% (9/23) on placebo (mean difference -10.3%, 95% CI - to 19.2%; p=0.75) ^[81]A1b. This contradiction underscores disease-specific heterogeneity and the need for subtype-targeted trials.

Non-Invasive Brain Stimulation

Transcranial direct current stimulation (tDCS) emerged as a symptomatic approach. In 2018, Benussi et al. conducted a double-blind, sham-controlled crossover trial of cerebello-spinal tDCS (5 days/week for 2 weeks) in 20 patients with neurodegenerative ataxia. Real stimulation improved all performance scores (SARA, ICARS, 9-Hole Peg Test, 8-m walking time) and restored cerebellar brain inhibition ^[3]A1b. A larger 2021 trial by the same group randomized 61 patients to real or sham tDCS for 2 weeks, followed by an open-label phase with repeated treatments. Motor and cognitive improvements (SARA, ICARS, cerebellar cognitive affective syndrome scale) were sustained up to 52 weeks, with an add-on effect after two treatment cycles ^[84]A1b. Despite these results, tDCS remains investigational and not yet standard of care.

Exercise and Rehabilitation

Exercise has shifted from compensatory balance training to high-intensity aerobic paradigms. In 2025, Barbuto et al. randomized 62 patients with various cerebellar ataxias to home high-intensity aerobic training (30 min, 5×/week at up to 85% predicted max heart rate) or dose-matched balance training. The aerobic group showed significantly greater SARA improvement at 6 months (β = -1.53, 95%; p=0.001). Those who continued training maintained benefits (SARA change -3.81, 95%), while those who stopped regressed ^[83]A1b. A 2024 multicenter trial by Milne et al. compared goal-directed rehabilitation (6 weeks outpatient physiotherapy + 24-week home program) to standard care in 71 patients with hereditary cerebellar ataxia. At 7 weeks, the rehabilitation group improved on the motor Functional Independence Measure (mean difference 2.26, 95% CI 0.26-4.26; p=0.028) and SARA (-1.21, 95% CI -2.32 to -0.11; p=0.032), with SARA benefit maintained at 30 weeks ^[85]A1b.

Failed Trials and Lessons Learned

Not all approaches succeeded. The PROMESA trial (2019) tested epigallocatechin gallate (up to 1200 mg/day) in 92 patients with multiple system atrophy (MSA). After 48 weeks, there was no difference in UMSARS motor score change (mean difference -0.94, 95% CI -3.71 to 1.83; p=0.51). Hepatotoxicity occurred in two patients, leading to a recommendation against doses >1200 mg ^[80]A1b. This trial highlighted the challenge of targeting α-synuclein aggregation in a rapidly progressive disease.

Natural History Informing Trial Design

Prospective natural history studies have provided essential benchmarks. The European MSA Study Group (EMSA-SG) followed 141 patients for 2 years, reporting median survival of 9.8 years from symptom onset and annualized UMSARS progression rates that enabled sample size calculations: 258 patients per group would detect a 30% reduction in UMSARS motor decline at 80% power ^[69]B2b. A 2026 analysis of 555 MSA patients (including 254 post-mortem confirmed) confirmed median survival of 8.25 years and identified time-dependent prognostic factors: at 3 years from onset, recurrent falls (HR 1.57), unintelligible speech (HR 3.29), and catheter use (HR 1.76) predicted shorter survival ^[91]B2b. These data are critical for stratifying patients in future trials.

Pearl: The riluzole story, positive in mixed ataxia and SCA/FRDA, negative in SCA2, teaches that treatment effects in cerebellar ataxia are disease-specific; future trials must target homogeneous genetic subtypes rather than lump all ataxias together.

Building on the historical evolution of treatment approaches, the current disease-modifying and immunotherapy program for autoimmune cerebellar ataxia requires a structured approach to sequencing, safety monitoring, and de-escalation. The evidence, drawn largely from retrospective cohort studies and case series, supports early, aggressive immunotherapy tailored to antibody subtype, onset tempo, and paraneoplastic status.

Step 1: Initial Assessment and Severity Classification

Classify patients by three axes: (1) antibody type, neuronal nuclear and/or cytoplasmic (NNC) antibodies (e.g., PCA-1/anti-Yo, anti-Hu) versus plasma membrane protein (PMP) antibodies (e.g., P/Q-type voltage-gated calcium channels, GAD65); (2) paraneoplastic versus nonparaneoplastic; and (3) onset tempo, subacute (weeks) versus insidious. In a cohort of 118 patients, neurologic improvements were significantly more common among patients with nonparaneoplastic disorders and those with exclusively PMP antibodies (P = 0.02) ^[130]C4. Progression to wheelchair dependence occurred significantly faster among patients with NNC antibody positivity only (P = 0.02), although those with GAD65 autoimmunity progressed at a rate similar to those with PMP autoimmunity ^[130]C4. Subacute onset of cerebellar ataxia (odds ratio [OR], 0.50; 95% CI, 0.25-0.99) and prompt immunotherapy (OR, 0.98; 95% CI, 0.96-0.99) are independent predictors of good outcome in GAD65-antibody-associated ataxia ^[57]C4.

Step 2: First-Line Intervention

Initiate first-line immunotherapy immediately after serologic confirmation. Options include high-dose intravenous methylprednisolone (IV-MTP), intravenous immunoglobulin (IVIG), plasma exchange (PE), or oral prednisolone. In a systematic review of 44 patients with anti-mGluR1 encephalitis, 37 of 43 (86.0%) received first-line immunotherapy (IV-MTP, IVIG, PE, or oral prednisolone) ^[140]C4. For SEZ6L2 autoimmunity, empiric IVIG 2 g/kg followed by monthly IVIg produced mild improvement (Brief Ataxia Rating Scale from 19.5 to 18 at 7 months) ^[138]C4. In GFAP astrocytopathy, neurologic improvement followed treatment with high-dose corticosteroids, with a tendency of patients to relapse without long-term immunosuppression ^[8]C4. For GAD65-antibody-associated ataxia, 20 of 25 patients with long-term follow-up received immunotherapy (IVIG in 10 and corticosteroids and IVIG or other immunosuppressors in 10), and 7 of them (35%) improved ^[57]C4.

Escalation trigger: If no improvement or worsening after 1-2 weeks of first-line therapy, proceed to second-line agents. In a cohort of 38 patients with acute CNS inflammation after COVID-19 vaccination (including one with primary autoimmune cerebellar ataxia [PACA]), 8 of 22 who received acute treatment had an inadequate response and went on to receive plasma exchange (n = 5), repeat IV methylprednisolone (n = 2), or IVIG (n = 1) ^[92]C4.

Step 3: Second-Line Intervention

For patients with inadequate response to first-line therapy or relapsing course, add or switch to second-line immunosuppressants. Options include cyclophosphamide (CYC), mycophenolate mofetil (MMF), azathioprine (AZA), tacrolimus (TAC), and rituximab (RTX). In the anti-mGluR1 encephalitis systematic review, 19 of 43 patients (44.2%) received second-line immunotherapy, and 18 (41.9%) received a combination of first- and second-line therapy ^[140]C4. The patient with PACA following COVID-19 vaccination was started on cyclophosphamide ^[92]C4. For septin-5 and septin-7 autoimmunity, 8 of 10 patients with data available improved after immunotherapy, and a further 2 improved spontaneously ^[137]D5. For DPPX-antibody-associated progressive encephalomyelitis with rigidity and myoclonus (PERM), response to immunotherapy was good, but constant and aggressive treatment may be required ^[22]C4.

Step 4: Safety Monitoring

Each immunosuppressive agent requires specific monitoring:

Step 5: De-escalation and Long-Term Maintenance

Once clinical stability is achieved (typically after 3-6 months of therapy), consider gradual tapering of corticosteroids and transition to a maintenance immunosuppressant. The risk of relapse without long-term immunosuppression is well documented. In GFAP astrocytopathy, patients tend to relapse without long-term immunosuppression ^[8]C4. In anti-mGluR1 encephalitis, relapses were documented in 5 patients, consistently occurring upon discontinuation of immunotherapy; symptoms improved or resolved upon resumption ^[140]C4. For GAD65-antibody-associated ataxia, long-term follow-up (median 5.4 years) showed that 35% of treated patients improved, but sustained response often required ongoing therapy ^[57]C4.

Treatment failure protocol: If no response after 3-6 months of second-line therapy, reassess for alternative diagnoses (e.g., missed malignancy, alternative antibody, non-immune etiology). Consider switching to a different second-line agent or adding a third agent. In the anti-mGluR1 series, some patients who failed first-line therapy achieved stabilization after second-line therapy, though not all achieved complete remission ^[140]C4.

What NOT to do: Do not delay immunotherapy while awaiting full serologic results if clinical suspicion is high, as prompt treatment is associated with better outcomes ^[57]C4. Do not rely solely on corticosteroids for long-term control in antibody-positive autoimmune ataxia without considering steroid-sparing immunosuppressants, given the high relapse rate upon steroid taper ^[8]C4. Do not use amantadine or leucine derivatives as first-line therapy for autoimmune ataxia; these agents have shown mixed results and insufficient evidence in pediatric ataxia-telangiectasia ^[113]A1a.

Controversies and Guideline Disagreement

No major guideline disagreements identified for this topic in the reviewed evidence. The available literature consists of retrospective cohort studies and case series, with no prospective randomized trials comparing immunotherapy regimens for autoimmune cerebellar ataxia. The evidence supports early, aggressive immunotherapy, but the optimal agent, dose, and duration remain undefined.

Pearl: Initiate first-line immunotherapy (high-dose corticosteroids, IVIG, or plasma exchange) promptly in patients with subacute-onset autoimmune cerebellar ataxia, especially those with nonparaneoplastic disorders or PMP/GAD65 antibodies, as early treatment is the strongest modifiable predictor of favorable outcome ^{[57]C4[130]C4}.

Alongside disease-modifying therapy, a structured rehabilitation program and targeted symptomatic management are essential to maintain function and quality of life in cerebellar ataxia. The evidence base for these interventions has grown substantially, with randomized trials now guiding exercise prescription, non-invasive brain stimulation, and pharmacologic symptom control.

Rehabilitation: Exercise and Physical Therapy

High-intensity aerobic training is the most strongly supported exercise modality. In a 12-month assessor-masked randomized trial of 62 individuals with various cerebellar ataxias (mean baseline SARA 12.1), home-based aerobic training (30 min/session, 5 days/week at up to 85% predicted maximum heart rate) produced significantly greater improvement in ataxia symptoms than dose-matched balance training: SARA β -1.53 (95% CI -2.44 to -0.61; P = 0.001) at 6 months ^[83]A1b. Fatigue also improved (β -9.38, 95% CI -15.1 to -3.7; P = 0.001) and VO₂max increased by 4.26 mL/kg/min (95% CI 2.1 to 6.4; P < 0.001) ^[83]A1b. Importantly, benefits were maintained only in participants who continued training regularly (SARA change -3.81, 95% CI -2.2 to -5.4); those who reduced or stopped training regressed toward baseline (SARA change 0.4, 95% CI -0.4 to 1.2) ^[83]A1b. No serious adverse events occurred; the most common minor adverse event in the aerobic group was mild pain, and in the balance group, falls without injury ^[83]A1b.

Goal-directed rehabilitation also improves function. In a multicenter single-blind randomized trial of 71 individuals with hereditary cerebellar ataxia, a 30-week program (6 weeks of outpatient physiotherapy followed by a 24-week home exercise program) improved the motor domain of the Functional Independence Measure (mFIM) by a mean 2.26 points (95% CI 0.26 to 4.26; P = 0.028) and SARA by -1.21 points (95% CI -2.32 to -0.11; P = 0.032) at 7 weeks compared with standard care ^[85]A1b. SARA improvement persisted at 30 weeks (mean difference -1.51, 95% CI -2.76 to -0.27; P = 0.017), though mFIM benefit was no longer significant ^[85]A1b. Frequent adverse events in both groups were fatigue, pain, and falls ^[85]A1b.

Telerehabilitation with robotic assistance is feasible. In a nonrandomized open-label trial of home-based telerehabilitation using a lumbar-type Hybrid Assistive Limb (HAL) three times weekly for 4 weeks in spinocerebellar ataxia, the Timed Up and Go test improved by a non-significant -1.3 s (P = 0.051), but the Berg Balance Scale improved by 2.6 points (P = 0.011) and the Cerebellar Cognitive Affective/Schmahmann Syndrome Scale (CCAS-S) by 3.7 points (P = 0.027) ^[146]A1b. No serious adverse events or dropouts occurred ^[146]A1b.

Non-Invasive Brain Stimulation

Cerebello-spinal transcranial direct current stimulation (tDCS) shows promise. In a randomized, double-blind, sham-controlled trial followed by an open-label phase, 61 patients with neurodegenerative ataxia received anodal cerebellar tDCS and cathodal spinal tDCS (or sham) for 5 days/week for 2 weeks ^[84]A1b. Real tDCS significantly improved motor scores (SARA, International Cooperative Ataxia Rating Scale), cognition (CCAS scale), and quality of life compared with sham, with effects persisting up to 52 weeks ^[84]A1b. A second treatment course produced an add-on effect ^[84]A1b. The improvement correlated with restoration of cerebellar inhibition measured by transcranial magnetic stimulation ^[84]A1b.

Pharmacologic Symptom Management

Acetyl-dl-leucine improved ataxia in Niemann-Pick type C (NPC). In a case series of 12 patients, treatment with acetyl-dl-leucine 3 g/d for 1 week then 5 g/d for 3 weeks reduced the median SARA score from 10.8 at baseline to 7.0 on medication (difference 3.8 points; P = 0.003), with scores returning to 10.5 after washout ^[89]C4. The 9-Hole Peg Test, modified Disability Rating Scale, and visual analog scale also improved ^[89]C4. No relevant side effects were reported except transient dizziness in one patient ^[89]C4.

Spasticity is a common accompaniment in many hereditary ataxias, including SPG7, SPG15, FAHN/SPG35, and ANO10-related ARCA3 ^{[88]C4[49]B3b[143]C4[145]D5}. Management follows standard guidelines with physical therapy, oral baclofen or tizanidine, and botulinum toxin injections for focal spasticity, though no ataxia-specific trials have tested these agents.

Neuropathic pain occurs in paraneoplastic ataxias such as CRMP5-autoimmune neuropathy, where 79% of patients had moderate to severe pain requiring a median of two neuropathic medications, with opioids used in 39% ^[51]D5. High-dose corticosteroids improved or stabilized neuropathy impairment scores (P = 0.012) ^[51]D5.

Dysphagia is a milestone in progressive ataxias. In the European MSA natural history study, the estimated probability of requiring nasogastric tube or gastrostomy increased with disease duration ^[69]B2b. Swallowing assessment and dietary modifications should be initiated early; speech-language pathology consultation is recommended.

Autonomic and Respiratory Support

Stridor in multiple system atrophy (MSA) occurs in 12%-42% of patients and may develop at any disease stage ^[2]A1c. Continuous positive airway pressure (CPAP) during sleep is recommended as first-line symptomatic therapy; it can eliminate stridor initially in most patients ^[2]A1c. Tracheostomy is effective for persistent or severe stridor and may improve survival ^[2]A1c. Home audio recording and video polysomnography help confirm the diagnosis ^[2]A1c.

Orthostatic hypotension and neurogenic bladder are common in MSA and other ataxias with autonomic involvement. In the EMSA-SG cohort, urinary incontinence was present in 73% and incomplete bladder emptying in 51% at baseline ^[69]B2b. Management includes volume expansion, compression garments, midodrine or fludrocortisone for hypotension, and intermittent catheterization or anticholinergics for bladder dysfunction.

Cognitive and Behavioral Sequelae

Impulsivity is increased in cerebellar ataxia. In a cross-sectional study using the Barratt Impulsivity Scale (BIS-11), 50 individuals with cerebellar ataxia scored 9.7% higher than controls (P < 0.001), driven primarily by the nonplanning domain ^[142]B3b. This contrasts with Parkinson disease , where impulsivity spans attentional, motor, and nonplanning domains ^[142]B3b. Clinicians should screen for impulsive behaviors and counsel patients and families.

Cognitive impairment (cerebellar cognitive affective syndrome) is common and may improve with rehabilitation. In the HAL telerehabilitation trial, CCAS-S scores improved by 3.7 points (P = 0.027) ^[146]A1b. Cognitive rehabilitation strategies should be incorporated into multidisciplinary care.

Hospital-Acquired Complications and Prophylaxis

Patients with advanced ataxia are at risk for aspiration pneumonia , pressure injuries, urinary tract infections, and falls. Preventive measures include:

• Aspiration precautions: swallowing assessment, modified diet, upright positioning during meals.
• Fall prevention: gait aids, home safety evaluation, balance training.
• Pressure injury prevention: regular turning, pressure-relieving mattresses.
• Urinary tract infection prevention: intermittent catheterization technique, adequate hydration.

Pearl: High-intensity aerobic training (30 min, 5 days/week at 85% max HR) improves ataxia symptoms more than balance training alone, but gains are lost if training is discontinued, sustained adherence is critical ^[83]A1b.

While neurorehabilitation addresses functional deficits, the clinician must also anticipate and intercept disease- and treatment-related complications that threaten survival and quality of life.

Disease-Specific Complications

Vanishing white matter disease (VWM) is notable for its sensitivity to febrile infections, minor head trauma, and acute fright, which may cause rapid neurological deterioration and unexplained coma ^[24]D5. Most patients die within a few years of onset ^[24]D5.

Ataxia-telangiectasia (A-T) carries a high burden of infectious, autoimmune, and malignant complications. Recurrent sinopulmonary infections are common, and immunoglobulin replacement therapy (0.6 g/kg every 4 weeks) with trimethoprim-sulfamethoxazole prophylaxis is standard ^[156]B3b. Autoimmune hemolytic anemia and juvenile idiopathic arthritis have been reported ^[156]B3b. Lymphoma occurs in up to 25% of patients; reduced-intensity chemotherapy is used due to radiosensitivity ^[156]B3b. Lupus vulgaris, a cutaneous tuberculosis, was documented for the first time in an A-T patient ^[156]B3b.

Paraneoplastic cerebellar degeneration may progress despite successful tumor treatment. In refractory cases, autologous hematopoietic stem cell transplantation has led to sustained neurologic improvement ^[20]C4.

Treatment-Related Complications

Immune checkpoint inhibitors (ICIs) cause neurologic adverse events in 2.9% of patients treated with anti-PD-1 antibodies, with subacute onset after a median of 5.5 cycles ^[90]C4. Phenotypes include cerebellar ataxia, myopathy, and varied neuropathies ^[90]C4. Management requires prompt ICI discontinuation; corticosteroids were used in 7 of 10 patients, IVIG in 3, and plasma exchange in 1 ^[90]C4. Nine of ten patients improved; one died of severe necrotizing myopathy ^[90]C4. Among nonfulminant neurologic immune-related adverse events (n-irAEs), 57% become chronic, and cerebellar ataxia is a sequela in 33% of chronic inactive cases ^[98]C4. Chronic n-irAEs are associated with higher mortality, primarily due to cancer progression ^[98]C4.

Deep brain stimulation targeting the dentate nucleus has shown modest improvements in selected hereditary and post-lesional ataxias, but adverse effects related to surgery and stimulation occur ^[151]B2a.

Scoliosis surgery in pediatric spinocerebellar ataxia and Friedreich ataxia carries specific risks: proximal junctional kyphosis in 23-25%, somatosensory evoked potential failure rates up to 91%, infection ≤4%, and cardiopulmonary death has been reported ^[153]B2a. Motor evoked potentials may offer greater sensitivity ^[153]B2a.

Hospital-Acquired Complications

General preventive measures for pneumonia, pressure injuries, and urinary tract infections apply, though specific evidence in cerebellar ataxia populations is not reported in the provided literature.

Complication Overview

Pearl: In any patient with cerebellar ataxia receiving immunotherapy, suspect an immune-related adverse event even with subacute onset; prompt ICI discontinuation and corticosteroids can be life-saving ^[90]C4.

The complications of cerebellar ataxia, particularly respiratory infection, falls, and dysphagia, define the terminal phase of the disease. Understanding the natural history and factors that modify it is essential for prognostication and shared decision-making.

Survival and Disease Progression in Multiple System Atrophy

MSA is the most extensively studied degenerative ataxia. Median survival from symptom onset is 8.25 to 9.8 years across large cohorts ^{[69]B2b[70]B2b[91]B2b}. Survival from diagnosis is shorter, 3.33 years in one series ^[60]B3b, reflecting the late stage at which diagnostic criteria are met. The commonest cause of death is respiratory infection (67%) ^[91]B2b.

Predictors of shorter survival include parkinsonian phenotype (MSA-P) ^[69]B2b, early autonomic failure (symptomatic orthostatic hypotension, urinary incontinence) ^[70]B2b, falls within 3 years of onset ^[60]B3b, bladder catheterization ^{[60]B3b[91]B2b}, and higher baseline UMSARS scores ^[91]B2b. Later disease onset also shortens survival (HR 1.04 per year, P<0.001) ^[91]B2b.

Clinical milestones mark the preterminal phase. Median survival from indoor wheelchair use, gastrostomy insertion, or development of unintelligible speech is consistently <1.5 years ^[91]B2b. The probability of falling at least once daily increases sharply during the second year of follow-up ^[69]B2b.

Disease progression rates are rapid. The UMSARS total score increases by 21.9 points (57%) over 2 years ^[69]B2b. Annualized UMSARS progression averages 10.27 points/year ^[91]B2b. Progression is fastest in the first year and attenuates in the second, likely reflecting a floor effect in advanced disease ^[69]B2b. Shorter symptom duration at baseline and absent levodopa response predict faster UMSARS decline ^[69]B2b.

Modifying the Trajectory

Pharmacotherapy. Riluzole 50 mg twice daily improved SARA scores in a mixed hereditary ataxia cohort (14/28 vs 3/27 improved; OR 8.00, 95% CI 1.95-32.83) ^[82]A1b, but a subsequent trial restricted to SCA2 found no benefit (7/22 vs 9/23; p=0.75) ^[81]A1b. Epigallocatechin gallate (up to 1200 mg/day) did not slow UMSARS progression in MSA (mean difference -0.94, 95% CI -3.71 to 1.83) ^[80]A1b.

Exercise and rehabilitation. Home high-intensity aerobic training (30 min, 5×/week at up to 85% max heart rate) improved SARA by -1.53 points (95% CI -2.44 to -0.61) compared with balance training at 6 months; benefits were maintained at 1 year only in those who continued training ^[83]A1b. Goal-directed rehabilitation improved function (mFIM +2.26, p=0.028) and ataxia (SARA -1.21, p=0.032) at 7 weeks, with SARA benefit sustained at 30 weeks (-1.51, 95% CI -2.76 to -0.27) ^[85]A1b. Intensive coordinative training also reduces ataxia symptoms, with effects sustained at 8-week follow-up ^[87]B2b.

Immunotherapy in autoimmune ataxia. Among 118 patients with autoimmune cerebellar ataxia, 54 (46%) improved with immunotherapy; nonparaneoplastic disorders and plasma membrane protein antibodies predicted better response ^[130]C4. In GAD65-antibody ataxia, subacute onset and prompt immunotherapy were associated with good outcome (OR 0.98, 95% CI 0.96-0.99) ^[57]C4. Patients with neuronal nuclear/cytoplasmic antibodies (e.g., anti-Yo) progressed to wheelchair dependence significantly faster ^[130]C4.

Prognosis in Other Ataxia Types

Spinocerebellar ataxias. Progression is slower than in MSA. In SCA2, median SARA score at enrolment was 13.5 (IQR 9.5-16.5) and worsened by 0.5 points over 12 months in the placebo group ^[81]A1b.

Episodic ataxia type 1. First episode occurs before age 20 (mean 7.9 years). Attack frequency ranges from daily to monthly; 26% develop permanent cerebellar signs, with SARA averaging 7.7 in those with progressive ataxia vs 2.0 in pure episodic cases ^[58]C4. Quality of life is reduced across all SF-36 domains, with mental health most affected ^[58]C4.

Genetic ataxias. In Twinkle-related disorders (TWNK mutations), mean onset is 40.3 years; 70.4% were alive at analysis, with progressive external ophthalmoplegia (84.7%) and skeletal myopathy (55.6%) as dominant features ^[42]B3b. Bi-allelic HMBS variants cause a slowly progressive leukoencephalopathy with ataxia and long life expectancy ^[43]C4.

Pearl: In MSA, the combination of falls within 3 years, bladder catheterization, and unintelligible speech identifies patients with median survival <1.5 years, a threshold that should trigger goals-of-care discussions ^[91]B2b.

Prognosis varies substantially across etiologies and age groups, but special populations require tailored diagnostic and therapeutic approaches.

Pediatrics

Acute ataxia in children is common, but age ≥ 5 years (OR 22.2, 95% CI 1.8-640.2) and symptom persistence beyond 3 days (OR 8.1, 95% CI 1.5-68.6) independently predict clinically urgent neurological pathology requiring intervention ^[167]B3b. Immune-mediated causes are increasingly recognized:

• MOG antibody-associated ataxia presents as a severe encephalitic syndrome with multifocal MRI lesions involving cerebral hemispheres, brainstem, and cerebellum; independent predictors include younger age (OR 0.718), higher acute mRS (OR 5.917), and cranial MRI abnormalities (OR 27.746) ^[162]B3b.
• Isolated CSF oligoclonal bands define a clinically mild but relapsing-prone subgroup with predominant cerebellar MRI findings ^[162]B3b.
• Opsoclonus-myoclonus syndrome (OMS) is frequently misdiagnosed as acute cerebellar ataxia; opsoclonus is recognized later in delayed-diagnosis cases (median 96 days vs 11.5 days, p < 0.01) ^[166]C4. OMS and rapidly progressive cerebellar ataxia are the most common pediatric paraneoplastic syndromes, associated with neuroblastoma in 88% ^[165]C4.
• Seronegative immune-mediated ataxia can relapse despite normal MRI and negative antibody panels; serial CSF may show evolving pleocytosis and intrathecal oligoclonal bands ^[168]C4.

Treatment modifications: N-acetyl-L-leucine for Niemann-Pick type C is dosed by weight: 4 g/day (≥ 13 years or 4-12 years ≥ 35 kg), 3 g/day (4-12 years, 25-< 35 kg), 2 g/day (4-12 years, 15-< 25 kg) ^[127]A1b. Betamethasone shows transient SARA/ICARS improvement in ataxia-telangiectasia but with dose-dependent systemic toxicity ^[113]A1a. Rituximab consolidation (four weekly infusions with B-cell monitoring) has been used in relapsing seronegative cases ^[168]C4. Multidisciplinary care includes scoliosis surveillance (proximal junctional kyphosis in 23-25% of surgical cases) ^[153]B2a, pulmonary management of bronchiectasis ^[169]C4, and immunoglobulin replacement (0.6 g/kg every 4 weeks) with trimethoprim-sulfamethoxazole prophylaxis for those with immunodeficiency ^[156]B3b.

Pregnancy

No controlled studies of cerebellar ataxia treatments in pregnancy exist in the reviewed literature; the N-acetyl-L-leucine phase III trial explicitly excludes pregnant or breastfeeding women and requires highly effective contraception ^[127]A1b. Many immunotherapies used for autoimmune ataxia, rituximab, mycophenolate , cyclophosphamide , are teratogenic and should be avoided. Corticosteroids and intravenous immunoglobulin are relatively safer but require individualized risk-benefit assessment. Delivery planning should involve neurology, obstetrics, and anesthesia. Breastfeeding safety data are absent; IVIG and prednisone are considered compatible, but rituximab is contraindicated.

Elderly

In patients ≥ 60 years, autoimmune cerebellar ataxia frequently presents without inflammatory MRI or CSF abnormalities (22.6% of 155 antibody-positive patients) ^[147]D5. Key syndromes include:

• LGI1 antibodies: faciobrachial dystonic seizures (FBDS) as isolated or predominant symptom in 85% of those without inflammation; all improved with immunotherapy ^[147]D5.
• IgLON5 antibodies: gait instability, bulbar symptoms, sleep dysfunction; 93% lack inflammatory markers and only 10% respond to immunotherapy ^[147]D5.
• Rapidly progressive cognitive decline occurred in 14.3%, mimicking neurodegenerative disease ^[147]D5.
• SCA27B (GAA-FGF14): neuropathy in 25% (42/170), predominantly length-dependent axonal, driven by aging and risk factors rather than expansion size; neuropathy associates with more severe disability (median SARA 12 vs 8, p = 0.0024) ^[163]B3b.

Diagnostic delay is longer in elderly patients without inflammation (median 3 vs 1 month, p < 0.005) ^[147]D5. Antibody testing (LGI1, IgLON5, CASPR2, DPPX) should be pursued even with normal MRI/CSF when characteristic syndromes are present.

Immunocompromised

Ataxia-telangiectasia (A-T) is the prototypical immunocompromised host with cerebellar ataxia. Combined immunodeficiency affects approximately two-thirds of patients, manifesting as recurrent sinopulmonary infections, bronchiectasis, and elevated IgM (hyper-IgM phenotype) in a minority ^[156]B3b. Management includes:

• Immunoglobulin replacement therapy 0.6 g/kg every 4 weeks and TMP-SMX prophylaxis ^[156]B3b.
• Surveillance for malignancy (lymphoma in 10-15% of children) and atypical infections such as lupus vulgaris, reported for the first time in A-T ^[156]B3b.
• Multidisciplinary follow-up with immunology, neurology, and pulmonary rehabilitation ^[169]C4.

In other immunocompromised states, Mycoplasma pneumoniae-associated cerebellar ataxia (parainfectious or post-infectious in 83.4%) responds to antibiotics and immunotherapy, with complete or near-complete recovery in 83.4% ^[158]B2a. MOGAD-related ataxia improves with steroids in approximately half of patients ^[159]B2a.

Pearl: In elderly patients with cerebellar ataxia and normal MRI/CSF, test for LGI1 and IgLON5 antibodies, LGI1-associated FBDS is highly treatable, while IgLON5 syndrome is immunotherapy-refractory and requires supportive care. In children, age ≥ 5 years and ataxia persisting > 3 days mandate urgent evaluation for immune-mediated or structural causes.

Beyond the unique considerations of pregnancy and special populations, the clinician's role extends to preventing disease progression through early detection of treatable causes and systematic surveillance for complications. Primary prevention of genetic cerebellar ataxias is not possible, but for autoimmune ataxias early recognition and immunotherapy may halt or slow progression. Anti-metabotropic glutamate receptor-1 (mGluR1) encephalitis, a treatable cause of subacute cerebellar ataxia, shows clinical improvement in the majority of patients after immunotherapy, with glucocorticoids being the most common first-line agent (used in 80.6% of cases) ^[183]C4. Similarly, SEZ6L2 autoimmunity is immunotherapy-responsive: among 19 reported patients, 47.4% showed partial improvement after treatment including IVIg 2 g/kg ^[138]C4. For Lambert-Eaton myasthenic syndrome (LEMS) with coexisting cerebellar ataxia, early immunotherapy produced marked neuromuscular and partial cerebellar improvement, but mortality was significantly higher in malignancy-associated cases (27.1% vs 5.3%, p = 0.04) ^[139]B2a. These observations support urgent antibody testing and early immunotherapy in subacute ataxia with red-flag features (cognitive decline, parkinsonism , ophthalmoparesis, or autonomic dysfunction).

Screening Recommendations

Genetic screening for RFC1 repeat expansions should be a priority in adults with late-onset cerebellar ataxia, especially when accompanied by chronic cough, sensory neuropathy, or bilateral vestibulopathy. In a consecutive unselected cohort of patients with late-onset ataxia, biallelic RFC1 expansions were found in 14% (95% CI 9%-22%) ^[46]B3b. Among patients with idiopathic bilateral vestibulopathy (BVP), the prevalence of biallelic RFC1 expansions was 8% (10/127) ^[171]D5. In idiopathic peripheral neuropathy , biallelic RFC1 expansions were present in 2.3% (18/788) of patients, including 6.9% of those with pure sensory neuropathy ^[179]D5. Chronic cough, sensory ataxia with use of visual control, and axonal sensory neuropathy yield the highest predictive values to discriminate RFC1-positive from RFC1-negative ataxia ^[46]B3b. For patients with early-onset or complex multisystemic ataxia, short-read genome sequencing (SR-GS) provided a genetic diagnosis in 38% of a cerebellar ataxia cohort (40/110), identifying repeat expansions and non-repeat variants in 356 ataxia genes ^[182]B2b.

Surveillance

For progressive ataxias, serial monitoring of the Scale for the Assessment and Rating of Ataxia (SARA) score quantifies progression: in RFC1 disease, the mean progression rate is 1.3 SARA points per year (95% CI 1.1-1.6), with nonlinear phases of rapid progression up to 5.5 points per year ^[46]B3b. Video-head-impulse testing (vHIT) is a potential biomarker for vestibular decline: horizontal-canal vestibulo-ocular reflex gains are significantly associated with disease duration (p = 0.031), and intra-individual declines over time were noted in 15/21 patients after a mean follow-up of 33.4 months ^[180]B2a. Surveillance for dysphagia, falls, and respiratory dysfunction (especially in SYNE1 ataxia, where premature death due to respiratory failure occurs ^[176]D5) is essential.

Vaccine Considerations

COVID-19 vaccination can rarely trigger acute CNS inflammation, including cerebellar ataxia. Among 38 patients with acute CNS inflammation within 60 days of vaccination (median 15 days), one patient was diagnosed with primary autoimmune cerebellar ataxia ^[92]C4. However, of 16 patients who received subsequent COVID-19 vaccination, 87.5% had no new or worsening neurological symptoms ^[92]C4. The benefits of vaccination against COVID-19 outweigh the risk of developing CNS inflammatory disease ^[92]C4.

Patient Education

Patients with biallelic RFC1 expansions should be counseled that chronic cough often precedes gait ataxia by a median of 16 years (range up to 36 years) and is a key early symptom ^[46]B3b. Genetic counseling is indicated for autosomal-recessive inheritance. For autoimmune ataxias, patients should be educated about relapsing symptoms and the need for prompt medical attention.

Pearl: In any adult with late-onset ataxia and chronic cough, sensory neuropathy, or vestibulopathy, genetic testing for RFC1 repeat expansions is warranted, the diagnostic yield in unselected cohorts is 14%, and a positive result changes prognosis and surveillance planning ^[46]B3b.