Early-Onset Hereditary Cerebellar Ataxias — General Approach

A heterogeneous group of genetic disorders in which cerebellar ataxia manifests before the age of 25 years, in congenital, progressive or episodic forms; a clinical guidance article focusing on classification, the European diagnostic algorithm, treatable causes and the principal recessive disorders.

Literature and regulatory information updated: 9 October 2026.

0. Initial Reference Table

1. Introduction

1.1. Definition and limits of the concept

Ataxia is impaired coordination of voluntary movements occurring with preserved strength and sensation; it manifests as an unsteady wide-based gait and may affect speech and eye movements [18, 20]. It results from lesions of the cerebellum and its pathways, the spinal cord or the peripheral sensory system, so the first clinical distinction separates cerebellar ataxia from sensory ataxia caused by loss of proprioception [33]. In children, ataxia must also be distinguished from gait instability due to motor immaturity and from muscle weakness [18]. Early-onset cerebellar ataxia (EOA) denotes a heterogeneous group of rare disorders of the central and peripheral nervous systems, often involving other organs, which manifest through cerebellar degeneration or abnormal development before the age of 25 years and are mostly autosomal recessive [1]. The age limit of 25 years is conventional and serves as an operational criterion in paediatric algorithms and European cohorts of recessive and early-onset ataxias [2, 36].

The group includes three temporal patterns that shape the entire approach. Congenital ataxia, usually non-progressive, presents with hypotonia, poor sucking and motor delay in infancy, and ataxia only becomes evident between 1 and 2 years, when the child stands and walks; a significant proportion of these children have learning difficulties [18]. Progressive ataxia results from degenerative or metabolic processes and is the most common form of ataxia in inborn errors of metabolism [20]. Intermittent or episodic ataxia suggests metabolic decompensation or, more often, a channelopathy without a biochemical marker [18, 20]. The 2026 French PNDS (French National Diagnosis and Care Protocol) defines non-progressive congenital ataxias (NPCA) as an early cerebellar syndrome that first manifests as hypotonia from birth or in the first months, with psychomotor delay, and only later as the classic cerebellar signs, in a child who continues to acquire skills; they are regarded as developmental disorders, mainly of genetic origin, and are distinguished from acquired cerebellar lesions (prenatal, perinatal or postnatal infectious or vascular sequelae) and from malformations identifiable on MRI (Joubert syndrome, rhombencephalosynapsis, Dandy-Walker malformation, tubulinopathies, muscle-eye-brain disease and multiple malformation syndromes), which fall outside that protocol [23]. The same document stresses that the boundary with some degenerative ataxias of very early onset and very slow progression, which may appear stable throughout childhood, can only be established through follow-up [23]. Ataxia is the most frequent movement disorder in inborn errors of metabolism, and movement disorders are present in approximately 70% of these patients, according to the synthesis in Saudubray’s textbook [20].

The concept of “primary hereditary ataxia” used by GeneReviews encompasses disorders in which an adult or a child’s carers seek diagnosis and treatment from a neurologist because ataxia is the principal manifestation; explicitly excluded from that scope are infantile-onset multisystem disorders, epileptic encephalopathies, cerebellar malformations (cerebellar hypoplasia, vermian hypoplasia and the Joubert spectrum), leukodystrophies, global developmental delay, inborn errors of metabolism such as peroxisomal biogenesis disorders and congenital disorders of glycosylation (CDG), primary mitochondrial disorders and multisystem diseases such as Niemann-Pick disease type C (NPC) and late-onset Tay-Sachs disease [7]. In paediatric practice, however, these exclusions constitute a substantial part of the differential diagnosis, so this article incorporates them from clinical genetics and metabolic disease textbooks and disease-specific GeneReviews [18–20].

1.2. Glossary of structural cerebellar abnormalities

Interpreting magnetic resonance imaging (MRI) requires precise terminology. Cerebellar hypoplasia is a congenitally small cerebellum resulting from deficient primary development; cerebellar atrophy is a progressive reduction in volume accompanied by increasing neurological sequelae; without serial MRI it is often difficult to distinguish them [19]. Pontocerebellar hypoplasia (PCH) denotes a group of rare genetic disorders, mostly autosomal recessive, with underdevelopment of the cerebellum and brainstem [19]. Dandy-Walker malformation combines cystic dilatation of the fourth ventricle, total or partial vermian agenesis and an enlarged posterior fossa with displacement of the tentorium; vermian aplasia is a midline defect without enlargement of the posterior fossa [19]. The molar tooth sign combines a deep interpeduncular fossa, thickened and elongated superior cerebellar peduncles and a hypoplastic or aplastic superior vermis; it is characteristic of Joubert syndrome but occurs in other ciliopathies [19]. Chiari malformations comprise types I to IV, with type IV corresponding to a form of cerebellar hypoplasia [19]. The French NPCA PNDS uses compatible radiological definitions but adds the mixed forms: atrophy as volume loss with a thin cortex and widened interfolial fissures, sometimes with T2/FLAIR hyperintensity of the cerebellar cortex or white matter; hypoplasia as a small cerebellum with normal foliation and non-widened fissures; “hypo-atrophy” when both coexist; and “hypo-dysplasia” when hypoplasia is associated with abnormal foliation or cortex; the abnormality may be focal (vermis or one hemisphere) or generalised [23]. It diverges from the Oxford Desk Reference definition on one practical point: in NPCA, atrophy may appear only on the second MRI, performed years after a normal first scan, without any clinical deterioration, so that in these children radiological atrophy does not imply increasing neurological sequelae [19, 23].

1.3. Historical overview and classification milestones

The separation between progressive and non-progressive childhood ataxias predates genetics. After Friedreich described the degenerative form in 1863, Freud described in 1893 two patients with intention tremor, nystagmus and ataxia present from birth; Batten described in 1903 eight children with congenital ataxia, hypotonia, language delay and improvement with age, and suggested that these were developmental anomalies rather than perinatal sequelae; and Ingram recorded in 1962 that 60% of patients with congenital ataxia had no history of perinatal distress and, given familial cases, proposed a genetic aetiology [23]. The modern clinical classification of hereditary ataxias and spastic paraplegias was proposed by Anita Harding in 1983, using age at onset, inheritance pattern and associated signs, before the genes had been identified [37]. That scheme gave rise to the designation early-onset cerebellar ataxia with retained tendon reflexes (EOCARR), distinguishing patients with a Friedreich-like presentation but without areflexia; an Italian population study in 1993 estimated a birth prevalence of 2.08/100 000 and a point prevalence of 1.0/100 000 [4, 22]. With gene identification, EOCARR proved to be a heterogeneous clinical category, and Friedreich ataxia itself was found to include late-onset forms with retained reflexes [7, 33].

The 1990s ushered in the molecular era of treatable recessive ataxias. In 1993, Ben Hamida and colleagues used homozygosity mapping to localise the Friedreich phenotype with selective vitamin E deficiency to chromosome 8q, and in 1995 Ouahchi and colleagues showed that the disorder results from variants in the alpha-tocopherol transfer protein gene; the 744delA variant accounted for 68% of mutant alleles in 17 families from North Africa and Italy [38, 39]. In 2000, Engert and colleagues identified SACS, with an open reading frame of 11.5 kb, as the gene responsible for spastic ataxia of Charlevoix-Saguenay (ARSACS), common in north-eastern Quebec [40]. In 2001, Moreira and colleagues showed that ataxia with oculomotor apraxia type 1 (AOA1) results from variants in the gene encoding aprataxin, a protein with an HIT domain and zinc finger [41]. Refsum disease had been described in 1946 by Sigwald Refsum in Norway as heredopathia atactica polyneuritiformis [10], abetalipoproteinaemia was known as Bassen-Kornzweig syndrome [9], and cerebrotendinous xanthomatosis (CTX) received the historical names cerebral cholesterinosis and Van Bogaert-Scherer-Epstein disease [11].

1.4. Current nomenclature and reclassifications

The growing number of genes prompted two organisational proposals. In 2018, the Movement Disorder Society’s Task Force on Classification and Nomenclature of Genetic Movement Disorders proposed naming recessive ataxias with the prefix ATX followed by the gene: 62 entities receive the single prefix because ataxia is predominant and consistent, 30 receive a double prefix (for example ATX/HSP, for hereditary spastic paraplegia) because ataxia coexists with another predominant movement disorder, and 89 disorders with complex phenotypes occasionally featuring ataxia are listed without a prefix [42]. In 2019, a Society for Research on the Cerebellum and Ataxias consensus, based on an earlier systematic review, classified 59 disorders as primary ARCA in both a clinical and a pathophysiological classification, and added 48 complex multisystem disorders with ataxia for inclusion in the differential diagnosis [3, 43]. In parallel, the genetic and phenotypic overlap between ataxias and hereditary spastic paraplegias, with genes such as SACS, SPG7, SETX and SYNE1 in both groups, led to their being considered a continuum [7, 44]. Historical numbered designations (SCAR, ARCA, AOA, SPAX), given in parentheses in this article, also remain in use.

The designations of some entities have changed recently. Saudubray’s textbook assigns “SCA4” to RUBCN, but this gene corresponds to autosomal recessive ataxia SCAR15, whereas SCA4 is associated with ZFHX3 in the 2025 GeneReviews revision [7, 20]. The Oxford Desk Reference lists “ATP6A2” among the genes for cerebellar ataxia, intellectual disability and disequilibrium syndrome (CAMRQ); the corresponding gene is ATP8A2 (CAMRQ4, OMIM 615268) [15, 18].

1.5. Entities covered

This article is organised into clinically distinct groups: treatable ataxias caused by metabolic or ion-channel defects (AVED, abetalipoproteinaemia and hypobetalipoproteinaemias, adult Refsum disease, CTX, primary coenzyme Q10 deficiency, NPC, GLUT1 deficiency, biotinidase deficiency, pyruvate dehydrogenase deficiency and episodic ataxias); early-onset degenerative recessive ataxias (ataxias with oculomotor apraxia, ARSACS, SYNE1, ANO10, SPG7, POLG and TWNK ataxias, Marinesco-Sjögren syndrome); congenital and malformation ataxias (Joubert, PCH, CAMRQ, CASK, Gillespie, CDG); and infantile-onset dominant ataxias, particularly spinocerebellar ataxia type 7 (SCA7) and infantile SCA2 [7, 18–20]. Friedreich ataxia, identified by the Oxford Desk Reference as the most common hereditary ataxia, and ataxia-telangiectasia (A-T) are addressed here only as relevant to differential diagnosis and testing strategy [18].

2. Molecular Genetics and Pathophysiology

2.1. Two molecular classes with different laboratory requirements

From a laboratory perspective, hereditary ataxias divide into nucleotide repeat expansion disorders and disorders caused by conventional variants (substitutions, small insertions and deletions, copy-number variants). Across all ages, expansions are the most common cause of hereditary ataxia and may be autosomal dominant, recessive or X-linked; in dominant forms, anticipation may be so pronounced that a child with early, severe and phenotypically different disease dies before the transmitting parent or grandparent becomes symptomatic [7]. In children and young adults, by contrast, recessive disorders caused by conventional variants predominate, except for Friedreich ataxia, which is recessive and caused in 98% of cases by GAA expansions in intron 1 of FXN [1, 18]. This distinction has immediate practical consequences: expansions require specific assays with sizing, and commercial sequencing panels without these assays do not identify them [7].

Source: summarised from Table 1 of the GeneReviews overview [7, 18].

According to the same source, the most frequent non-repeat hereditary ataxias encompass recessive and dominant genes with overlap between ataxia and spasticity; several dominant genes produce congenital non-progressive forms and may arise de novo [7, 44].

Source: Table 2 of the GeneReviews overview [7].

For non-progressive congenital ataxias, the 2026 French PNDS (French National Diagnosis and Care Protocol) lists more than 30 genes whose proteins have very diverse functions: glutamatergic transmission (GRM1, SPTBN2, GRID2), ion channels and calcium signalling (CACNA1A, ITPR1, CA8, KCNJ10, KCNC3), mitochondrial function (SACS, PMPCA, COQ8A, formerly ADCK3), neuronal migration (VLDLR), chaperones (SIL1), autophagy (RUBCN, SNX14) and mechanisms still unknown (WDR81, ABCB7, ATCAY, ATP2B3) [23]. Several dominant genes produce, depending on the family and the variant, either a congenital ataxia or an adult-onset ataxia (ITPR1, SPTBN2, KCNC3, CACNA1A), and some cause recessive or dominant forms depending on the nature of the variant (SPTBN2, ITPR1, GRID2) [23]. Microdeletions removing all or part of a gene, and more rarely microduplications, are rare causes [23]. The metabolic causes with a biochemical marker, PMM2-CDG and GLUT1 deficiency, which until a few years ago were the only identifiable ones, account for no more than 1% of congenital ataxias, but they must be excluded because of genetic counselling and specific treatment [23].

Source: Table 1 of the PNDS for non-progressive congenital ataxias, a non-exhaustive list that excludes the Joubert syndrome genes; the progressive ataxia genes are included because onset may mimic a congenital ataxia and progression only becomes evident later [23].

2.2. Pathophysiological mechanisms by group

Vitamin E deficiency. Alpha-tocopherol transfer protein (α-TTP), encoded by TTPA on chromosome 8, binds 2R-alpha-tocopherols stereoselectively, and this selectivity is lost with some variants [8, 18]. Pathogenic variants act through loss of function, leaving plasma vitamin E very low despite normal lipids [8]. The same neurological deficiency arises through another pathway in disorders lacking apolipoprotein B (apoB)-containing lipoproteins: in abetalipoproteinaemia, deficiency of the large subunit of microsomal triglyceride transfer protein (MTTP) reduces apoB-containing lipoproteins and abolishes chylomicrons, with marked hypocholesterolaemia and deficiency of all fat-soluble vitamins [9, 45]. GeneReviews lists the genetic causes of vitamin E deficiency treatable by replacement: ANGPTL3 (familial combined hypolipidaemia, AR), APOB (familial hypobetalipoproteinaemia, AR when homozygous or compound heterozygous), MTTP (abetalipoproteinaemia) and TTPA (AVED) [7]. In chylomicron retention disease, caused by SAR1B (a GTPase), apoB48 is reduced and triglycerides are normal [9, 45].

Phytanic acid. Phytanic acid is exclusively dietary in origin and accumulates according to diet and age when phytanoyl-CoA 2-hydroxylase (PHYH) is deficient, accounting for more than 90% of adult Refsum disease; in fewer than 10%, the cause is PEX7 variants, producing a milder form and allelic to rhizomelic chondrodysplasia punctata type 1 [10, 46].

Bile acid synthesis. In CTX, biallelic CYP27A1 variants inactivate mitochondrial sterol 27-hydroxylase. 5β-cholestane-3α,7α,12α-triol accumulates and is diverted through an alternative C25 pathway into bile alcohols excreted as urinary glucuronides; feedback inhibition of cholesterol 7α-hydroxylase is lost, and cholestanol and cholesterol accumulate in almost all tissues, with cerebral xanthomas preferentially located in cerebellar white matter [47]. The variants cause loss of function, many affecting splicing [11].

Coenzyme Q10. Primary CoQ10 deficiency results from biallelic variants in one of ten biosynthesis genes (COQ2, COQ4, COQ5, COQ6, COQ7, COQ8A [ADCK3], COQ8B [ADCK4], COQ9, PDSS1, PDSS2) and is classified among respiratory chain disorders; a case caused by a de novo contiguous-gene deletion including COQ4 has been described [12]. Secondary CoQ10 deficiencies occur in respiratory chain disorders, multiple acyl-CoA dehydrogenase deficiency and AOA1, and only molecular testing distinguishes them from primary deficiency [7, 12]. Transmission is nuclear and autosomal recessive; the chapter in the Brazilian book Síndromes Genéticas: Reconhecer para Tratar describes it as “linked to mitochondrial DNA” and gives incorrect sibling recurrence risks, errors that are not reproduced here [12, 48].

Lysosomes. NPC is an autosomal recessive, progressive, life-limiting neurovisceral lysosomal disorder, caused in at least 95% of cases by NPC1 variants and in the remainder by NPC2 [29]. Other lysosomal disorders with childhood- or adolescent-onset ataxia include neuronal ceroid lipofuscinoses (CLN2, 3, 5, 6, 7, 8, 10 and 14), Gaucher disease type 3, GM2 gangliosidoses, Krabbe disease, sialidosis and mannosidoses [20].

DNA repair. Aprataxin participates in single-strand break repair; AOA1 shares early ataxia, oculomotor apraxia and cerebellar atrophy with A-T, but not immunodeficiency, chromosomal instability or radiosensitivity [18, 41]. Senataxin (SETX) is a helicase, and missense variants in the helicase domain are associated with a less severe phenotype [13]. PNKP (AOA4) and MRE11 (ataxia-telangiectasia-like disorder) complete this group in the differential diagnosis [13]. Biallelic XRCC1 variants cause progressive cerebellar ataxia with oculomotor apraxia and axonal neuropathy [49].

Mitochondria. Ataxias caused by defects in mitochondrial DNA homeostasis or replication include Friedreich ataxia (FXN), the polymerase gamma spectrum (POLG; sensory ataxic neuropathy with dysarthria and ophthalmoparesis, SANDO, and mitochondrial recessive ataxia syndrome, MIRAS) and infantile-onset spinocerebellar ataxia (IOSCA, TWNK, formerly C10orf2); disorders of mitochondrial DNA itself with progressive ataxia include myoclonic epilepsy with ragged-red fibres (MERRF), mitochondrial encephalomyopathy with lactic acidosis and stroke-like episodes (MELAS), neuropathy, ataxia and retinitis pigmentosa (NARP), maternally inherited Leigh syndrome (MILS) and Kearns-Sayre syndrome [20].

Complex lipids, cellular trafficking and translation. Saudubray’s textbook identifies emerging metabolic categories: complex lipid synthesis and remodelling (ABHD5 in Chanarin-Dorfman syndrome; PNPLA6 in Laurence-Moon, Boucher-Neuhäuser and Gordon Holmes syndromes; ABHD12 in polyneuropathy, hearing loss, ataxia, retinitis pigmentosa and cataract syndrome (PHARC); CYP7B1; ceramide synthases with myoclonic epilepsy), cellular trafficking and autophagy (SNX14, WDR45, SIL1, TANGO2, VPS13D, SCYL1, KIF1C), aminoacyl-tRNA synthetases and ribosome biogenesis (POLR3A, POLR3B, POLR1C), the latter with cerebellar and brainstem atrophy and hypomyelination [20]. Hypogonadotrophic hypogonadism is a key sign of the PNPLA6 spectrum [20].

Cerebellar development. Joubert syndrome is a genetically heterogeneous autosomal recessive ciliopathy [19]. Pontocerebellar hypoplasias involve genes for neuronal development and survival and RNA processing, including TSEN54 [16, 19]. VLDLR deficiency disrupts the reelin signalling pathway and produces CAMRQ1; microlissencephaly with severe cerebellar abnormality may be caused by RELN, and tubulinopathies are a major cause of lissencephaly with cerebellar hypoplasia [15, 19]. In Gillespie syndrome, a restricted repertoire of de novo ITPR1 variants acts through a dominant-negative effect [19]. Glycosylphosphatidylinositol (GPI) anchor defects, involving more than 150 anchored proteins and at least 27 genes, include PIGG and PIGT with cerebellar hypoplasia, while PMM2-CDG (CDG-Ia), the most common CDG type, features cerebellar atrophy [19].

Ion channels. Episodic ataxias are channelopathies without a metabolic marker: KCNA1 (EA1, with interictal myokymia), CACNA1A, CACNB4 and SLC1A3 (with interictal nystagmus), PRRT2 (with attacks of dystonia or chorea) and ATP1A3, whose spectrum includes cerebellar ataxia, areflexia, pes cavus, optic atrophy and sensorineural hearing loss syndrome (CAPOS) and relapsing encephalopathy with cerebellar ataxia [17, 20].

2.3. Inheritance patterns, penetrance and expressivity

Autosomal recessive inheritance predominates in early-onset ataxia, particularly with consanguinity or affected siblings [1, 18]. Dominant inheritance manifests through anticipation in expansion disorders and de novo variants in dominant genes with congenital forms (ITPR1, KCNC3, AFG3L2); parental mosaicism has been described in SCA29 [7]. X-linked transmission is represented by CASK, whose form with microcephaly and pontocerebellar hypoplasia predominantly affects girls and is more severe in boys, and by ornithine transcarbamylase (OTC) deficiency, in which symptomatic female carriers may have intermittent ataxia with protein intolerance [18, 19]. In abetalipoproteinaemia, biochemical penetrance is 100% in biallelic carriers, but symptoms vary and increase with age [9]. In CTX, intrafamilial variability is considerable [11], and in AVED age at onset and course are more uniform within a family, although they vary between siblings [8]. Among non-progressive congenital ataxias with a molecular diagnosis at the French reference centre, the variant arose de novo in more than half of cases (approximately 60% in the distribution by mode of transmission); dominant inheritance transmitted by a parent mainly involved CACNA1A, ESRRG, SPTBN2, ITPR1 and KCNC3, and the most represented recessive forms were those due to GEMIN5, SACS and COQ8A [23]. Parental mosaicism, germline or somatic, in an asymptomatic parent or one with minimal signs, is particularly frequent with KIF1A and has also been described with ITPR1, SPTBN2, DAGLA, SLC2A1 and STXBP1; the proband may also be mosaic, with a milder phenotype [23].

2.4. Variant spectrum, nomenclature and founder variants

Variants cited in the sources were normalised in VariantValidator to the MANE Select transcript (Matched Annotation from NCBI and EMBL-EBI; GRCh38), using Human Genome Variation Society (HGVS) nomenclature, on 9 October 2026. Several GeneReviews use earlier transcript versions (for example NM_000370.2 for TTPA, NM_207346.2 for TSEN54, NM_015046.6 for SETX), without changes to the numbering of the variants listed.

The equivalence between the historical SACS designation “6594delT” and the current variant c.8844del is that provided by GeneReviews; the old numbering could not be reproduced in VariantValidator, which uses current transcripts [14]. For SETX p.Glu65Lys, TTPA p.His101Gln and the Hutterite VLDLR deletion, the sources do not provide the c. nomenclature needed for normalisation.

2.5. Genotype–phenotype correlations

In AVED, variants affecting highly conserved amino acids (p.Arg59Trp, p.Asp64Gly, p.Glu141Lys, p.Leu183Pro, p.Arg221Trp) and truncating variants are associated with early onset, by approximately 10 years, and a severe course; missense variants in non-conserved or semi-conserved amino acids (p.His101Gln, p.Ala120Thr, p.Arg192His, p.Gly246Arg) are associated with later onset and a mild course [8]. p.Arg134Ter (c.400C>T), c.486delT, c.513_514insTT and c.530_531delAGinsGTAAGT have early onset; p.Ala120Thr (c.358G>A) is associated with onset after 20 years; c.744del confers a slightly greater cardiomyopathy risk, and onset varied from 3 to 12 years within the same family [8]. In AOA2, in a series of 90 patients, helicase-domain missense variants were associated with a less severe phenotype [13]. In COQ8A ataxia, the largest cohort (59 patients, 44 variants) showed more frequent multisystem involvement with missense variants (82–93%) than with biallelic loss of function (53%) [32]. In ANO10 ataxia, the Romani c.1150_1151del variant is associated with onset 14 years earlier, according to the MDSGene database of 94 patients [31]. In primary CoQ10 deficiency, treatment response correlates with the gene and variant severity: in Yu’s COQ4 series, responders had two missense variants and non-responders a null variant [12]. There is no reliable genotype–phenotype correlation in abetalipoproteinaemia, Refsum disease or CTX; in Refsum disease, variation depends mainly on diet [9–11].

For non-progressive congenital ataxias, the French PNDS brings together genotype–phenotype correlations that still rest on small series but are useful for guiding analysis: abnormal eye movements suggest CACNA1A (paroxysmal tonic upgaze), DAGLA or SLC2A1; ITPR1 ataxia is usually pure, with significant motor impairment contrasting with a good cognitive level; the association with optic atrophy and neuropathy points to KIF1A; ESRRG is the most frequent cause of congenital ataxia without intellectual disability and with normal or near-normal MRI; early ataxia without hypotonia in the first months, with brainstem signal changes, spasticity, neuropathy or prominent myelinated fibres on fundoscopy, points to SACS; and recurrence within a sibship favours a recessive form, with GEMIN5 the first gene to consider on frequency grounds, after parental mosaicism and X-linked forms in two affected brothers have been considered [23].

2.6. Allelic disorders, heterozygosity and modifiers

Several early-onset ataxia genes produce distinct allelic phenotypes that matter when interpreting sequencing results. CACNA1A causes EA2 through conventional variants, SCA6 through CAG expansion, and familial hemiplegic migraine type 1 [7, 26]. SETX causes AOA2 in the biallelic state and autosomal dominant juvenile amyotrophic lateral sclerosis type 4 through variants such as p.Thr3Ile, p.Leu389Ser and p.Arg2136His [13]. ITPR1 causes adult-onset SCA15/16, congenital non-progressive SCA29 and Gillespie syndrome (iris hypoplasia, hypotonia and ataxia) [7, 19]. KCND3 is allelic to Brugada syndrome; SYNE1 produces adult recessive ataxia, a childhood multisystem form and arthrogryposis, and, in heterozygosity, Emery-Dreifuss muscular dystrophy type 4 [7, 31]. PEX7 causes Refsum disease and rhizomelic chondrodysplasia punctata type 1; PNKP causes AOA4, microcephaly with seizures and Charcot-Marie-Tooth disease type 2B2 [10, 31]. A small number of patients of Japanese ancestry with cerebellar-type multiple system atrophy have biallelic or heterozygous COQ2 variants [7]. Heterozygous carriers of recessive ataxias do not develop ataxia, with relevant exceptions: ATM heterozygotes have increased cancer risk [7]; in heterozygous familial hypobetalipoproteinaemia, vitamin E is recommended to prevent adult neuropathy [45]; and TTPA heterozygotes may have moderately low vitamin E, insufficient to identify carrier status [8].

3. Clinical Spectrum and Manifestations by System

3.1. Cerebellar signs and initial manifestations

The typical cerebellar examination reveals a wide-based, staggering gait with difficulty walking in tandem; truncal instability when sitting unsupported; dysmetria and terminal tremor on finger-to-nose and heel-to-shin testing; dysdiadochokinesia; slow, slurred, monotonous or scanning dysarthria with changes in pitch and volume; and oculomotor abnormalities such as saccadic intrusions, horizontal or vertical nystagmus and hypermetric saccades [7, 18]. The initial manifestation is almost always slowly progressive gait instability with unexpected falls; imbalance described as dizziness, hand clumsiness or tremor, dysarthria or choking may occur, and diplopia rarely; initially, complaints may be intermittent, appearing at the end of the day, with fatigue or alcohol, before becoming constant [7]. Non-cerebellar abnormalities may mimic or worsen ataxia: lesions of frontal and parietal motor pathways, central or peripheral vestibular dysfunction, posterior column or peripheral sensory loss, weakness, rigidity, spasticity and joint disease [7].

Non-ataxic signs have greater diagnostic value than ataxia itself. The Inventory of Non-Ataxia Signs (INAS) standardises their recording; it was validated in 140 patients with spinocerebellar ataxia for inter-rater agreement and in 28 for intra-rater agreement, with good reproducibility but insufficient responsiveness to serve as a primary endpoint [24]. The Scale for the Assessment and Rating of Ataxia (SARA) quantifies severity on a score from 0 to 40; developed in two trials with 167 and 119 patients, it took a mean of 14.2 ± 7.5 minutes to administer, with inter-rater intraclass correlation of 0.98 and test–retest correlation of 0.90 [51]. In children, SARA scores depend on age: in 156 typically developing children aged 4 to 16 years, scored by 22 experts from 15 institutions, age was the only variable associated with score, and inter-observer agreement also varied with age, so age-specific reference values are required [52].

3.2. Congenital ataxias and cerebellar malformations

Joubert syndrome. All patients have hypotonia; 75% can sit, at a mean age of 19 months, and 50% walk, at a mean of 4 years, with highly variable developmental delay [19]. Episodic hyperpnoea or apnoea occurs in 50–75%, particularly in infancy; facial features include high rounded eyebrows, a broad nasal bridge, mild epicanthic folds, anteverted nares, an open triangular mouth with irregular tongue protrusion and low-set ears; retinal dysplasia, colobomas, nystagmus, strabismus, ptosis and oculomotor apraxia, microcystic renal disease and sometimes polydactyly occur [19]. Related cerebro-oculo-renal syndromes (Arima, Senior-Løken, COACH) combine retinal dysplasia and cystic dysplastic kidneys, and orofaciodigital syndromes types 2 (Mohr) and 6 (Váradi, with a Y-shaped central metacarpal) share the molar tooth sign [19].

Pontocerebellar hypoplasias. PCH disorders have a neurodegenerative phenotype with microcephaly and severe delay; feeding and swallowing difficulties are typical, and weakness resembling spinal muscular atrophy may occur [19]. TSEN54 forms (PCH2, PCH4 and PCH5) differ mainly in life expectancy. In PCH2, death usually occurs before 10 years; patients have generalised clonus, uncoordinated sucking and swallowing, absent voluntary motor development, cognitive impairment, cortical blindness, epilepsy in approximately half and a risk of rhabdomyolysis during severe infections; PCH4 presents in the newborn with seizures, arthrogryposis, clonus and central respiratory failure, with neonatal death [16]. Pregnancy is usually unremarkable, the newborn has no dysmorphism, and head circumference is always below the 50th centile, progressing to microcephaly (< −2 standard deviations) within the first year [16]. In a cohort of 33 p.Ala307Ser homozygotes, born to non-consanguineous parents and surviving up to 11 years, pyramidal signs were recorded in 62%, seizures in 82%, dystonic attacks in 33%, feeding difficulties in 100%, gastro-oesophageal reflux in 73%, sleep disturbance in 96%, apnoea in 67% and recurrent infections in 52% [16].

CASK-linked microcephaly with pontocerebellar hypoplasia. This features marked congenital and postnatal microcephaly, severe intellectual disability and disproportionate pontine and cerebellar hypoplasia; it predominantly affects females, while males have more severe disease and reduced survival [19].

CAMRQ and VLDLR cerebellar hypoplasia. CAMRQ 1 to 4 are congenital non-progressive ataxias with cerebellar hypoplasia, with or without gyral simplification, caused by biallelic variants in VLDLR, CA8, ATP8A2 and WDR81; they enter the differential diagnosis of ataxic cerebral palsy [15, 18, 19]. In CAMRQ1, ataxia is predominantly axial; walking occurs very late, often after 6 years, or never becomes independent; moderate to profound intellectual disability is universal, with dysarthria, strabismus in most and seizures in 40% of Hutterite patients, less often in others; Turkish patients exhibited quadrupedal locomotion [15]. MRI shows hypoplasia of the inferior vermis and hemispheres, gyral simplification and a small brainstem, particularly the pons; life expectancy does not appear reduced [15].

Other forms. Gillespie syndrome combines bilateral iris hypoplasia (partial aniridia), congenital hypotonia, non-progressive ataxia, progressive cerebellar atrophy and mild to moderate delay [19]. Macrocephaly with progressive cerebellar hypoplasia suggests SNX14 or PIGT [19]. PMM2-CDG combines cerebellar atrophy and encephalopathy with hepatic, renal, cardiac, skeletal and genital involvement, and lipodystrophy with fat pads [19]. In PMM2-CDG, ataxia may be recognised from 7 months, with cerebellar and pontine atrophy on MRI, which may appear particularly after episodes of acute deterioration; reduced nerve conduction velocity in the lower limbs may manifest at 6–8 months and tendon reflexes may disappear at around 1 year; between 3 and 12 years ataxia and intellectual disability (IQ reported between 40 and 60) predominate, with stroke-like episodes during febrile infections, and few patients achieve walking; mortality in infancy and early childhood is 15–20%, mainly from infection, and some less severely affected patients are diagnosed in adulthood because of cerebellar ataxia [53]. Congenital non-progressive forms caused by ITPR1 (SCA29) and KCNC3 (SCA13) may be de novo [7]. In congenital non-progressive metabolic ataxias (CDG, succinic semialdehyde dehydrogenase deficiency, ITPR signalling defects), ataxia may even improve over time [20]. “Ataxic cerebral palsy” is a descriptive designation covering heterogeneous patients with frequent global delay and hypotonia; it is a diagnosis of exclusion requiring review of the birth history and high-resolution MRI before definitive counselling [18]. Angelman syndrome enters this differential diagnosis, with very marked language delay, jerky movements, hand flapping and frequent laughter [18].

Non-progressive congenital ataxias (NPCA). The cerebellum grows and matures intensively after birth, beyond 2 years of age, and the clinical signs follow this maturation: the classic cerebellar signs only become manifest after the first or second year and are preceded by hypotonia responsible for psychomotor delay [23]. The hypotonia, global, axial and of the limbs, without obvious weakness, appears from birth or in the first weeks or months; tendon reflexes are present, although difficult to elicit, and the child progresses steadily but behind their age [23]. At the end of the first year or during the second, tone improves and the cerebellar syndrome appears: oscillations of the head and trunk when sitting, with an absent or delayed lateral protective reaction of the arms, instability when standing and a staggering broad-based gait, which does not worsen with eye closure and causes many falls; independent walking is acquired at very variable ages, sometimes after 10 years, and some children never acquire it [23]. Dysmetria, hypermetria and intention tremor frequently accompany the static ataxia, fine grasp is delayed and may remain impaired, and the child is described as slow, clumsy and easily fatigued; language delay with dysarthria is frequent, and hypotonia may persist for a long time [23]. Oculomotor abnormalities related to the cerebellar syndrome are very frequent (irregular or saccadic slow pursuit, slow or dysmetric saccades, nystagmus typically downbeating, gaze-evoked nystagmus, strabismus) and are mostly described after 18 months; oculomotor recording is not part of routine practice in children [23]. Depending on the gene, there may be associated epilepsy (particularly with CACNA1A, often drug-responsive), pyramidal syndrome, dystonia, peripheral neuropathy, intellectual disability of variable degree, autism spectrum disorder, attention deficit hyperactivity disorder, episodes of hemiplegic migraine or episodic ataxia, hearing loss and ophthalmological abnormalities [23]. Severity can be quantified with the SARA from 4 years, repeated to document improvement and interpreted in the light of developmental effects on the score [23, 52].

Some abnormal eye movements point to specific genes. Paroxysmal tonic upgaze, in clusters of attacks lasting a few seconds with compensatory neck flexion (“eyes up, chin down”), may be the first manifestation of a CACNA1A variant, from 2–3 months, and herald episodes of episodic ataxia type 2 and, later, chronic ataxia; brief attacks of multidirectional eye movements accompanied by head movements in the same direction, before 6 months, suggest GLUT1 deficiency (SLC2A1) and should be sought in the history, because they may disappear over time [23]. Rarer ophthalmological manifestations include strabismic or refractive amblyopia, retinopathy with poor visual behaviour, the aniridia of Gillespie syndrome (ITPR1) and optic atrophy [23]. Orthopaedic complications result from hypotonia and the other neurological signs: flat valgus feet (pes planovalgus), kyphoscoliosis and hip dysplasia [23]. Cerebellar MRI abnormalities are rarely gene-specific (exceptions such as VLDLR and SACS), their extent does not correlate with clinical severity in most cases, and some children have cerebellar imaging abnormalities without clinical ataxia [23]. In adulthood, the improved ataxia may persist to a variable degree, with slowness, fatigability, clumsiness, visual difficulties and affective changes, and intellectual disability, when present, becomes the main disability [23].

3.3. Treatable metabolic ataxias: untreated natural history

AVED. Without treatment, onset usually occurs between 5 and 15 years; in a series of 132 North African patients, it ranged from 2 to 37 years [8, 54]. The presentation closely resembles Friedreich ataxia: gait ataxia, hand clumsiness, distal proprioceptive and vibration loss, dysdiadochokinesia, Romberg sign, head titubation, areflexia and Babinski sign, sometimes with reduced visual acuity due to macular degeneration or pigmentary retinopathy [8, 18]. Head tremor is typical of vitamin E deficiency [20]. Most untreated patients become wheelchair-dependent between 11 and 50 years; psychotic episodes and intellectual decline occur in some, while cervical or upper-limb dystonia is rare [8, 55].

Source: El Euch-Fayache 2014, as reported in GeneReviews [8, 54]. In another series, 8 of 11 patients had reduced visual acuity [8].

Abetalipoproteinaemia. The classic form presents from birth or in infancy with growth faltering, severe diarrhoea, vomiting, steatorrhoea and hepatomegaly [9, 45]. Without treatment, vitamin E deficiency produces, from the first or second decade, progressive loss of tendon reflexes and vibration and proprioceptive sensation, weakness, dysarthria and a wide-based “Friedreich-like” ataxia, with a high-stepping gait in young adults [9]. Atypical retinal pigmentation is acquired, many patients remain asymptomatic until adulthood (night and colour vision), progressive scotomas develop and blindness is possible without treatment; ptosis and ophthalmoplegia due to cranial nerve demyelination, and corneal ulcers due to vitamin A deficiency, are rare [9]. Hepatic involvement is often stable, but steatosis may rarely progress to steatohepatitis, fibrosis and cirrhosis; acanthocytosis, anaemia, haemolysis, a low erythrocyte sedimentation rate and coagulopathy from vitamin K deficiency occur; cardiomegaly is rare and late [9]. Historically, without high-dose fat-soluble vitamins, survival rarely extended beyond the third decade because of severe neuromyopathy and respiratory failure [9]. Chylomicron retention disease causes malabsorption with chronic diarrhoea, vomiting and distension, and, in children and adults, areflexia, ataxia, myopathy and sensory neuropathy, with elevated creatine kinase (CK) and invariable vitamin E deficiency [45].

Adult Refsum disease. Patients are asymptomatic at birth; the usual presentation is in late childhood with progressive night blindness, declining vision and anosmia; after 10 to 15 years or more, hearing loss, ataxia, demyelinating polyneuropathy, ichthyosis, fatigue and cardiac conduction abnormalities appear [10, 46]. Onset ranges from 7 months to over 50 years, mostly between 10 and 20 years, and early onset does not imply a worse prognosis for survival [10]. The complete triad of retinitis pigmentosa, cerebellar ataxia and chronic polyneuropathy is rarely seen in a single patient; ataxia is relatively late and disproportionate to weakness and sensory loss [10, 46]. An acute form, with polyneuropathy, weakness, ataxia, sudden visual and hearing deterioration, ichthyosis, arrhythmias and elevated transaminases and bilirubin, triggered by weight loss, stress, trauma or infection, is distinguished from a chronic form dominated by retinitis pigmentosa [10]. In 23 patients, the delay between the first ophthalmological assessment and diagnosis ranged from 1 to 28 years, with a mean of 11 [10]. Arrhythmias and cardiomyopathy are frequent causes of death [10].

Source: Wierzbicki 2002, as reported in GeneReviews [10].

Cerebrotendinous xanthomatosis. Mean age at diagnosis is approximately 35 years, with a diagnostic delay of around 16 years; this is a paediatric disorder usually diagnosed in adulthood [47]. Cholestatic jaundice in infancy may be the first manifestation, improving spontaneously; chronic diarrhoea from childhood and cataracts follow, the latter being the first finding in approximately 75% of patients and often appearing in the first decade, although in 25% only after 40 years [11, 47]. Intellectual disability may appear in the first decade; most, however, have normal intellect until puberty, and more than half develop intellectual disability or dementia in the third decade [11, 47]. Pyramidal and cerebellar signs are almost invariable between 20 and 30 years; tendon xanthomas become detectable in the second decade and usually in the third or fourth, particularly in the Achilles tendon [11, 47]. In a review of 79 patients, extrapyramidal movement disorders (parkinsonism more often than dystonia, myoclonus and postural tremor) began at a mean age of 40 ± 12 years and were the presenting feature in 18% [11]. Without treatment, death occurs between 30 and 60 years from neurological dysfunction or acute myocardial infarction [47].

Source: GeneReviews [11].

Primary CoQ10 deficiency. Age at onset ranges from birth to the seventh decade, and the spectrum extends from fatal neonatal encephalopathy to a late phenotype resembling multiple system atrophy (COQ2) [12]. The ataxic form is associated mainly with COQ8A: muscle weakness and exercise intolerance between 18 months and 3 years, followed by predominant cerebellar ataxia with severe cerebellar atrophy, following a progressive or apparently self-limiting course, isolated or accompanied by intellectual disability, epilepsy and stroke-like episodes [12]. In the cohort of 59 patients with COQ8A ataxia, cerebellar atrophy was universal; cerebral atrophy or T2 hyperintensity in the dentate nucleus or pons occurred in 28%, cognitive impairment in 49%, hyperkinetic disorders in 41% (dystonia and myoclonus may be the presenting feature), epilepsy in 32% and exercise intolerance in 25%, with mean progression of 0.45 SARA points per year [32]. Steroid-resistant nephrotic syndrome, a hallmark of primary deficiency in PDSS1, PDSS2, COQ2, COQ6 and COQ8B, has not been described in COQ4, COQ8A or COQ9; ataxia also occurs in COQ4, PDSS1 and PDSS2 [12].

Niemann-Pick type C. The late-infantile form (2–6 years) presents with clumsiness, impaired gait and fine motor skills, speech delay, a history of neonatal cholestasis, seizures, variable visceromegaly, vertical supranuclear saccadic palsy followed by vertical gaze palsy, gelastic cataplexy and hearing loss; the juvenile form (6–16 years) presents with cognitive and school difficulties, clumsiness, falls, progressive ataxia, dystonia, seizures and vertical gaze palsy [29]. In the International Niemann-Pick Disease Registry (INPDR; 203 patients from 6 European countries, 2014–2019), mean age at diagnosis was 11.2 years (standard deviation 14.2); among 168 patients with neurological disease, 24.2% had early-infantile, 26.4% late-infantile, 23.0% juvenile and 20.8% adult forms, and 5.6% had a rapidly fatal neonatal systemic form; cognitive impairment (78.5%), dysarthria and ataxia (75.9%), vertical gaze palsy (70.9%) and dysphagia (69.6%) were the most frequent symptoms [29]. In a UK cohort (146 patients), the proportions were 5% early-infantile, 35% late-infantile, 29% juvenile and 17% adolescent or adult [29].

3.4. Early-onset degenerative recessive ataxias

Sources: [7, 13, 14, 18, 20, 31, 33].

AOA2 warrants more detailed description because of its frequency in Europe. It develops after a period of normal development; oculomotor apraxia was found in 51% of 90 patients of various origins, 20% in Italy and 32% of 19 Algerian patients, and staircase hypometric saccades are more reliable than compensatory head movements; convergent strabismus occurred in 37% of Algerian patients and 12.3% of the series of 90 [13]. Tremor was described in 57% of French Canadian patients; mild cognitive impairment occurs in some, without severe intellectual disability or dementia, and there is no heart disease, cancer predisposition or immunodeficiency [13]. Outside the nervous system, premature menopause, ovarian failure, hypogonadotrophic hypogonadism, polycystic ovary syndrome and dermatofibrosarcoma protuberans have been described [13]. In a French series of 158 families with progressive non-Friedreich recessive ataxia, AOA1 accounted for approximately 5% [31].

In ARSACS, neuropathy may be the first sign, and a thickened peripapillary retinal nerve fibre layer on optical coherence tomography (OCT) is a useful marker when hypermyelinated fibres are not visible on fundoscopy [14]. In ANO10 ataxia, a cohort of 10 patients recorded bradykinesia in 7, mild vertical gaze paresis in 5, pes cavus in 4, sphincter disturbance in 3 and executive dysfunction with a normal Mini-Mental State Examination; in an Italian cohort of adult sporadic or recessive ataxia, ANO10 explained 10 of 80 cases (10%) [31]. SYNE1 ataxia explains 5.3–6% of recessive ataxias beginning before 40 years that remain unexplained after excluding Friedreich ataxia [31].

Friedreich ataxia and A-T serve as reference points for differential diagnosis. In Friedreich ataxia, mean age at onset is 15.5 ± 8 years (2–51), loss of walking occurs approximately 15 years after onset and death, at a mean of 37.5 years, is mainly due to cardiomyopathy [18]. In A-T, ataxia is evident from approximately 1 year of age and telangiectasias appear between 4 and 8 years [18].

3.5. Acute, intermittent and episodic ataxias

Acute or recurrent ataxia, often with behavioural change and triggered by fever or infection, should suggest inborn errors of small-molecule metabolism: late-onset maple syrup urine disease (MSUD) and organic acidurias, with ketoacidosis that may mimic diabetic ketoacidosis; late-onset urea cycle defects (OTC and argininosuccinate synthetase); GLUT1 deficiency, in which intermittent ataxia or gait dyspraxia may be the sole finding, worse before meals and during fasting; mild biotinidase deficiency; Hartnup disease; and pyruvate dehydrogenase (PDH) deficiency, with acute ataxia, peripheral neuropathy and moderate hyperlactataemia, sometimes only postprandial, with a normal lactate/pyruvate ratio and no ketosis [18, 20]. In adults, fever-triggered acute ataxia also occurs in respiratory chain disorders and SLC19A3 deficiency [56].

Source: Table 1.7 in Saudubray’s textbook, with the gene designation in the last row completed from the ATP1A3 GeneReviews [17, 20].

Episodic ataxias (EA1 to EA8) are autosomal dominant, potentially treatable ion-channel disorders; CACNA1A EA2 is allelic to SCA6 and familial hemiplegic migraine, and SCA6 may begin with ataxic attacks [7, 33]. SCA27B caused by an FGF14 GAA expansion has a median onset at 60 years and an episodic component, but childhood onset has been described: a 53-year-old woman with progressive cerebellar ataxia since childhood, with negative Friedreich, common SCA and exome tests, had a heterozygous (GAA)302 expansion and bilateral dentate nucleus hyperintensities on MRI [57, 58]. A study of 85 patients with undiagnosed early-onset cerebellar ataxia found no association with the FGF14 GAA expansion [59].

3.6. Infantile-onset dominant forms

Some dominant spinocerebellar ataxias may begin in childhood, particularly SCA7, whose repeat is highly unstable and may manifest in a child before the parent shows signs; infantile SCA2 and DRPLA belong to this group, and epilepsy is a warning sign for SCA10, infantile SCA7 and DRPLA [7, 18]. In older children, the Oxford Desk Reference recommends considering SCAs and DRPLA [18]. FXTAS, the most common X-linked ataxia, affects approximately 45% of male premutation carriers over 50 years and fewer than 20% of females, and onset before 50 years is rare; its relevance to early-onset ataxia lies in family history, when a grandfather has late ataxia and a grandchild has intellectual disability [7, 33].

3.7. Cognition, behaviour and other neurological comorbidities

Cerebellar involvement is associated with cerebellar cognitive affective syndrome, affecting executive function, language, visuospatial processing and emotional regulation, and quantifiable using the CCAS scale [7, 60]. The Groningen group described the frequency of dystonia and myoclonus with epilepsy as comorbidities of early-onset ataxia, with shared molecular pathways, and studied cognition in patients with early-onset ataxia and dystonia [61–63]. Distinguishing early-onset ataxia, developmental coordination disorder and central hypotonia is clinically difficult in young children and has been the subject of a specific phenotypic study [64]. In non-progressive congenital ataxias, the neurocognitive profile ranges from mild to moderate intellectual disability to normal intelligence; even without intellectual disability there are learning difficulties, attributed to dysfunction of the dentato-thalamo-cortical circuit, and the cerebellar cognitive affective syndrome described by Schmahmann in 1998 includes impairments of attention without hyperactivity, planning, abstract reasoning, visuospatial skills, fluency and verbal memory, with difficulty in managing emotions [23]. The CCAS scale quantifies this pattern in adults; in children the same domains are explored with standardised tests, taking into account slowness and fatigability during assessment [23, 60]. Some children also have autism spectrum disorder or psychiatric disorders (depression, hallucinations, anxiety), which warrant child psychiatric or psychiatric assessment [23].

3.8. Progression and natural history in cohorts

In primary hereditary ataxias, particularly adult-onset forms, walking aids become necessary 5 to 10 years after onset and wheelchair use later; choking and falls, with head injuries and hip fractures, are common causes of morbidity and death, while aspiration pneumonia, urinary tract infections and pressure ulcers predominate in late stages [7]. The general GeneReviews statement that these patients rarely live more than 25 years after onset does not apply to treatable ataxias or congenital non-progressive forms [7]. The PREPARE consortium collected 1637 SARA assessments in 884 patients with recessive or early-onset ataxia, 370 of whom had 2 to 8 longitudinal assessments; gait and stance progressed on a robust linear scale, and an optimised SARA version, excluding finger chase and nose–finger tests, reduced required trial sample sizes by 20–25% [36]. Published progression rates differ between entities: 0.45 SARA points per year in COQ8A ataxia [32], 0.12 points per year on a composite index in a multicentre cohort of 57 patients with NPC [29], and 0.29 SARA points per year in adult SCA27B [57].

4. Clinical Recognition and Warning Signs

4.1. When to suspect ataxia, by age

In infants, congenital ataxia presents as hypotonia, poor sucking and motor delay; ataxia itself only becomes visible when the child sits unsupported and starts walking, between 1 and 2 years [18]. The French PNDS for non-progressive congenital ataxias gives as reasons for referral head control after 3 months, sitting after 10 months and independent walking after 18 months; language delay with dysarthria and, at school age, learning difficulties complete the picture, the history should record birth parameters and growth and head circumference curves, and improvement over time, which defines the non-progressive form, can only be confirmed through follow-up [23]. Progressive microcephaly, severe feeding difficulties, apnoeas or episodic hyperpnoea, ocular abnormalities and associated malformations point to PCH, Joubert and other malformation syndromes, while macrocephaly (rare) suggests SNX14 or PIGT [19]. In children already walking, progressive gait instability, falls, loss of motor acquisitions and hand clumsiness should be distinguished from instability due to immaturity, developmental coordination disorder and muscle weakness [18, 64]. In adolescents and young adults, progressive ataxia with sensory neuropathy and areflexia first suggests Friedreich ataxia and, when excluded, AVED, abetalipoproteinaemia and ataxias with oculomotor apraxia [8, 18].

The history should include a three-generation pedigree (signs in parents and siblings; some dominant SCAs, particularly SCA7, may begin in childhood), consanguinity, suggesting recessive or metabolic disease, ethnic origin, because several ataxias cluster geographically, pregnancy, delivery and the perinatal period, motor milestones and walking age, distinguishing congenital from later-onset disease, speech and swallowing, and visual loss [18]. Family history guides the inheritance hypothesis: anticipation suggests a dominant expansion; an affected parent or grandparent suggests dominant inheritance; absence of male-to-male transmission suggests X-linked inheritance; affected siblings or consanguinity suggest recessive inheritance, although communities with a high carrier frequency, such as Saguenay–Lac-Saint-Jean for ARSACS, may have affected individuals in two generations; and a grandfather with late ataxia and a grandchild with intellectual disability suggests FXTAS [7]. Without molecular confirmation, balance complaints in relatives do not imply a shared cause, and multifactorial or acquired ataxias are 4 to 5 times more common than hereditary forms [7].

4.2. Diagnostic warning signs

4.3. Emergency signs

Ataxia UK recommends urgent paediatric referral for a child with ataxic symptoms [27]. Acute or subacute ataxia requires immediate exclusion of acquired causes, such as tumour, stroke, trauma, hypoxia, intoxication, acute post-infectious cerebellitis (for example after varicella), encephalitis and autoimmune causes, and metabolic decompensation, with ammonia, blood gases, glucose, lactate and ketones measured in the acute phase [20, 26, 65]. Rapid progression over weeks or months in adults suggests paraneoplastic or prion disease or multiple system atrophy and requires urgent investigation [27]. In Refsum disease, the acute form with arrhythmias or extreme weakness warrants emergency apheresis [10], and in PCH2 prolonged fever requires hydration to prevent rhabdomyolysis [16].

4.4. Newborn screening

None of the early-onset recessive ataxias is included in national newborn screening programmes in the sources consulted. For CTX, dried blood spot methods based on bile alcohol glucuronides and metabolite ratios have been developed and tested in biomarker studies of 20,000 and 32,000 newborns, and the expert Delphi panel considered dried blood spots the optimal screening method, with 71% agreement; at the time of that consensus, no national programme had been implemented [66–70]. The argument for screening is the dependence of treatment response on age at initiation, discussed in section 11 [71].

5. Diagnosis: Criteria, Tests and Algorithms

5.1. General principles

No formal diagnostic criteria exist for early-onset ataxia as a group; diagnosis is aetiological and relies on combining clinical phenotype, biochemistry, electrophysiology, neuroimaging and molecular analysis [1]. Several treatable entities also lack consensus criteria: AVED, abetalipoproteinaemia and adult Refsum disease are diagnosed by a characteristic biomarker together with biallelic variants in the corresponding gene [8–10]. CTX and NPC have expert consensuses with diagnostic recommendations [29, 70]. Detailed phenotyping remains essential in the sequencing era: in a French validation study of a clinical algorithm for recessive ataxias, two assessors blinded to molecular results predicted the correct gene in 61% and 78% of 23 patients diagnosed through capture of 57 genes, with a kappa coefficient of 0.85 (0.55–0.98) [72].

5.2. ERN-RND flowchart for early-onset ataxia

Between September 2019 and July 2020, the ERN-RND Ataxias and Hereditary Spastic Paraplegias disease group developed a diagnostic flowchart for early-onset ataxia, with patient representatives participating, and agreed it in July 2020; the network recommends its use and cautions that the document is not continuously updated [25]. The steps are summarised in the following table.

Source: [25].

5.3. CACG-EPNS paediatric algorithm

In 2019, the European Paediatric Neurology Society’s Childhood Ataxia and Cerebellar Group published an algorithm with seven consecutive steps: identifying early-onset ataxia, applying INAS, considering family history, neuroimaging, laboratory investigations, genetic testing by array-CGH and next-generation sequencing; the algorithm also aims to standardise data entry into early-onset ataxia patient databases [2]. The sources consulted include only the abstract and reference list of this article; the distinctive phenotype and targeted investigation tables to which the ERN-RND flowchart refers have been replaced in this article by equivalent tables from clinical genetics and metabolic disease textbooks and the Ataxia UK guideline [18, 20, 27].

5.4. Adolescents and young adults: adult flowchart, EFNS/ENS and Ataxia UK

At transition to adulthood, the ERN-RND adult ataxia flowchart, published in 2019 and revised in September 2024, which refers early-onset ataxia to the dedicated flowchart, and the 2014 European Federation of Neurological Societies and European Neurological Society (EFNS/ENS) consensus apply [26, 73]. The adult flowchart directs exclusion of acquired causes when onset is acute or subacute, there is a specific medical history or MRI findings (autoimmune disorders, toxins, trauma, hypoxia, tumour, stroke, infections, vitamin deficiencies, paraneoplastic syndromes), assessment for peripheral neuropathy or sensory neuronopathy, and identification of MRI warning signs; first-line tests depend on inheritance pattern [26].

Source: [26].

The EFNS/ENS consensus recommends first-line screening for SCA1, 2, 3, 6, 7 and 17 (level B) with a dominant history, and DRPLA in Asian patients; with a recessive history, a stepwise assessment using biochemical markers and targeted testing for Friedreich ataxia, A-T, AVED, POLG, ARSACS, AOA1 and AOA2; and with a negative family history, full assessment of recessive causes when onset is before 45 years, FMR1 premutation testing after 45 years and consideration of multiple system atrophy in sporadic cases beginning after 30 years [73]. The British Ataxia UK guideline, in its third edition published in 2019 and endorsed by ERN-RND as ready for dissemination, organises investigations by level of care [27, 74].

Source: [27].

5.5. Biochemical investigation in children

The Oxford Desk Reference recommends investigating children with ataxia in conjunction with a paediatric neurologist, except where there is a strong family history of a hereditary form with a similar age at onset [18]. Basic biochemistry includes electrolytes, liver tests, blood gases, fasting glucose, pyruvate and lactate, ammonia, plasma amino acids and urinary amino and organic acids; second-line tests include white-cell enzymes (metachromatic leukodystrophy, Krabbe, gangliosidoses, galactosialidosis), very-long-chain fatty acids and phytanic acid, alpha-fetoprotein, vitamin E, cholesterol and albumin, lipids with apolipoprotein B and transferrin isoelectric focusing, together with investigation for lysosomal inclusions when a metabolic disorder is suspected [18]. Structural cerebellar abnormalities warrant additional full blood count (aplastic anaemia in Hoyeraal-Hreidarsson syndrome), CK, renal function and ultrasound (Joubert), 7-dehydrocholesterol (Smith-Lemli-Opitz), alkaline phosphatase (high or low in GPI anchor defects) and mitochondrial studies when cerebellar atrophy accompanies neuromuscular or other organ signs [19].

For non-progressive congenital ataxia, the French PNDS proposes a minimum set of tests prior to genetic analysis, intended mainly to exclude differential diagnoses: CK, for the hypothesis of myopathy, with electroneuromyography as clinically indicated; thyroid function, looking for the hypothyroidism of benign hereditary chorea; alpha-fetoprotein, to exclude early-onset A-T still without telangiectasias or oculomotor apraxia; and anti-transglutaminase antibodies, for the rare ataxic forms of coeliac disease, even without growth failure [23]. Depending on the context, in a paediatric neurology or neurology department, a metabolic work-up is added: transferrin electrophoresis (PMM2-CDG, which has its own PNDS); the METAglut1 blood test, which measures by immunofluorescence the expression of the glucose transporter on erythrocytes, or, if this is unavailable or normal, lumbar puncture with CSF/blood glucose ratio (below 0.4 in GLUT1 deficiency); lactate and pyruvate in blood and cerebrospinal fluid, with redox status (mitochondrial disorders); beta-galactosidase (GM1 gangliosidosis); very long-chain fatty acids (rare peroxisomal disorders); and screening for neuronal ceroid lipofuscinoses [23]. A normal METAglut1 does not exclude GLUT1 deficiency, and most genetic panels for congenital or very early-onset ataxia include the genes for these disorders [23].

5.6. Imaging, neurophysiology and ophthalmological assessment

MRI distinguishes hypoplasia, suggesting a congenital cause, from progressive atrophy, and identifies white matter abnormalities in metabolic disorders [18]. Some patterns are highly suggestive: the molar tooth sign in Joubert syndrome [19]; a “dragonfly” cerebellum, with hemispheres more affected than the vermis and ventral pontine atrophy, in TSEN54 PCH [16]; paramedian hypointense pontine stripes and hyperintense rims around the thalami in ARSACS [14]; bilateral dentate nucleus and cerebral and cerebellar white matter hyperintensity, with elevated lactate and low N-acetylaspartate on spectroscopy, in CTX [11, 47]; marked cerebellar atrophy with relative brainstem sparing in AOA2 [13]; inferior vermian hypoplasia with gyral simplification in CAMRQ1 [15]; and universal cerebellar atrophy, sometimes with dentate nucleus or pontine T2 hyperintensity, in COQ8A ataxia [32]. In AVED, cerebellar atrophy is present in approximately half of patients and there are no specific signs [8]. MRI should be reviewed with a neuroradiologist, establishing whether the abnormality is restricted to the cerebellum or involves the cortex, and whether hypoplasia affects only the vermis or also the hemispheres [19].

For non-progressive congenital ataxia, the French PNDS considers MRI the reference examination and the first to request, in a radiology department with experience: transfontanellar ultrasound explores the posterior fossa poorly and does not contribute, and computed tomography detects severe abnormalities but may miss mild or moderate ones [23]. The protocol includes a high-resolution 3D T1 sequence and thin sagittal T2 slices centred on the midline, in addition to the usual sequences (coronal T2 or 3D T2, sagittal T1, axial or 3D FLAIR and a magnetic susceptibility sequence, T2* or SWI), with spectroscopy or diffusion tensor imaging when a differential diagnosis is suspected [23]. On qualitative analysis, a normal-sized vermis extends, in the sagittal plane, from the inferior colliculi to the obex in the newborn and from the intercollicular sulcus to the obex in infants and children, the anterior lobe occupies approximately one third of the vermian surface, and the midbrain, pons and medulla have similar heights; quantitative analysis uses the reference values for vermian height and anteroposterior diameter, brainstem and midbrain published by Jandeaux and colleagues in 2019, based on 718 children [23]. The atrophy of NPCA cannot be distinguished from that of degenerative ataxias; a normal initial MRI should be repeated approximately 2 years later, even if the child continues to progress, and imaging has neither clinico-radiological correlation nor prognostic value [23]. Associated abnormalities of the brainstem, of the posterior fossa or supratentorial fluid spaces point to other diagnoses, and the pontine striations of ARSACS may be absent early on [23]. A complete ophthalmological and orthoptic examination (visual acuity, ocular motility, slit lamp, fundoscopy, with visual evoked potentials and electroretinography as clinically indicated) is systematic; electroneuromyography, auditory evoked potentials and electroencephalography are requested when neuropathy, hearing loss or epilepsy is suspected [23].

Nerve conduction studies and electromyography identify the neuropathy of Friedreich ataxia, ARSACS, Roussy-Lévy syndrome, NARP, ataxias with oculomotor apraxia (axonal neuropathy in 90–100% in AOA2) and AVED, although electrophysiology in the latter is non-specific and severe neuropathy does not exclude the diagnosis; somatosensory evoked potentials may show increased central conduction time [8, 13, 18]. Ophthalmological assessment includes electroretinography (retinopathies) and OCT (peripapillary fibre layer > 119 µm in ARSACS) [14, 18]. Electrocardiography and echocardiography assess for Friedreich cardiomyopathy, and electroencephalography is useful when Angelman syndrome, IOSCA, Unverricht-Lundborg disease or ceroid lipofuscinosis is suspected [18].

5.7. Genetic strategy and diagnostic yield

The molecular strategy combines two types of tests. Expansion tests with sizing include Friedreich ataxia in children and SCAs and DRPLA in older children and adolescents; these can be performed as expansion panels or mixed panels combining expansions and sequencing [7, 18]. For conventional variants, the choice is between a multigene panel, maximising detection and limiting variants of uncertain significance (VUS) and non-explanatory findings, and exome or genome sequencing [7]. In children with associated developmental delay or intellectual disability, the 2021 American College of Medical Genetics and Genomics (ACMG) guideline strongly recommends considering exome or genome sequencing as a first- or second-line test, with higher yield than conventional testing and potential cost-effectiveness when requested early; this guideline is not ataxia-specific, and applying it to early-onset ataxia is an extrapolation [75]. The European flowchart permits starting sequencing in parallel with investigations after MRI [25]. First-line strategy depends on local epidemiology: in a Chinese cohort of 2096 probands of all ages, yield was 84.9%, dominated by SCA3 (59.6%), with no FXN expansion carriers, leading the authors to propose a regional stepwise algorithm [76].

The UCLA, 100,000 Genomes, Australian and Campinas cohorts are predominantly adult, and their yields cannot be transferred to paediatric ataxia; paediatric data are scarcer and heterogeneous in design [77, 78, 82–85].

The French NPCA PNDS proposes its own decision tree: after clinical and radiological diagnosis, information for the family and consent, a congenital ataxia gene panel is requested, with a yield of approximately 30% in 2026; if the result is conclusive, it is completed where necessary by parental or transcript studies and information is given on prenatal diagnosis; if not, exome or genome sequencing is requested [23]. The choice of a panel as first test is justified by its better detection of mosaic variants in blood, which exome and genome sequencing detect inconsistently, and the strategy can be discussed at the national multidisciplinary consultation meeting of the reference centre [23]. Chromosomal microarray is reserved for second line, only when the sequencing technique used does not detect copy number variants, which is rare [23].

5.8. Test limitations

Technical limitations govern interpretation of any negative result. Repeat expansions and mitochondrial DNA disorders may escape next-generation sequencing, and exome sequencing does not readily detect expansions [7, 33]. Short-read genome sequencing has limitations in identifying complex structural variants and sizing large expansions; it missed all FGF14 GAA expansions in a British cohort and produced 15% false-positive SCA27B results in an Australian cohort, so these expansions require confirmation by long-range PCR, repeat-primed PCR or long-read sequencing [84, 85]. Copy-number variants require dedicated methods: in abetalipoproteinaemia, sequencing detects variants in 59 of 60 probands and deletion/duplication analysis should follow when 0 or 1 variants are found [9]; in ARSACS, sequencing detects approximately 95%, while deletions, often involving the entire gene and neighbouring genes, account for approximately 5% and are detectable by microarray [14]; the Hutterite VLDLR founder variant is a whole-gene deletion [15]. Sequencing analytical sensitivity is high for other treatable genes: more than 97% of TTPA variants (with one whole-gene deletion described), more than 95% for PHYH and PEX7 and 99% for CYP27A1 (1% by deletion or duplication) [8, 10, 11]. These figures measure the proportion of variants detectable by the technique among molecularly diagnosed patients, not the clinical sensitivity of the test in a patient with unexplained ataxia. In NPC, deep intronic and structural variants may escape sequencing and require multiplex ligation-dependent probe amplification (MLPA), complementary DNA analysis or genome sequencing [29].

Apparent homozygosity for a variant may represent an undetected deletion of one or more exons on the other allele (false homozygosity) or uniparental isodisomy, and parental testing resolves phase [7]. Parental somatic or germline mosaicism may remain undetected in blood [7]. In congenital ataxias, a targeted panel detects a mosaic variant in the proband’s blood cells better than exome or genome sequencing [23]. Even after comprehensive assessment, up to 40% of adults with late-onset cerebellar ataxia without a family history have no identified genetic cause, and a negative result does not exclude hereditary disease [7]. Conversely, sporadic ataxia does not exclude a genetic cause: according to the Oxford Desk Reference, citing Abele and colleagues (2002), a patient with sporadic ataxia has a 13% probability of SCA1, 2, 3, 6, 7 or 17 or Friedreich ataxia [33]. Intermediate alleles in expansion disorders complicate counselling and predictive testing [33].

5.9. Interpretation, reporting and reanalysis

The molecular report should state the gene, reference transcript with version, HGVS nomenclature at complementary DNA and protein levels, zygosity, classification under ACMG and Association for Molecular Pathology (ACMG/AMP) criteria and, in expansion disorders, the repeat count for each allele. VUS are frequent in ataxias: at the Sheffield paediatric centre, 46.7% of variants identified by sequencing were classified as VUS [78]. In AVED, two biallelic VUS neither confirm nor exclude diagnosis [8], and plasma vitamin E may serve as a functional assay, as in the case described in section 10, where a novel missense variant in trans with a pathogenic variant coexisted with very low vitamin E [50]. Parental segregation confirms phase for biallelic variants, and detailed phenotyping (for example OCT in ARSACS) helps interpretation [27, 29]. Periodic exome data reanalysis increased pathogenic diagnoses by 7% in the UCLA cohort and is included in the European flowchart [25, 83]. The Ataxia Global Initiative has published standards for sharing and analysing next-generation sequencing data in ataxias [86], and the British guideline emphasises that sequencing requires specialist bioinformatics and that VUS incur investigation costs [27].

5.10. Confirmation in treatable entities

In CTX, the Delphi panel considered genetic analysis the most important diagnostic element, followed by serum cholestanol, which experts stated is always elevated at diagnosis [70]. In NPC, the 2025 guideline recommends biomarkers (oxysterols, PPCS with lysosphingomyelin, bile acids, also in dried blood spots) as first-line screening and confirmation by NPC1 and NPC2 sequencing (strength 1, evidence A); approximately 700 NPC1 variants have been described, around 420 pathogenic [29]. In primary CoQ10 deficiency, diagnosis uses a panel (steroid-resistant nephrotic syndrome, nuclear mitochondrial disease, retinopathy, ataxia, neuropathy) including the ten genes, or exome or genome sequencing, and muscle biochemistry is reserved for cases without a molecular diagnosis or with VUS; the fibroblast enzyme assay for Refsum disease is no longer clinically available [10, 12]. In abetalipoproteinaemia, diagnosis rests on absent or extremely low LDL, triglycerides and apoB with biallelic MTTP variants [9].

6. Differential Diagnosis

6.1. Acquired causes and mimics

The first step in all algorithms is to exclude ataxia mimics and acquired ataxias [25, 26]. Acquired causes include immune-mediated ataxias (gluten ataxia, anti-GAD antibody ataxia, steroid-responsive encephalopathy associated with autoimmune thyroiditis, Miller-Fisher syndrome, multiple sclerosis, paraneoplastic ataxia), toxic ataxias (alcohol, antiseizure medicines, cytotoxic agents, insecticides, heavy metals, drugs) and infectious or post-infectious ataxias (acute cerebellitis, post-infectious ataxia, Whipple disease, meningoencephalitis, progressive multifocal leukoencephalopathy), as well as structural, vascular, traumatic and hypoxic lesions [26, 65]. Antigliadin antibodies are positive in 10–15% of patients with ataxia, with uncertainty as to whether they are cause or consequence [33]; Ataxia UK recommends testing gluten sensitivity in idiopathic ataxia, with transglutaminase 6 antibodies being more sensitive [27]. In young children, developmental coordination disorder and central hypotonia are the main mimics [64], while Angelman syndrome and ataxic cerebral palsy are differential diagnoses of congenital ataxia [18]. Acquired vitamin E deficiency (unbalanced diets such as Zen macrobiotic diets, cholestasis, short bowel syndrome, cystic fibrosis, Crohn disease) should be distinguished from AVED; it is corrected by supplementation without requiring the high doses used in AVED [8].

Faced with a congenital ataxia, the French PNDS organises the differential diagnosis around hypotonia and MRI [23]. Neuromuscular disorders (Steinert myotonic dystrophy, sensory neuropathies, spinal muscular atrophy) combine weakness, sometimes difficult to demonstrate in infants, with areflexia in neuropathies; reflexes that are difficult to elicit are, however, frequent in cerebellar hypotonia, and elevated CK points to myopathy, with electroneuromyography as clinically indicated [23]. Prader-Willi syndrome and other genetic syndromes with early hypotonia are distinguished by the clinical picture, course, MRI and genetics, and sequelae of anoxia-ischaemia require documented prenatal and perinatal criteria and imaging, bearing in mind that an NPCA may itself favour fetal distress [23]. Benign hereditary chorea, almost always due to NKX2-1, also begins with global hypotonia without weakness, but gait is unstable without a widened base and there are continuous, random choreic movements of the trunk and extremities, worsened by action; it is associated with hypothyroidism, sometimes compensated (elevated TSH with normal T3 and T4), and lung disease (neonatal surfactant deficiency, which may progress to interstitial lung disease, recurrent respiratory infections, asthma) [23]. Early vestibular ataxia is generally accompanied by hearing loss, notably in congenital cytomegalovirus infection, and the proprioceptive ataxia of neuropathies worsens with eye closure [23]. Acquired cerebellar lesions (infectious or ischaemic-haemorrhagic sequelae) usually have a richer clinical picture, with spastic paraparesis or dystonia, and MRI with magnetic susceptibility shows tissue loss, asymmetry or haemorrhagic stigmata; an abnormality confined to one cerebellar hemisphere suggests first a clastic lesion, which COL4A1 or COL4A2 variants may favour, and more rarely PHACE syndrome, with a facial haemangioma appearing in the first week of life [23]. Finally, very early-onset progressive ataxias begin after a free interval, without initial hypotonia: ARSACS between 12 and 18 months, with a cerebellar-spastic syndrome and neuropathy; A-T and AOA2, with elevated alpha-fetoprotein; AOA1, with usually normal alpha-fetoprotein and hypoalbuminaemia and hypercholesterolaemia proportional to disease duration; Friedreich ataxia, rarely confused because it begins after 2 years and has a normal MRI; and insidiously progressive metabolic disorders, particularly neuronal ceroid lipofuscinosis type 6, NPC and GM1 gangliosidosis [23].

6.2. Differential diagnosis between hereditary entities

Sources: [7–11, 13, 14, 19, 20, 23, 29, 31, 32, 45, 46].

6.3. Differentials with the greatest clinical consequences

Distinguishing Friedreich ataxia from AVED has the greatest consequence, because their presentations are clinically indistinguishable and only AVED has effective inexpensive treatment; vitamin E should therefore be measured in every Friedreich-like or unexplained progressive ataxia and whenever FXN testing is negative [8, 18, 31]. In Morocco, 13 of 29 patients with a Friedreich-like phenotype had AVED [8]. In abetalipoproteinaemia, distinction from homozygous hypobetalipoproteinaemia requires the parents’ lipid profile, normal in the former and with LDL below half of normal in the latter [9].

Oculomotor apraxia with early ataxia brings together A-T, AOA1, AOA2 and AOA4 and A-T-like disorder; alpha-fetoprotein, albumin, cholesterol and age at onset guide assessment, and AOA2 diagnosis requires exclusion of A-T, AOA1 and AOA4 [13]. AOA1 differs from A-T by the absence of immunodeficiency, chromosomal instability and radiosensitivity [18]. In spastic ataxia, ARSACS, SPG7, SYNE1 ataxia and CTX are the main differentials, and the ataxia-paraplegia continuum supports panels including both groups [14, 44]. In congenital ataxia, high-resolution MRI is decisive before counselling because it distinguishes malformations with their own recurrence risks from non-malformation forms [18, 19]. Secondary CoQ10 deficiency in AOA1 and respiratory chain disorders can only be distinguished from primary deficiency through molecular testing [12].

7. Treatments and Risk Management Strategies

7.1. Principles

Treatment depends on aetiological diagnosis. A small number of ataxias have specific treatment, and efficacy depends on early initiation: in treatable ataxias, established neurological damage reverses little, whereas presymptomatic treatment may prevent manifestations [8, 12, 71]. Treatment should therefore begin as soon as diagnosis is probable, and siblings should be tested [8, 70]. Other ataxias receive supportive treatment from a multidisciplinary team experienced in disability, with a paediatric neurologist and childhood disability team for children [18, 33]. The British guideline summarises this position by stating that a nihilistic approach is no longer justified [74].

7.2. Specific treatment of treatable ataxias

AVED. Lifelong high-dose oral vitamin E may normalise plasma concentrations; presymptomatic siblings do not develop manifestations [8, 89]. In symptomatic patients treated early, ataxia and intellectual deterioration may reverse; in older patients, progression may stop, but proprioceptive deficits and instability persist [8, 89]. Because α-TTP binds 2R-alpha-tocopherols stereoselectively and some variants lose that selectivity, GeneReviews considers natural RRR-alpha-tocopherol appropriate despite its cost, although synthetic all-rac-alpha-tocopheryl acetate is also used [8]. Even temporary interruption lowers vitamin E within 2 to 3 days and plasma antioxidant capacity for a prolonged period, making adherence critical [8]. Neither formulation has an approved AVED indication in the regulatory sources consulted, and the dose derives from case series [8].

Abetalipoproteinaemia. Calorie intake should be adequate; normal growth is possible but may not reach potential [9]. Dietary fat should supply 10–20% of calories, avoiding long-chain fatty acids, with essential fatty acids supplied by up to one teaspoon daily of oils rich in polyunsaturated fatty acids (soybean, olive oil); fat intake above 20% of calories is generally not tolerated [9, 31]. Vitamin A is titrated using serum β-carotene, targeting low-normal levels to avoid hepatotoxicity (described in one patient); plasma vitamin E always remains low despite supplementation, and early vitamin E may delay or prevent neurological dysfunction, ophthalmoplegia and ptosis, whereas vitamin A may arrest visual progression and prevents corneal ulcers [9]. Hepatic steatosis without fibrosis requires fat restriction and avoidance of anti-inflammatory medicines; in very rare fibrosis or cirrhosis, liver transplantation may be considered [9, 45]. No drug targets the molecular defect; Saudubray’s textbook includes “MTTP inhibitors” among therapeutic options, an error because MTTP is already deficient in abetalipoproteinaemia and MTTP inhibitors (lomitapide) are intended for homozygous familial hypercholesterolaemia [9, 31, 45]. Rare hypothyroidism is treated with levothyroxine, and mild anaemia generally requires no treatment [9].

Refsum disease. The phytanic acid-restricted diet excludes dairy products, ruminant meat and fats (cattle, sheep, goats) and certain fish, such as cod; walnuts contain phytanic acid and phytol, whereas almonds, coconut and peanuts test negative; green leafy vegetables need not be restricted because humans do not release phytol from chlorophyll [10, 46]. The diet should be high in calories to avoid mobilisation of lipid stores, and, if intake is limited, parenteral nutrition or fluids maintaining blood glucose and preventing ketosis are administered [10]. Phytanic acid levels never fully normalise; reduction resolves, approximately in this order, ichthyosis, sensory neuropathy and ataxia, and may stabilise cardiac abnormalities, while retinitis pigmentosa, hearing loss and anosmia are more refractory [10, 46]. During acute crisis with arrhythmias or extreme weakness, plasmapheresis or lipid apheresis reduces phytanic acid by 50–70%, typically to 100–300 µmol/L [10]. Cataracts require adapted surgical techniques (iris hooks, iris-fixated anterior chamber intraocular lenses, capsular tension rings), ichthyosis is treated with emollients, and severe cardiomyopathy may require heart transplantation, successfully performed in two patients [10]. Restrictive diets require vitamin and trace-element monitoring [31].

CTX. Chenodeoxycholic acid (CDCA) has been used since 1984; it reduces cholestanol and cholesterol synthesis and reverses neurological disability in a significant number of patients (dementia, orientation, intelligence quotient, strength), without evident MRI improvement; ursodeoxycholic acid is ineffective [47]. In a series treated with 750 mg/day in adults and 5–15 mg/kg/day in children, neurological improvement occurred in 85% on the Rankin scale and 85% on the Expanded Disability Status Scale (EDSS), diarrhoea, epilepsy and neuropathy resolved in all, pyramidal dysfunction improved or stabilised in 10/15 and cerebellar dysfunction in 12/14, psychiatric symptoms in 6/7, and parkinsonism (2 patients) did not respond [47]. In a cohort of 14 patients, cholestanol normalised within months, patients aged up to 25 years treated less than 15 years after neurological onset improved, and those treated more than 25 years after onset continued to deteriorate [47]. CDCA normalises plasma and cerebrospinal fluid cholestanol, normalises nerve conduction velocity and improves osteoporosis [11]. The Delphi panel ranked CDCA alone first choice and CDCA with a statin second, without consensus on cholic acid alone (33% agreement and 50% disagreement) [70]. Cholic acid, tested in 12 adults (10 stable or improved, 2 deteriorated), is an alternative when CDCA causes adverse effects and is probably preferable in infantile cholestasis [11, 47]. Statins (simvastatin, pravastatin; one ineffective lovastatin study) and LDL apheresis have been used, cautiously because of muscle injury and rhabdomyolysis risk [11, 47]. Cataracts are usually operated on in at least one eye by 50 years; epilepsy, spasticity and parkinsonism are treated symptomatically, and parkinsonism responds poorly to levodopa [11].

Primary CoQ10 deficiency. There is no cure or published guideline; high-dose oral CoQ10 should start as early as possible because it may limit progression and reverse some manifestations, but established severe neurological or renal damage does not reverse [12]. Response varies by gene: in COQ8A ataxia, most respond poorly, with three favourable responses described (objective stabilisation of ataxia; marked sustained improvement in dystonia and myoclonus after 6 months; improved tremor and handwriting); in COQ2, response is poor in severe forms of the first months and good in later forms; in COQ4, non-responders and partial responses occur; COQ6, COQ8B and PDSS2 responses are mainly renal [12]. Ubiquinol dose and efficacy remain undefined, and idebenone and other short-chain quinone analogues have been associated with clinical deterioration; bypass therapies and caspofungin remain preclinical [12]. In the USA, ubidecarenone is not an approved medicine and is marketed as a supplement, without formulation or bioavailability standards, so preparations are not interchangeable [31]. A 2026 case report described improved COQ8A ataxia with intermittent ketogenic fasting using medium-chain triglycerides, without constituting a recommendation [90].

NPC. The 2025 consensus guideline recommends considering miglustat in all patients with confirmed diagnosis (strength 1, evidence B), levacetylleucine in all patients with neurological symptoms (strength 2, evidence B), arimoclomol in patients with neurological symptoms receiving miglustat (strength 2, evidence B) and combination therapy in all (strength 2, evidence C), recognising that access is resource-limited [29]. In the phase I/II miglustat trial (29 patients aged 12 years or older, 200 mg 3 times daily versus standard care for one year), horizontal saccadic velocity improved significantly only in the analysis excluding patients receiving benzodiazepines (p = 0.028); adverse effects were diarrhoea (85%), flatulence (70%) and weight loss (65%) [29]. In an observational study of 789 patients from five national cohorts, miglustat was associated with lower mortality, with median survival approximately 10 years longer from neurological onset and approximately 5 years from diagnosis; in early-infantile disease, in a French series of 10 treated and 16 untreated patients (1990–2013), median survival was 5.56 versus 4.42 years (p = 0.11) and none survived beyond 7.4 years, whereas 75% of 20 children with late-infantile disease stabilised or improved [29]. Experimental therapies include intrathecal and intravenous 2-hydroxypropyl-β-cyclodextrin, efavirenz, nizubaglustat, gene therapy, antisense oligonucleotides and gene editing in animal models [29]. Brazil has its own practical recommendations for NPC diagnosis and follow-up [91].

Episodic ataxias. Ataxia UK recommends avoiding triggers (stress, caffeine, alcohol, excessive exertion), acetazolamide as first line in EA1 and EA2, recognising that not all respond, with hydration and annual renal ultrasound because of stone risk and caution in sulphonamide hypersensitivity, and 4-aminopyridine as second line in EA2 on a named-patient basis [27]. The reference dose comes from Muth and colleagues’ phase III trial: acetazolamide 250 mg 3 times daily (750 mg/day) or prolonged-release fampridine 10 mg 2 times daily (20 mg/day), in adults [30]. The first 4-aminopyridine series used 5 mg 3 times daily in three patients, with attacks recurring on withdrawal [92]. No paediatric doses have been studied in trials: one report described relief in an 8-year-old refractory to acetazolamide with dalfampridine 0.3 mg/kg [93]. Acetazolamide and fampridine (dalfampridine) have no approved indication for episodic ataxia; fampridine is authorised for multiple sclerosis [30].

SCA27B. Although an adult ataxia, SCA27B enters the differential of rare early-onset cases and responds to 4-aminopyridine: 86% of treated patients in a cohort of 50 reported meaningful daily-life benefit, and three alternating-treatment n-of-1 trials showed reduced symptom duration and severity; in the childhood-onset case already cited, gait and falls improved with 4-aminopyridine [57, 58]. Use is off-label.

Non-progressive congenital ataxias. The French PNDS recognises no specific treatment other than the ketogenic diet in GLUT1 deficiency [23]. Epilepsy is treated with antiseizure medications chosen after electroclinical assessment, with no specificity for NPCA; in TRPM3-associated epilepsy, primidone may be beneficial, its appropriateness being assessed by the referring physician [23]. The paroxysmal episodes of CACNA1A variants, such as episodic ataxia type 2 and hemiplegic migraine, may warrant acetazolamide when frequent and disabling, with monitoring of serum potassium; when it is poorly tolerated, topiramate or 4-aminopyridine are proposed, on the basis of a 2025 case series [23]. Coenzyme Q10 has been suggested for COQ8A and GEMIN5 variants, without efficacy or safety demonstrated in large studies, and behavioural and sleep-onset disorders may require non-specific drugs, within psychological and educational support [23]. The PNDS gives no doses; in table 7.3, the paediatric doses of primidone, topiramate and acetazolamide are taken from the US labels for epilepsy (and, for topiramate, also for migraine prophylaxis), by extrapolation to an analogous condition treated with the same drug [94–96].

7.3. Dosage of labelled and off-label medicines

The US miglustat label describes it as a competitive reversible inhibitor of glucosylceramide synthase, the initial enzyme in synthesis of most glycosphingolipids; the levacetylleucine and arimoclomol labels state that the mechanism through which they exert clinical effects in NPC and A-T is unknown [35, 98, 100].

7.4. Comparative regulatory status

Sources: [28, 31, 35, 87, 97–100, 102–104]. The consulted sources do not include information on reimbursement or national access to these medicines in Portugal (INFARMED) or Brazil (ANVISA).

7.5. Clinical trials

The trials differ in population, endpoint and duration, precluding indirect comparisons between medicines. The 12-week levacetylleucine crossover trials measure a short-term symptomatic effect on a functional scale, and the 18-month open-label NPC extension used natural history controls (difference in the 5-domain index with 95% CI −4.17 to −0.46), a design at risk of bias [29]. In the arimoclomol trial, benefit was concentrated in patients receiving miglustat (−0.06 versus +4.2 with arimoclomol alone), and the US label considers data without miglustat insufficient; serious adverse events were less frequent with arimoclomol (14.7% versus 31.3%), but urticaria and angioedema were serious related events [29, 100]. The chenodiol trial measures a biomarker in 13 adult patients rather than a clinical endpoint [87]. In the riluzole trial, the proportion improving (50% versus 11%) represents an absolute difference of approximately 39 percentage points in a small, aetiologically mixed sample [101]. In the EA2 trial, only 18 of 30 patients had genetic confirmation [30]. The NEAT trial of erythrocyte-encapsulated dexamethasone is included because its registry result showed no demonstrated benefit and led to discontinuation of development [105, 106].

7.6. Symptomatic treatment and rehabilitation

Rehabilitation underpins treatment of ataxias without specific therapy: rehabilitation medicine, occupational therapy and physiotherapy (balance, gait training, strengthening), adaptive equipment (sticks, walkers, ramps, weighted cutlery, dressing aids), weight control and home adaptations [7]. Inpatient rehabilitation with occupational therapy and physiotherapy may improve ataxia and function in degenerative ataxias; the EFNS/ENS consensus assigned class II evidence to physiotherapy and class III to occupational therapy [7, 73]. In ARSACS, recommendations include physiotherapy and exercise games, early baclofen and physiotherapy to prevent tendon shortening and contractures, support for dysarthria and hearing aids, and low-dose amitriptyline or oxybutynin for urinary urgency; a 4-week pilot study of home speech therapy in 7 patients improved intelligibility [14, 31]. In AOA2 and other progressive forms, educational support may include voice recognition and adapted keyboards [13]. In CAMRQ and congenital forms, early intervention with occupational therapy, physiotherapy, speech and language therapy, child mental health care and special education is standard [15].

For non-progressive congenital ataxias, the French PNDS details rehabilitation, without efficacy demonstrated in formal studies but with benefit supported by clinical experience [23]. Up to 6 years of age, early multidisciplinary follow-up ideally takes place in an early medico-social intervention service, coordinated by a paediatric neurologist or a physical and rehabilitation medicine physician, who identifies the needs for orthoses, moulded seating and adapted pushchairs [23]. Motor physiotherapy starts early, particularly when there is axial hypotonia and motor delay, works on motor developmental levels, axial tone and postural control, prevents orthopaedic deformities and teaches the child to anticipate risk situations, to fall and to get up from the floor; in severe forms, verticalisation with technical aids is always indicated, and physiotherapy may continue through adolescence [23]. Psychomotor therapy works on global coordination, balance, fine motor skills and visuomotor integration, and occupational therapy looks early for activity limitations caused by intention tremor, dysmetria and eye–hand incoordination, with adaptations for daily living and writing (adapted tools, non-slip supports, digital aids) [23]. Speech and language therapy starts very early when there are sucking and swallowing difficulties or orofacial hypotonia and then addresses dysarthria and language delay, in which expression is more affected than comprehension, with augmentative and alternative communication when needed and systematic screening for written language disorders [23]. The adapted physical activity teacher supports balance, tone, movement control and endurance, with continuation in adapted or mainstream sports clubs, and educational and psychological support promotes autonomy, sleep, feeding, socialisation and parenting [23]. Oculomotor problems disrupt learning even with normal acuity and call for school adaptations (tilting desk, additional lighting, enlarged documents) and good positioning of the child; refractive errors are corrected with glasses, strabismic amblyopia is treated by occlusion, the more effectively the earlier it is started, persistent strabismus may be operated on, and retinal or optic nerve lesions, which have no curative treatment, should be sought early, particularly in children with cognitive impairment, in whom they go unnoticed [23]. The orthopaedic complications of hypotonia (pes planovalgus, kyphoscoliosis, hip dysplasia) are screened for early, for orthoses, rehabilitation and, if necessary, surgery [23].

Riluzole is the only medicine that GeneReviews considers to have demonstrated improvement in ataxic symptoms in mixed-aetiology ataxias, based on an 8-week pilot trial and the 12-month randomised trial [7, 101, 107]. Cerebellar transcranial magnetic stimulation may improve motor signs after 21 daily sessions, according to a 2018 American Academy of Neurology (AAN) systematic review subsequently retired by the society; its conclusions (for example 4-aminopyridine 15 mg/day in EA2, riluzole in Friedreich ataxia and SCAs, no lithium benefit in SCA3 and possible worsening with deferiprone in Friedreich ataxia) no longer constitute current guidance and serve only as a historical record [7, 108]. The EFNS/ENS consensus considered idebenone ineffective in Friedreich ataxia (level A) and varenicline useful in SCA3 (level B) [73]. A 2026 review of treatment for adult-onset neurodegenerative ataxias adds transcranial direct current stimulation and repetitive magnetic stimulation as promising and places antisense oligonucleotides, viral vectors and gene editing at preclinical or early clinical stages [109]. In children, a 2026 review concludes that emerging therapies have not yet substantially altered survival in paediatric ataxia and that the focus remains on complications, multidisciplinary teams and palliative care [110]. In an Ataxia UK survey of 426 people, balance, coordination, speech and fatigue were the symptoms with the greatest impact [27].

7.7. Developments in the last twelve months (October 2025 to October 2026)

The search for developments did not identify specific therapeutic trials for AVED, ARSACS, ataxias with oculomotor apraxia or COQ8A ataxia within the period considered, which does not demonstrate their non-existence.

7.8. Agents and situations to avoid

In Refsum disease, avoid phytanic acid-rich foods, fasting and sudden weight loss, ibuprofen, which interferes with phytanic acid metabolism, and amiodarone because of hyperthyroidism and catabolism risk [10]. In AVED, smoking lowers antioxidant capacity and vitamin E, and occupations requiring rapid reactions or good balance should be avoided [8]. In abetalipoproteinaemia, avoid long-chain fatty acids and, with steatosis, anti-inflammatory medicines [9]. In CTX, statins require caution because of muscle risk, and chenodeoxycholic acid interacts with ciclosporin, sirolimus, phenobarbital, oral contraceptives, bile acid sequestrants and aluminium-containing antacids [11, 28]. In episodic ataxias, avoid stress, caffeine, alcohol and excessive exertion [27]. In CoQ10 deficiency, idebenone and short-chain quinone analogues have been associated with deterioration [12]. In patients with cognitive impairment requiring a bladder antimuscarinic, trospium or darifenacin are preferred [27]. In ataxias due to EXOSC3 or EXOSC5, very rare cases of thrombotic microangiopathy after viral infections have been described; an increased risk has not been demonstrated, but the French PNDS recommends flagging it so that the complication is recognised and treated if it arises, and EXOSC5 variants are associated with cardiac conduction disorders that may require a pacemaker [23].

7.9. Pregnancy and lactation

In AVED, vitamin E should remain high-normal during pregnancy because α-TTP is expressed in the placenta and deficient mice have low fertility and embryonic resorption [8]. In abetalipoproteinaemia, excess vitamin A is teratogenic: the dose should be reduced by 50% in pregnant women and those planning pregnancy, with β-carotene monitoring, without stopping supplementation because deficiency causes maternal morbidity [9]. In Refsum disease, pregnancy is a catabolic state requiring phytanic acid control; rapid visual field reduction in the third trimester was observed in a series without events despite elevated phytanic acid, and unaffected offspring had no problems; breastfeeding is discouraged because phytanic acid enters milk and because not breastfeeding reduces maternal catabolism [10]. In CTX, chenodeoxycholic acid should not be stopped during pregnancy according to GeneReviews; the US label reports decades of cases without increased major malformations or miscarriage, with hepatic, renal and adrenal lesions in monkey fetuses exposed to 1–2 times the human dose; the Delphi panel did not reach consensus on CDCA for fetal protection (67% agreement), and the European summary of product characteristics (SmPC) requires effective contraception and additional pregnancy monitoring [11, 28, 70, 87]. Levacetylleucine and arimoclomol cause embryo-fetal toxicity in animals; exclude pregnancy before levacetylleucine and use effective contraception until 7 days after the last dose [35, 100].

7.10. Perioperative care and catabolic situations

In Refsum disease, the postoperative period requires phytanic acid-free parenteral nutrition (for example soybean oil emulsions with egg-yolk phospholipids) and high calorie intake, and heart transplantation has been performed with perioperative and postoperative high-calorie infusions [10]. In PCH2, prolonged fever requires hydration to prevent rhabdomyolysis [16]. In urea cycle disorders, organic acidurias and PDH deficiency, fever, fasting, infections and protein meals may trigger decompensation with acute ataxia [18, 20].

7.11. Transition to adult care

Transition requires moving from the paediatric to the ERN-RND adult flowchart, which includes RFC1, SCA27B and, above 50 years, FXTAS in first-line testing [25, 26]. The British guideline recommends referral to a specialist ataxia centre and coordination with clinical genetics [27]. In treatable ataxias, transition must not interrupt therapy: in AVED, interruption lowers vitamin E within days, and poor adherence is a documented problem [8].

For non-progressive congenital ataxias, the French PNDS recommends preparing transition from 16 years, with the referring paediatric neurologist, reviewing all manifestations and their surveillance and addressing contraception, sexuality, pregnancy and social and occupational integration [23]. It proposes four modalities of liaison with adult neurology (reference centre for neurogenetic disorders or, if intellectual disability predominates, for intellectual disability): dual follow-up between 16 and 18 years, in separate consultations or a joint consultation; a consultation meeting dedicated to transition with the reference centre; a therapeutic education programme with a transition session (for example “CervEduc”, aimed at patients with a congenital cerebellar anomaly and their carers); and, at a minimum, a detailed letter with the clinical history, given to the patient and sent to the physical and rehabilitation medicine physician, the general practitioner and the specialists involved [23]. Adult physicians value seeing the young person without the parents, and transition is an opportunity for a new multidisciplinary assessment (neuropsychological, speech and language therapy, occupational therapy, social rights), for information on adult benefits and on possible legal protection when there is intellectual disability and, without a genetic diagnosis, for a new genetics clinic appointment, particularly when pregnancy is desired [23]. In adulthood, few patients are followed in specialised neurology, and coordination often falls to the physical and rehabilitation medicine physician, the psychiatrist or the general practitioner [23].

7.12. Guidelines and divergences

Sources: [2, 11, 23, 25–29, 70, 73, 75, 108]. Spanish practical recommendations exist for clinical assessment of hereditary ataxias and spastic paraplegias [118]. The search found no European Academy of Neurology (EAN) ataxia guideline; the European reference remains the EFNS/ENS consensus, ERN-RND flowcharts and Ataxia UK guideline together.

The most relevant dosage divergence for early-onset ataxia concerns paediatric CDCA. GeneReviews gives 10–20 mg/kg/day, whereas the European SmPC, the only regulatory source with a paediatric indication, gives an initial 5 mg/kg/day in three doses and a maximum 15 mg/kg/day, titrated against cholestanol and bile alcohols; Saudubray’s textbook gives 5–15 mg/kg/day [11, 28, 47]. This article adopts the European SmPC as its dosage reference. In adults, the European dose of 750 mg/day, adjustable to 1000 mg/day, and the fixed US dose of 250 mg 3 times daily share the same starting value [28, 87].

The French NPCA PNDS diverges from the other sources on five points. First, in genetic strategy: it favours a targeted panel as first test, because of better detection of mosaicism, whereas the ACMG guideline recommends exome or genome as first or second line in children with developmental delay or intellectual disability and the ERN-RND flowchart allows exome or genome in parallel immediately after MRI [23, 25, 75]. Second, on microarray, which the PNDS reserves for second line when sequencing does not detect copy number variants and which the ERN-RND includes in steps 5 and 6 [23, 25]. Third, on imaging: the Oxford Desk Reference associates atrophy with increasing neurological sequelae and considers early ultrasound unreliable, whereas the PNDS describes atrophies that only appear on the second MRI without clinical deterioration and, without a known familial variant, proposes additional ultrasound scans and systematic fetal MRI, acknowledging that normal findings do not exclude recurrence [19, 23]. Fourth, on surveillance: GeneReviews proposes annual neurological assessment in hereditary ataxias in general, and the PNDS requests paediatric neurology and ophthalmology consultations twice a year in infancy [7, 23]. Fifth, on drugs for CACNA1A episodes: the PNDS gives no doses, the reference dose of acetazolamide comes from the phase III trial in adults (250 mg three times daily) and, in children, only the epilepsy label (8–30 mg/kg/day) provides a range, applied by extrapolation [23, 30, 96].

8. Reproductive Options and Family Counselling

8.1. Recurrence risks by inheritance pattern

In autosomal recessive forms, the parents of an affected individual are presumed heterozygous and both should be tested; each sibling has a 25% probability of being affected, 50% of being an asymptomatic carrier and 25% of being a non-carrier, and the children of an affected individual are obligate carriers [7]. If only one parent has the variant, consider a de novo variant in the proband or a postzygotic variant in a mosaic parent; apparent homozygosity may reflect an undetected deletion of one or more exons on the other allele or uniparental isodisomy [7, 12]. Carriers do not develop recessive ataxia, with the exception of increased cancer risk in ATM heterozygotes [7]. The aunts and uncles of an affected individual have a 50% carrier probability, and carrier detection requires prior identification of familial variants [7].

In autosomal dominant forms, most patients have an affected parent, but some have de novo variants; an apparently negative family history may result from unrecognised disease, early parental death or late onset [7]. If a parent carries the variant, each sibling and child has a 50% risk; if the variant is not detected in parents, sibling risk is slightly above population risk because of germline mosaicism, described in SCA29; and if parents have not been tested but are clinically healthy, siblings remain at increased risk because of age-dependent penetrance and mosaicism [7]. In non-progressive congenital ataxias, more than half of the variants in solved cases arise de novo, which is reassuring but does not eliminate the risk of recurrence, because of parental mosaicism, more frequent with KIF1A; recurrence within a sibship suggests recessive inheritance, in particular due to GEMIN5 [23]. In X-linked forms, risk depends on sex and maternal carrier status (CASK; OTC deficiency) [18, 19]. Mitochondrial DNA disorders are maternally transmitted, and prenatal diagnosis is rarely straightforward [18].

8.2. Empirical risks when the cause is unknown

When investigation of a child does not identify the aetiology, precise risks are unknown. Recessive inheritance is assumed with an affected sibling or consanguinity, and progressive ataxia suggests a metabolic or genetic aetiology with high recurrence risk [18]. In “ataxic cerebral palsy”, a genetic subgroup exists and recurrence risk may reach 25%; without family history or consanguinity, an empirical estimate of 10–15% is used [18]. In apparently isolated cerebellar hypoplasia or vermian agenesis, risk may reach 25%, without empirical data; dominantly inherited vermian aplasia is described, so parents should be examined for ataxia or nystagmus; tubulinopathies may be dominant; and in non-Mendelian, non-chromosomal Dandy-Walker malformation, recurrence is 1–5% [19]. In adults with sporadic progressive ataxia after syndromic and acquired causes are excluded, offspring risk is unknown and 5% is commonly used without firm data; non-paternity should be considered [33]. These estimates are counselling conventions rather than measured risks.

8.3. Cascade testing, carriers and founder populations

In AVED, all siblings should be tested because presymptomatic individuals benefit from treatment; predictive sibling testing is appropriate, moderately low vitamin E in heterozygotes cannot identify carriers, and partners of the same ethnicity may be tested [8]. In abetalipoproteinaemia, older and younger siblings undergo a lipid profile with apoB, with or without molecular testing [9]. In Refsum disease, siblings undergo molecular testing or phytanic acid measurement before symptoms, and in CTX molecular testing or cholestanol measurement [10, 11]. In CoQ10 deficiency diagnosed only biochemically, siblings may be assessed through fibroblast CoQ10, and DNA banking is recommended without a molecular diagnosis [12].

In founder populations, carrier screening has its own rationale. In Saguenay–Lac-Saint-Jean, ARSACS carrier frequency is 1:21 according to GeneReviews (1:22 in another source), and the birth incidence of 1:1932 has declined with voluntary carrier screening; 12 SACS variants have been described in Quebec, so testing should not be restricted to the two ancestral variants [14, 31]. In Hutterites, with VLDLR carriers at approximately 1:15, population reproductive screening is justified and targeted founder-deletion testing should come first [15]. In CTX, carriers are more frequent in Israel, Morocco and Druze populations, increasing offspring risk for a proband with a partner of these origins; carrier frequency in Ashkenazi Jews is approximately 0.002 [11, 31]. In Marinesco-Sjögren syndrome, carrier frequency is approximately 1:96 in Finland versus 1:700 worldwide [31].

8.4. Predictive testing and minors

In dominant SCA families, predictive testing requires prior confirmation of the variant in an affected individual, follows dedicated protocols and does not predict onset age, severity or progression [33]. Ataxia UK recommends informed consent and genetic counselling before any test, including familial and reproductive implications, offering counselling to asymptomatic at-risk relatives and generally avoiding testing asymptomatic at-risk minors, with case-by-case decisions [27]. Exceptions are treatable disorders in which presymptomatic treatment changes prognosis, such as AVED and CTX, where testing younger siblings is appropriate [8, 70]. Carrier detection in children follows the usual restrictions [18].

8.5. Prenatal and preimplantation diagnosis

When familial variants are known, prenatal diagnosis by chorionic villus sampling and preimplantation genetic testing are possible [8, 9, 18, 33]. Early ultrasound is unreliable for prenatal diagnosis of cerebellar hypoplasia; in a subsequent pregnancy, detailed ultrasound may be offered with discussion of its limitations, and fetal ultrasound or MRI is useful only with structural cerebellar abnormalities [18, 19]. Prenatal diagnosis of mitochondrial disorders is rarely straightforward [18].

The French NPCA PNDS distinguishes two situations in a couple with an affected child [23]. When the causal variants are known, molecular prenatal diagnosis is performed on fetal DNA by chorionic villus sampling at around 12 weeks or, later, by amniocentesis; preimplantation genetic diagnosis is possible in particular cases, with prior agreement of the relevant centre, and non-invasive prenatal diagnosis on maternal blood may be offered in certain situations; in the presence of a mosaic variant, prenatal diagnosis is accepted even with a very low risk; all these procedures require prior preparation and the agreement of a multidisciplinary prenatal diagnosis centre [23]. When the genetic cause has not been identified, diagnosis is only morphological and difficult, because the cerebellum grows mainly at the end of pregnancy and atrophy may only appear years after birth: depending on the proband’s imaging, two ultrasound scans at around 18 and 26 weeks are added to the usual scans at 22 and 32 weeks, all performed by a reference sonographer of a prenatal diagnosis centre, looking for vermian abnormalities, enlargement of the posterior fossa fluid spaces and extracerebral signs; fetal MRI is systematic in this indication, from 25–26 weeks and more often at around 32, repeatable, with midsagittal and parasagittal slices of the vermis, superior cerebellar peduncles and brainstem; normal imaging does not exclude recurrence, and, if an abnormality is detected and the pregnancy continues, postnatal MRI may be performed depending on the course [23]. The couple may also consider sperm donation or adoption [23].

In the general population, the French recommendations for screening obstetric ultrasound, updated in 2023 by the Conférence Nationale de l'Échographie Obstétricale et Fœtale (CNEOF), now include measurement of the transverse cerebellar diameter in the second and third trimesters, usually with the charts of Chavez and colleagues, and qualitative analysis of the posterior fossa and cerebellum (two hemispheres, dimensions and contours, vermis and primary fissure); a diameter below that expected or an index case in the family leads to a reference ultrasound scan and discussion of fetal MRI at a prenatal diagnosis centre [23]. Even so, a large proportion of hypoplasias escape screening, because cerebellar growth only slows at the end of pregnancy or in the first years of life, and most NPCA cannot be diagnosed before birth [23]. On fetal MRI, the norms of Parazzini and colleagues are used before 24 weeks and those of Tilea and colleagues after 26; magnetic susceptibility sequences reveal haemorrhagic stigmata that ultrasound does not show, and foliation is well analysed, although the nine vermian lobules cannot be distinguished [23]. Hypoplasia of a single cerebellar hemisphere generally corresponds to a clastic, haemorrhagic or ischaemic lesion, which should be termed a focal lesion rather than hypoplasia; more rarely it is part of PHACE syndrome, suggested by hemispheres shaped like a tilted telephone receiver on coronal section; hypoplasia of the vermis, of the brainstem or bilateral and symmetrical hypoplasia of the hemispheres suggests first a malformative or genetic cause, or else an infectious, alcohol-related or other exogenous cause; and there is significant discordance between prenatal and postnatal diagnosis of vermian hypoplasias and dysplasias [23].

8.6. Counselling content

Counselling should explain inheritance and risks to siblings, offspring and other relatives; treatment possibilities in treatable forms, with urgent sibling testing; test limitations and the meaning of a negative result or VUS; reproductive options; and family support resources [7, 27]. In congenital ataxia, high-resolution MRI should precede definitive counselling because hypoplasia or Dandy-Walker malformation changes the recurrence estimate [18]. After a diagnosis of NPCA, the French PNDS recommends referring the parents to a genetic counselling consultation (clinical geneticist or genetic counsellor), refined once a molecular diagnosis is available, with confirmation of the mode of transmission, the recurrence risk according to whether the variant is de novo or inherited, and options in a future pregnancy; the diagnosis is disclosed by the referring physician, who explains the improvement expected over time and states that the risk of early death is exceptional [23].

9. Special Considerations: Population Diversity and Epidemiology

9.1. What each number measures

There is no prevalence estimate for early-onset ataxia as a whole; available data measure different things (prevalence of all hereditary ataxias, ARCA-group prevalence, prevalence of individual entities, allele frequency estimated from genomic databases or diagnostic proportions in a clinical cohort) and are not interchangeable.

The prevalence of NPCA, as defined by the French PNDS, is poorly known and probably underestimated, because recent studies tend not to separate non-progressive from progressive forms; the Swedish and European figures include acquired or progressive forms, and the French figure is an incidence estimate from a reference centre, not a population study [23].

The Orphanet figure of 3.3/100 000 for Portugal corresponds to recessive ataxias in the national survey, including Friedreich ataxia and patients of all ages; it is not an early-onset ataxia prevalence [5, 21]. In the Portuguese survey, ataxia with oculomotor apraxia was among the most frequent hereditary ataxias after Machado-Joseph disease and Friedreich ataxia, and the Oxford Desk Reference identifies AOA1 as the second cause of recessive ataxia in Portugal after Friedreich ataxia, and the first in Japan [18, 21]. Machado-Joseph disease has a large Portuguese founder population [7].

9.2. Geography, ancestry and founder effect

Several recessive ataxias cluster in founder populations: ARSACS and ARCA1 (SYNE1, Beauce region) in Quebec; AOA2 in the Gaspé Peninsula; CAMRQ1 in Hutterites; the TTPA c.744del variant in North Africa and Italy; TTPA p.His101Gln in Japan; CYP27A1 c.355del in Druze populations; ANO10 c.1150_1151del in Romani communities; and Marinesco-Sjögren syndrome in Finland [8, 11, 13–15, 31]. SYNE1 ataxia has also been described in Japan, Europe, the Middle East and Brazil [31].

The aetiological spectrum varies by population. In an Omani paediatric series, in a population with high consanguinity, recessive inheritance accounted for 65.7%, and A-T (17%) and ARSACS (7.1%) were the most frequent causes [80]. In India, SACS and SETX led diagnoses in an early-onset ataxia cohort [79]. In the UK, Friedreich ataxia and CACNA1A were the main paediatric genetic causes at the Sheffield centre [78]. In China, SCAs dominate and no FXN expansions were found [76].

Brazil. In Porto Alegre between 2002 and 2020, in a series of ataxias without a vertical family history, the most common diagnoses were Friedreich ataxia, A-T, CoQ10 deficiency, NPC, AOA2 and SCA2, and access barriers in the Unified Health System resulted in loss to follow-up and incomplete investigation in more than half of patients [81]. In Campinas and São Paulo, exome sequencing diagnosed 35.5% of adults with idiopathic ataxia, with SACS and SPG7 the most frequent genes [82]. In a Brazilian reference centre, SCA27B accounted for 2.8% of SCA probands, a lower relative frequency than in Europe and Canada [124].

Gaps. The consulted sources provide no prevalence for early-onset ataxia as a whole in any country, or population-level paediatric data for Portugal or Brazil; absence from these sources does not demonstrate absence of published studies. Several rare-entity estimates derive from founder populations and cannot be extrapolated to other populations.

9.3. Special populations

During pregnancy, treatable ataxias require specific adjustments (vitamin A in abetalipoproteinaemia, vitamin E in AVED, phytanic acid control in Refsum disease, continued CDCA in CTX), described in section 7.9 [8–11]. In children, labelled medicine doses differ between jurisdictions (CDCA, levacetylleucine), and SARA requires age-specific reference interpretation [28, 35, 52, 99]. In patients with intellectual disability, adherence to restrictive diets and lifelong treatment requires family and social support, included in the GeneReviews assessment tables [7].

9.4. Reference networks and patient associations

The relevant European network is ERN-RND (https://www.ern-rnd.eu/), whose Ataxias and Hereditary Spastic Paraplegias disease group brings together 32 centres and 10 affiliated partners in 20 Member States and published the diagnostic flowcharts cited [25]. Patient associations include Euro-ATAXIA, the European federation of hereditary ataxia associations (https://euro-ataxia.eu/); Ataxia UK (https://www.ataxia.org.uk/); the National Ataxia Foundation in the USA (https://www.ataxia.org/); Friedreich's Ataxia Research Alliance (https://www.curefa.org/); the A-T Children's Project (https://atcp.org/); FEDAES (Spanish Ataxia Federation) and the Italian association AISA; APAHE, Associação Portuguesa de Ataxias Hereditárias (https://www.apahe.pt/); and ABAHE, Associação Brasileira de Ataxias Hereditárias e Adquiridas (https://abahe.org.br/) [7, 18, 27]. For families with cerebellar malformations, the Oxford Desk Reference lists Contact a Family (http://www.cafamily.org.uk) [19].

In France, congenital ataxias are followed by the reference network for cerebellar malformations and congenital disorders (Centre de Référence Malformations et Maladies Congénitales du Cervelet, 2M2C; https://anomaliescervelet.aphp.fr/), part of the national rare neurodevelopmental disease network DéfiScience (https://defiscience.fr/), with three reference centres (Paris-Hôpital Trousseau, coordinator, Lille and Lyon) and five competence centres (Besançon, Montpellier, Paris-Pitié-Salpêtrière, Reims and Tours), in liaison with the networks for neurogenetic disorders, developmental anomalies and malformation syndromes, and intellectual disability [23]. The centre organises a national multidisciplinary consultation meeting and one-day multidisciplinary consultations; the general practitioner refers the child to the centre, provides local follow-up, vaccinations and visual and hearing screening, and contributes to the care protocol that underpins exemption from patient cost-sharing as an off-list long-term condition [23]. Families receive information on benefits and recognition of disability through the departmental house for disabled persons (MDPH, now the departmental autonomy house, MDA): education allowance for a disabled child (AEEH), disabled adult allowance (AAH), disability compensation benefit (PCH) and recognition of disabled worker status (RQTH); schooling may benefit from teaching adaptations, extra time, a support worker for pupils with disabilities (AESH), school inclusion units (ULIS) or medico-educational institutes (IME) and motor education institutes (IEM), and visual impairment from specialised early family support services (SAFEP) [23]. The French associations are CSC (Connaître les Syndromes Cérébelleux), with a paediatric branch (https://www.csc.asso.fr/), and KIF1A France (https://kif1a.fr/); the therapeutic education programme “CervEduc” is being rolled out in the reference centres [23].

10. Illustrative Clinical Case

Baso and colleagues described in 2026 two Italian siblings, born to healthy non-consanguineous parents, assessed in the neuromuscular disease clinic at Milan’s Policlinico for progressive gait instability [50]. Symptoms began subtly in adulthood as clumsiness in sports compared with peers; neither sibling had onset before 15 years. The 28-year-old brother reported progressive arm clumsiness, lower-limb weakness, falls and altered speech from age 20. The 36-year-old sister had recurrent neck pain from age 21, followed by gait instability with falls during hazardous tasks, particularly in darkness, and a nasal voice, without visual complaints [50].

Examination in both showed scanning dysarthria, mild head titubation, dysdiadochokinesia, diffuse areflexia, marked distal proprioceptive loss and Romberg sign; bilateral Babinski signs were present only in the brother. SARA scores were 13/40 in the brother and 19/40 in the sister. Brain MRI and cerebellar volumes were normal, sensory and motor potentials were preserved, electromyography showed motor unit potentials with slightly increased amplitude and duration, and fundoscopy and visual acuity were normal [50].

Given the similar presentations, the authors requested a clinical exome targeted to central and peripheral ataxic syndromes. Both had biallelic TTPA variants: NM_000370.3:c.513_514insTT, p.(Thr172LeufsTer5), maternally inherited, pathogenic, frequent in northern European AVED families and usually associated with onset by approximately 10 years; and NM_000370.3:c.158T>C, p.(Leu53Pro), paternally inherited, previously undescribed, absent from gnomAD, with a PolyPhen-2 score of 1.00, located in the CRAL_TRIO_N domain near residues forming the wall of the alpha-tocopherol binding pocket (Figure 1, pedigree showing variants and vitamin E levels: https://pmc.ncbi.nlm.nih.gov/articles/PMC13186537/figure/F1/; Figure 2, α-TTP structure with p.Leu53Pro substitution: https://pmc.ncbi.nlm.nih.gov/articles/PMC13186537/figure/F2/) [50]. Sanger sequencing confirmed variants and segregation, each parent carrying one. With a balanced diet, normal lipids and no ileal disease, plasma vitamin E was 1.36 mg/L in the brother and 1.26 mg/L in the sister (reference 5.6–14.8 mg/L), versus 11.16 and 8.35 mg/L in the parents [50]. Without functional studies, the missense variant falls between likely pathogenic and uncertain significance under ACMG/AMP criteria; very low vitamin E in both siblings serves here as specific phenotypic evidence.

Patients began oral vitamin E supplementation at 800 mg daily, with the article not specifying the chemical form. At 3 months, vitamin E was 6.35 mg/L in the brother and 14.4 mg/L in the sister; at 6 months, with physiotherapy, SARA improved by 1 point in each (12/40 and 18/40), subjective improvement occurred and conduction studies and electromyography were unchanged; progression was arrested without clear examination improvement [50]. The authors suggest the missense variant may attenuate the truncating variant’s effect, explaining late-adolescent onset and preserved walking around 30 years, and recommend including TTPA in hereditary ataxia investigations even in mild phenotypes, to begin supplementation promptly [50].

The case illustrates four points from this article. First, AVED reproduces the Friedreich ataxia phenotype, with areflexia, proprioceptive loss and Babinski signs, and head titubation is a clue [8, 20]. Second, inexpensive accessible vitamin E measurement would have guided diagnosis before sequencing, provided preanalytical requirements were observed [8]. Third, the dose used in the adult siblings (800 mg/day) is at the lower end of the 800–1500 mg/day range described in GeneReviews for adults, which recommends maintaining high-normal vitamin E and favours RRR-alpha-tocopherol; the brother’s vitamin E at 3 months was near the laboratory’s lower reference limit, which, according to GeneReviews, would justify six-monthly monitoring and possible adjustment [8]. Fourth, limited response in adults with several years of disease agrees with observations that late treatment stabilises but rarely reverses proprioceptive deficits [8, 89]. Other published reports describe AVED initially diagnosed as Friedreich ataxia and AVED with a predominantly cervicobrachial dystonic presentation [55, 125].

11. Prognosis and Long-Term Follow-up

11.1. Prognosis by group, with and without treatment

Prognosis differs radically between entities and should be communicated after aetiological diagnosis rather than from the category “early-onset ataxia”. In treatable ataxias, age at treatment initiation is the main determinant. In abetalipoproteinaemia, survival rarely extended beyond the third decade without high-dose fat-soluble vitamins, whereas lifelong oral treatment permits survival into the seventh or eighth decade with minimal symptoms [9]. In CTX, a retrospective Dutch cohort of 56 patients with median follow-up of 8 years showed that, among 43 adherent patients, neurological symptoms disappeared in all treated before 24 years and no new symptoms developed, whereas 61% treated after 24 years deteriorated with treatment-resistant parkinsonism; improvement or stabilisation on the modified Rankin scale was 100% versus 58% and on EDSS 100% versus 50% [71]. Experts considered cerebellar vacuolisation on MRI a poor prognostic marker (88% agreement) [70]. In AVED, presymptomatic treatment prevents manifestations and late treatment stabilises without reversing proprioceptive deficits [8]. In Refsum disease, diet stabilises neuropathy, ataxia and ichthyosis, but retinopathy continues to progress slowly; arrhythmias and cardiomyopathy are frequent causes of death [10]. In primary CoQ10 deficiency, severe multisystem forms respond poorly and cause death in the neonatal period or first year, whereas later forms respond better [12]. In NPC, miglustat was associated with approximately 10 years of additional median survival from neurological onset in an observational study, but not in early-infantile disease [29].

In degenerative ataxias without specific treatment, progression varies: in AOA1, walking is lost 7 to 10 years after onset; in AOA2, wheelchair use usually begins around 30 years, with reported disease duration of 2 to 53 years and one patient examined at 79 years; in Friedreich ataxia, walking is lost approximately 15 years after onset and death occurs at a mean of 37.5 years [13, 18]. In congenital forms, the course may be stable, as in CAMRQ1, where life expectancy does not appear reduced, or rapidly fatal, as in PCH2, with death usually before 10 years, and PCH4, with neonatal death [15, 16]. In Marinesco-Sjögren syndrome, motor function deteriorates for several years then stabilises [31]. In paediatric ataxia overall, emerging therapies have not yet substantially changed survival [110]. In non-progressive congenital ataxias, the cerebellar syndrome behaves as a fixed neurological disorder, without regression: ataxia and hypotonia improve with age, falls disappear and the widened base usually disappears, sometimes leaving only instability in tandem gait, single-leg stance and turning; most children achieve independent walking, sometimes late, the risk of early death is exceptional and imaging has no prognostic value [23].

11.2. Surveillance

GeneReviews notes the absence of published surveillance guidelines for hereditary ataxias generally and proposes annual neurological assessment, more frequently with acute worsening, including SARA, rehabilitation medicine, occupational therapy and physiotherapy; dysarthria, dysphagia, mood, psychosis and cognition are assessed according to evolution and symptoms, while family needs, social work, palliative care and carer respite are reviewed at each consultation [7]. Initial assessment should include baseline SARA, electrophysiology, MRI, speech, nutritional status and aspiration risk assessment, respiratory testing and sleep studies if symptomatic, cognitive assessment with the CCAS scale, musculoskeletal assessment (scoliosis, pes cavus) and genetic counselling [7, 60]. In non-progressive congenital ataxias, the French PNDS defines follow-up common to all patients, to which gene-specific surveillance is added: the paediatric neurologist coordinates overall follow-up, checks that extraneurological follow-up is carried out and teaches families the signs of an epileptic seizure when the gene carries that risk; children with marked axial hypotonia need equipment for sitting or standing and neuro-orthopaedic surveillance of the spine and joints, in physical and rehabilitation medicine and orthopaedics; and consultations at a reference or competence centre should take place at least at the key ages of the care pathway [23].

11.3. Lifelong follow-up and care

Follow-up should continue at a centre experienced in ataxias, coordinating paediatric neurology or neurology, clinical genetics, metabolic medicine, rehabilitation, ophthalmology, cardiology and nutrition according to the entity [7, 27]. In patients without a molecular diagnosis, genomic reanalysis and clinical phenotype reassessment form part of follow-up, and DNA should be banked [12, 18, 25]. Ataxia UK includes palliative care in progressive ataxia management, and the 2026 paediatric review emphasises its role in children [74, 110]. SARA, the International Cooperative Ataxia Rating Scale (ICARS) and the Friedreich Ataxia Rating Scale (FARS) support clinical follow-up, and paediatric SARA requires age-specific reference interpretation [27, 52].

12. Summary and Final Considerations

Early-onset cerebellar ataxia brings together dozens of genetic disorders, predominantly autosomal recessive, manifesting before 25 years in congenital, progressive or episodic forms. Harding’s clinical classification has given way to gene-based nomenclature and pathophysiological classifications, but practice still starts with the patient: age and mode of onset, three-generation family history and consanguinity, and above all non-ataxic signs, which guide aetiological grouping. The first task is to exclude acquired causes and mimics, with urgent referral for a child with acute ataxia.

The ERN-RND early-onset ataxia flowchart and European paediatric group algorithm structure investigation into steps: history, INAS and SARA examination, MRI and neurophysiology, targeted biochemistry, expansion testing and next-generation sequencing, which may proceed in parallel, followed by mitochondrial DNA, microarray and reanalysis if negative. Biochemistry remains essential because it identifies treatable causes and helps interpret variants: vitamin E, lipid profile with apoB, phytanic acid, cholestanol, oxysterols, alpha-fetoprotein, albumin, cholesterol, lactate and cerebrospinal fluid glucose. Genetic tests have limitations governing interpretation of a negative result: short-read sequencing misses or mis-sizes expansions and poorly detects structural variants and mitochondrial DNA, while apparent homozygosity may conceal a deletion or isodisomy; the proportion of detectable variants in a gene is not the test’s clinical sensitivity.

Treatable ataxias warrant systematic investigation: AVED responds to high-dose vitamin E, abetalipoproteinaemia to fat-soluble vitamins with a low-fat diet, Refsum disease to phytanic acid restriction, CTX to chenodeoxycholic acid, primary CoQ10 deficiency to supplementation, NPC to miglustat and newer medicines, episodic ataxias to acetazolamide and 4-aminopyridine, and GLUT1 deficiency to a ketogenic diet. In all, benefit depends on early initiation, as illustrated by CTX, where patients treated before 24 years had better outcomes, and siblings should be tested. Doses and indications differ between jurisdictions: chenodeoxycholic acid has a paediatric dose only in the European label, miglustat has no NPC indication in the USA and arimoclomol received a negative European Union opinion in July 2026, while levacetylleucine was authorised for NPC in Europe and extended to A-T in the USA in 2026.

In ataxias without specific treatment, rehabilitation, symptomatic treatment of spasticity, tremor, dysphagia and urinary and cognitive abnormalities, and family and palliative support underpin care, with riluzole the only medicine showing symptomatic benefit in randomised mixed-aetiology trials, used off-label. Genetic counselling rests on molecular diagnosis, empirical risks when unavailable and carrier screening in founder populations, and surveillance, without general guidelines, follows entity-specific recommendations. In non-progressive congenital ataxias, which the 2026 French PNDS treats as developmental disorders, hypotonia precedes ataxia and the course is one of improvement; a targeted panel identifies the cause in approximately 30% of patients, more than half through de novo variants, and follow-up combines early rehabilitation, neuropsychological assessments at school transitions, regular ophthalmology and a follow-up MRI when the first is normal.

13. References

1. Hosseinpour S, Bemanalizadeh M, Mohammadi P, Ashrafi MR, Heidari M. An overview of early-onset cerebellar ataxia: a practical guideline. Acta Neurol Belg. 2024;124(6):1791-1804. doi:10.1007/s13760-024-02595-w. PMID: 38951452.
2. Brandsma R, Verschuuren-Bemelmans CC, Amrom D, Barisic N, Baxter P, Bertini E, et al. A clinical diagnostic algorithm for early onset cerebellar ataxia. Eur J Paediatr Neurol. 2019;23(5):692-706. doi:10.1016/j.ejpn.2019.08.004. PMID: 31481303.
3. Beaudin M, Matilla-Dueñas A, Soong BW, Pedroso JL, Barsottini OG, Mitoma H, et al. The Classification of Autosomal Recessive Cerebellar Ataxias: a Consensus Statement from the Society for Research on the Cerebellum and Ataxias Task Force. Cerebellum. 2019;18(6):1098-1125. doi:10.1007/s12311-019-01052-2. PMID: 31267374. PMC6867988.
4. Orphanet. Early-onset cerebellar ataxia with retained tendon reflexes (ORPHA:1177). Orphadata, 2026-06-23. https://www.orpha.net/en/disease/detail/1177
5. Orphanet. Autosomal recessive cerebellar ataxia (ORPHA:1172). Orphadata epidemiology dataset, 2026-06-23. https://www.orpha.net/en/disease/detail/1172
6. Orphanet. Autosomal recessive congenital cerebellar ataxia (ORPHA:98095). https://www.orpha.net/en/disease/detail/98095
7. Perlman S. Hereditary Ataxia Overview. 1998 Oct 28 [updated 2022 Jun 16; revised 2026 Jul 9]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. https://www.ncbi.nlm.nih.gov/books/NBK1138/
8. Schuelke M. Ataxia with Vitamin E Deficiency. 2005 May 20 [updated 2023 Mar 16]. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. https://www.ncbi.nlm.nih.gov/books/NBK1241/
9. Burnett JR, Hooper AJ, Hegele RA. Abetalipoproteinemia. 2018 Oct 25 [revised 2022 May 19]. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. https://www.ncbi.nlm.nih.gov/books/NBK532447/
10. Waterham HR, Wanders RJA, Leroy BP. Adult Refsum Disease. 2006 Mar 20 [updated 2021 Sep 30]. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. https://www.ncbi.nlm.nih.gov/books/NBK1353/
11. Federico A, Gallus GN. Cerebrotendinous Xanthomatosis. 2003 Jul 16 [updated 2022 Mar 17; revised 2024 Nov 14]. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. https://www.ncbi.nlm.nih.gov/books/NBK1409/
12. Salviati L, Trevisson E, Agosto C, Doimo M, Navas P. Primary Coenzyme Q10 Deficiency Overview. 2017 Jan 26 [updated 2023 Jun 8]. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. https://www.ncbi.nlm.nih.gov/books/NBK410087/
13. Moreira MC, Koenig M. Ataxia with Oculomotor Apraxia Type 2. 2004 Nov 15 [updated 2018 Jul 12]. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. https://www.ncbi.nlm.nih.gov/books/NBK1154/
14. Vermeer S, van de Warrenburg BP, Kamsteeg EJ, Brais B, Synofzik M. ARSACS. 2003 Dec 9 [updated 2020 Jan 2]. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. https://www.ncbi.nlm.nih.gov/books/NBK1255/
15. Boycott KM, MacDonald SK, Parboosingh JS. VLDLR Cerebellar Hypoplasia. 2008 Aug 26 [updated 2020 Feb 27]. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. https://www.ncbi.nlm.nih.gov/books/NBK1874/
16. van Dijk T, Baas F. TSEN54 Pontocerebellar Hypoplasia. 2009 Sep 8 [updated 2020 May 28]. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. https://www.ncbi.nlm.nih.gov/books/NBK9673/
17. Brashear A, Sweadner KJ, Haq I, Napoli E, Ozelius L. ATP1A3-Related Disorder. 2008 Feb 7 [updated 2024 Dec 5]. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. https://www.ncbi.nlm.nih.gov/books/NBK1115/
18. Firth HV, Hurst JA. Ataxic child. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 74–77.
19. Firth HV, Hurst JA. Cerebellar anomalies. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 92–95.
20. García-Cazorla Á, Saudubray JM. Clinical approach to inborn errors of metabolism in paediatrics. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 1, sections 1.4.2 (Table 1.7) and 1.5.2.3 (Table 1.19).
21. Coutinho P, Ruano L, Loureiro JL, Cruz VT, Barros J, Tuna A, et al. Hereditary ataxia and spastic paraplegia in Portugal: a population-based prevalence study. JAMA Neurol. 2013;70(6):746-55. doi:10.1001/jamaneurol.2013.1707. PMID: 23609960.
22. Chiò A, Orsi L, Mortara P, Schiffer D. Early onset cerebellar ataxia with retained tendon reflexes: prevalence and gene frequency in an Italian population. Clin Genet. 1993;43(4):207-11. doi:10.1111/j.1399-0004.1993.tb04450.x. PMID: 8330454.
23. Haute Autorité de Santé; Centre de Référence Malformations et Maladies Congénitales du Cervelet; Filière DéfiScience. Protocole National de Diagnostic et de Soins (PNDS): Ataxies congénitales non progressives (ACNP). Saint-Denis La Plaine: HAS; 2026 Jun 26 (text of May 2026). Available from: https://www.has-sante.fr/jcms/p_4092418
24. Jacobi H, Rakowicz M, Rola R, Fancellu R, Mariotti C, Charles P, et al. Inventory of Non-Ataxia Signs (INAS): validation of a new clinical assessment instrument. Cerebellum. 2013;12(3):418-28. doi:10.1007/s12311-012-0421-3. PMID: 23090211.
25. European Reference Network for Rare Neurological Diseases (ERN-RND), Disease Group for Ataxia and Hereditary Spastic Paraplegias. Diagnostic flowchart for early-onset ataxias. Consented July 2020. Available from: https://www.ern-rnd.eu/wp-content/uploads/2020/11/ERN-RND-Diagnostic-flowchart-for-early-Ataxias_final.pdf
26. European Reference Network for Rare Neurological Diseases (ERN-RND). Diagnostic flowchart for adult ataxias. Published 11 February 2019; last revision 16 September 2024. Available from: https://www.ern-rnd.eu/wp-content/uploads/2024/09/Final-Flowchart-adult-ataxias_en-New.pdf
27. de Silva R, Greenfield J, Cook A, Bonney H, Vallortigara J, Hunt B, et al. Guidelines on the diagnosis and management of the progressive ataxias. Orphanet J Rare Dis. 2019;14(1):51. doi:10.1186/s13023-019-1013-9. PMID: 30786918. PMC6381619.
28. European Medicines Agency. Chenodeoxycholic acid Leadiant: EPAR – product information (SmPC). Authorised 10 Apr 2017; latest EC decision 12 Aug 2026. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/chenodeoxycholic-acid-leadiant
29. Hiwot T, Porter FD, Bremova-Ertl T, Ramaswami U, Hastings C, Héron B, et al. 2025 Consensus Clinical Management Guidelines for Niemann-Pick Disease Type C. J Inherit Metab Dis. 2026;49(3):e70185. doi:10.1002/jimd.70185. PMID: 42036217. PMC13110917.
30. Muth C, Teufel J, Schöls L, Synofzik M, Franke C, Timmann D, et al. Fampridine and Acetazolamide in EA2 and Related Familial EA: A Prospective Randomized Placebo-Controlled Trial. Neurol Clin Pract. 2021;11(4):e438-e446. doi:10.1212/CPJ.0000000000001017. PMID: 34484942. PMC8382428.
31. GeneGuidance (user database). Records for genes TTPA, MTTP, PHYH, PEX7, CYP27A1, COQ8A, COQ2, COQ4, APTX, SETX, SACS, SYNE1, ANO10, VLDLR, TSEN54, PNKP, SIL1 and ITPR1: identifiers, clinical curation, linked conditions and listed references. Accessed 9 Oct 2026.
32. Traschütz A, Schirinzi T, Laugwitz L, Murray NH, Bingman CA, Reich S, et al. Clinico-Genetic, Imaging and Molecular Delineation of COQ8A-Ataxia: A Multicenter Study of 59 Patients. Ann Neurol. 2020;88(2):251-263. doi:10.1002/ana.25751. PMID: 32337771. PMC7877690.
33. Firth HV, Hurst JA. Ataxic adult. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 70–73.
34. Orphanet. X-linked non progressive cerebellar ataxia (ORPHA:314978). https://www.orpha.net/en/disease/detail/314978
35. IntraBio Inc. AQNEURSA (levacetylleucine) for oral suspension: prescribing information. NDA 219132. Approved 24 Sep 2024 (NPC); supplement 2 (efficacy, A-T) approved 18 Sep 2026. DailyMed setid 0f248e55-d1bb-13f9-e063-6294a90a05ce, published 21 Sep 2026. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=0f248e55-d1bb-13f9-e063-6294a90a05ce
36. Traschütz A, Adarmes-Gómez AD, Anheim M, Baets J, Brais B, Gagnon C, et al. Responsiveness of the Scale for the Assessment and Rating of Ataxia and Natural History in 884 Recessive and Early Onset Ataxia Patients. Ann Neurol. 2023;94(3):470-485. doi:10.1002/ana.26712. PMID: 37243847.
37. Harding AE. Classification of the hereditary ataxias and paraplegias. Lancet. 1983;1(8334):1151-5. doi:10.1016/s0140-6736(83)92879-9. PMID: 6133167.
38. Ben Hamida C, Doerflinger N, Belal S, Linder C, Reutenauer L, Dib C, et al. Localization of Friedreich ataxia phenotype with selective vitamin E deficiency to chromosome 8q by homozygosity mapping. Nat Genet. 1993;5(2):195-200. doi:10.1038/ng1093-195. PMID: 8252047.
39. Ouahchi K, Arita M, Kayden H, Hentati F, Ben Hamida M, Sokol R, et al. Ataxia with isolated vitamin E deficiency is caused by mutations in the alpha-tocopherol transfer protein. Nat Genet. 1995;9(2):141-5. doi:10.1038/ng0295-141. PMID: 7719340.
40. Engert JC, Bérubé P, Mercier J, Doré C, Lepage P, Ge B, et al. ARSACS, a spastic ataxia common in northeastern Québec, is caused by mutations in a new gene encoding an 11.5-kb ORF. Nat Genet. 2000;24(2):120-5. doi:10.1038/72769. PMID: 10655055.
41. Moreira MC, Barbot C, Tachi N, Kozuka N, Uchida E, Gibson T, et al. The gene mutated in ataxia-ocular apraxia 1 encodes the new HIT/Zn-finger protein aprataxin. Nat Genet. 2001;29(2):189-93. doi:10.1038/ng1001-189. PMID: 11586300.
42. Rossi M, Anheim M, Durr A, Klein C, Koenig M, Synofzik M, et al. The genetic nomenclature of recessive cerebellar ataxias. Mov Disord. 2018;33(7):1056-1076. doi:10.1002/mds.27415. PMID: 29756227.
43. Beaudin M, Klein CJ, Rouleau GA, Dupré N. Systematic review of autosomal recessive ataxias and proposal for a classification. Cerebellum Ataxias. 2017;4:3. doi:10.1186/s40673-017-0061-y. PMID: 28250961. PMC5324265.
44. Synofzik M, Schüle R. Overcoming the divide between ataxias and spastic paraplegias: Shared phenotypes, genes, and pathways. Mov Disord. 2017;32(3):332-345. doi:10.1002/mds.26944. PMID: 28195350. PMC6287914.
45. Ramaswami U, Humphries SE. Inborn errors of lipoprotein metabolism presenting in childhood. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 36, section 36.1. doi:10.1007/978-3-662-63123-2_36
46. Wanders RJA, Engelen M, Vaz FM. Inborn errors of non-mitochondrial fatty acid metabolism including peroxisomal disorders. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 42, section 42.3 (Adult Refsum disease).
47. Clayton PT. Disorders of bile acid synthesis. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 38, section 38.3 (Cerebrotendinous xanthomatosis).
48. Maia Filho JM, Medeiros LO, Ribeiro EM. Deficiência de coenzima Q10. In: Ribeiro EM, coord. Síndromes Genéticas: Reconhecer para Tratar. Fortaleza: EdUnichristus; 2021. p. 127–129.
49. Firth HV, Hurst JA. DNA repair disorders. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 398–401.
50. Baso G, Magri F, Sciacco M, Corti SP, Comi GP, Ronchi D, et al. Ataxia With Vitamin E Deficiency Syndrome and a Novel TTPA Variant: A Paired Case Report. Neurol Genet. 2026;12(3):e200399. doi:10.1212/NXG.0000000000200399. PMID: 42164107. PMC13186537.
51. Schmitz-Hübsch T, du Montcel ST, Baliko L, Berciano J, Boesch S, Depondt C, et al. Scale for the assessment and rating of ataxia: development of a new clinical scale. Neurology. 2006;66(11):1717-20. doi:10.1212/01.wnl.0000219042.60538.92. PMID: 16769946.
52. Lawerman TF, Brandsma R, Burger H, Burgerhof JGM, Sival DA, the Childhood Ataxia and Cerebellar Group of the European Pediatric Neurology Society. Age-related reference values for the pediatric Scale for Assessment and Rating of Ataxia: a multicentre study. Dev Med Child Neurol. 2017;59(10):1077-1082. doi:10.1111/dmcn.13507. PMID: 28815574.
53. Nyhan WL, Hoffmann GF, Al-Aqeel AI, Barshop BA. PMM2-CDG (Congenital disorders of glycosylation, type Ia). In: Atlas of Inherited Metabolic Diseases. 4th ed. Boca Raton: CRC Press; 2020. p. 771–777.
54. El Euch-Fayache G, Bouhlal Y, Amouri R, Feki M, Hentati F. Molecular, clinical and peripheral neuropathy study of Tunisian patients with ataxia with vitamin E deficiency. Brain. 2014;137(Pt 2):402-10. doi:10.1093/brain/awt339. PMID: 24369383.
55. Holla VV, Gurram S, Kamath SD, Arunachal G, Kamble N, Yadav R, et al. Genetically Proven Ataxia With Vitamin E Deficiency With Predominant Cervicobrachial Dystonic Presentation: A Case Report From India. J Mov Disord. 2024;17(2):220-222. doi:10.14802/jmd.23227. PMID: 38105474. PMC11082611.
56. Mochel F, Sedel F. Inborn errors of metabolism in adults: a diagnostic approach to neurological and psychiatric presentations. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 2, section 2.3.9 (Table 2.11).
57. Wilke C, Pellerin D, Mengel D, Traschütz A, Danzi MC, Dicaire MJ, et al. GAA-FGF14 ataxia (SCA27B): phenotypic profile, natural history progression and 4-aminopyridine treatment response. Brain. 2023;146(10):4144-4157. doi:10.1093/brain/awad157. PMID: 37165652.
58. Mitrotti P, Vegezzi E, Zangaglia R, Palmieri I, Dicaire MJ, Gana S, et al. Bilateral Dentate Nuclei Hyperintensities and Response to 4-Aminopyridine in a Patient With Childhood-Onset GAA-FGF14-Related Ataxia. Neurol Genet. 2024;10(6):e200208. doi:10.1212/NXG.0000000000200208. PMID: 39512794. PMC11543268.
59. Delvallee C, Calmels N, Bogdan T, Tranchant C, Anheim M, Wirth T. No evidence for association between GAA-FGF14 expansion and early onset cerebellar ataxia: a study on 85 undiagnosed patients. J Neurol. 2024;272(1):32. doi:10.1007/s00415-024-12765-8. PMID: 39666057.
60. Hoche F, Guell X, Vangel MG, Sherman JC, Schmahmann JD. The cerebellar cognitive affective/Schmahmann syndrome scale. Brain. 2018;141(1):248-270. doi:10.1093/brain/awx317. PMID: 29206893. PMC5837248.
61. Sival DA, Garofalo M, Brandsma R, Bokkers TA, van den Berg M, de Koning TJ, et al. Early Onset Ataxia with Comorbid Dystonia: Clinical, Anatomical and Biological Pathway Analysis Expose Shared Pathophysiology. Diagnostics (Basel). 2020;10(12). doi:10.3390/diagnostics10120997. PMID: 33255407. PMC7760948.
62. van Noort SAM, van der Veen S, de Koning TJ, de Koning-Tijssen MAJ, Verbeek DS, Sival DA. Early onset ataxia with comorbid myoclonus and epilepsy: A disease spectrum with shared molecular pathways and cortico-thalamo-cerebellar network involvement. Eur J Paediatr Neurol. 2023;45:47-54. doi:10.1016/j.ejpn.2023.05.009. PMID: 37301083.
63. Coenen MA, Sival D, Brandsma R, Eggink H, Timmerman ME, Tijssen MA, et al. Moving across disorders: A cross-sectional study of cognition in early onset ataxia and dystonia. Eur J Paediatr Neurol. 2024;49:100-105. doi:10.1016/j.ejpn.2024.02.016. PMID: 38479210.
64. Lawerman TF, Brandsma R, Maurits NM, Martinez-Manzanera O, Verschuuren-Bemelmans CC, Lunsing RJ, et al. Paediatric motor phenotypes in early-onset ataxia, developmental coordination disorder, and central hypotonia. Dev Med Child Neurol. 2020;62(1):75-82. doi:10.1111/dmcn.14355. PMID: 31529709. PMC6916203.
65. Stezin A, Pal PK. Treatable Ataxias: How to Find the Needle in the Haystack?. J Mov Disord. 2022;15(3):206-226. doi:10.14802/jmd.22069. PMID: 36065614. PMC9536909.
66. Vaz FM, Bootsma AH, Kulik W, Verrips A, Wevers RA, Schielen PC, et al. A newborn screening method for cerebrotendinous xanthomatosis using bile alcohol glucuronides and metabolite ratios. J Lipid Res. 2017;58(5):1002-1007. doi:10.1194/jlr.P075051. PMID: 28314860. PMC5408618.
67. DeBarber AE, Kalfon L, Fedida A, Fleisher Sheffer V, Ben Haroush S, Chasnyk N, et al. Newborn screening for cerebrotendinous xanthomatosis is the solution for early identification and treatment. J Lipid Res. 2018;59(11):2214-2222. doi:10.1194/jlr.M087999. PMID: 30135217. PMC6210902.
68. Hong X, Daiker J, Sadilek M, DeBarber AE, Chiang J, Duan J, et al. Toward newborn screening of cerebrotendinous xanthomatosis: results of a biomarker research study using 32,000 newborn dried blood spots. Genet Med. 2020;22(10):1606-1612. doi:10.1038/s41436-020-0846-x. PMID: 32523054. PMC7529987.
69. Vaz FM, Jamal Y, Barto R, Gelb MH, DeBarber AE, Wevers RA, et al. Newborn screening for Cerebrotendinous Xanthomatosis: A retrospective biomarker study using both flow-injection and UPLC-MS/MS analysis in 20,000 newborns. Clin Chim Acta. 2023;539:170-174. doi:10.1016/j.cca.2022.12.011. PMID: 36529270. PMC10387442.
70. Stelten BML, Dotti MT, Verrips A, Elibol B, Falik-Zaccai TC, Hanman K, et al. Expert opinion on diagnosing, treating and managing patients with cerebrotendinous xanthomatosis (CTX): a modified Delphi study. Orphanet J Rare Dis. 2021;16(1):353. doi:10.1186/s13023-021-01980-5. PMID: 34362411. PMC8349076.
71. Stelten BML, Huidekoper HH, van de Warrenburg BPC, Brilstra EH, Hollak CEM, Haak HR, et al. Long-term treatment effect in cerebrotendinous xanthomatosis depends on age at treatment start. Neurology. 2019;92(2):e83-e95. doi:10.1212/WNL.0000000000006731. PMID: 30530799.
72. Mallaret M, Renaud M, Redin C, Drouot N, Muller J, Severac F, et al. Validation of a clinical practice-based algorithm for the diagnosis of autosomal recessive cerebellar ataxias based on NGS identified cases. J Neurol. 2016;263(7):1314-22. doi:10.1007/s00415-016-8112-5. PMID: 27142713.
73. van de Warrenburg BP, van Gaalen J, Boesch S, Burgunder JM, Dürr A, Giunti P, et al. EFNS/ENS Consensus on the diagnosis and management of chronic ataxias in adulthood. Eur J Neurol. 2014;21(4):552-62. doi:10.1111/ene.12341. PMID: 24418350.
74. 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. doi:10.1136/practneurol-2018-002096. PMID: 31048364. PMC6585307.
75. Manickam K, McClain MR, Demmer LA, Biswas S, Kearney HM, Malinowski J, et al. Exome and genome sequencing for pediatric patients with congenital anomalies or intellectual disability: an evidence-based clinical guideline of the American College of Medical Genetics and Genomics (ACMG). Genet Med. 2021;23(11):2029-2037. doi:10.1038/s41436-021-01242-6. PMID: 34211152.
76. Yu JY, Sun YM, Dong Y, Wu JJ, Zhu CH, Du YC, et al. Large-scale genetic profiling of hereditary ataxias in China: implications for a stepwise molecular diagnostic strategy. Transl Neurodegener. 2026;15(1). doi:10.1186/s40035-026-00584-y. PMID: 42806408. PMC13621623.
77. Sawyer SL, Schwartzentruber J, Beaulieu CL, Dyment D, Smith A, Warman Chardon J, et al. Exome sequencing as a diagnostic tool for pediatric-onset ataxia. Hum Mutat. 2014;35(1):45-9. doi:10.1002/humu.22451. PMID: 24108619. PMC4255313.
78. Garrard K, Beauchamps N, Connolly D, Secker S, Allen KE, Campbell S, et al. Genetic causes of early onset ataxia: experience from the National Sheffield Paediatric Ataxia Centre [preprint]. medRxiv. 2024 Jul 3. doi:10.1101/2024.07.03.24309227.
79. Shakya S, Kumari R, Suroliya V, Tyagi N, Joshi A, Garg A, et al. Whole exome and targeted gene sequencing to detect pathogenic recessive variants in early onset cerebellar ataxia. Clin Genet. 2019;96(6):566-574. doi:10.1111/cge.13625. PMID: 31429931.
80. Al-Habsi A, Al-Thihli K, Al-Ajmi E, Al-Futaisi A, Al-Badi A, Al-Amrani F. Spectrum of Hereditary Ataxia in Omani Children. J Clin Med. 2026;15(18). doi:10.3390/jcm15187216. PMID: 42795988. PMC13607818.
81. Aschoff CAM, Silva TO, Hasan A, Migliorini E, Donis KC, Rare Genomes Project Consortium, et al. The Real Life of Ataxia Patients Without a Vertical Family History: a Twenty-year Experience in South Brazil. Cerebellum. 2026;25(5). doi:10.1007/s12311-026-02067-2. PMID: 42616279. PMC13490217.
82. da Graça FF, Peluzzo TM, Bonadia LC, Martinez ARM, Diniz de Lima F, Pedroso JL, et al. Diagnostic Yield of Whole Exome Sequencing for Adults with Ataxia: a Brazilian Perspective. Cerebellum. 2022;21(1):49-54. doi:10.1007/s12311-021-01268-1. PMID: 33956305.
83. Ngo KJ, Rexach JE, Lee H, Petty LE, Perlman S, Valera JM, et al. A diagnostic ceiling for exome sequencing in cerebellar ataxia and related neurological disorders. Hum Mutat. 2020;41(2):487-501. doi:10.1002/humu.23946. PMID: 31692161. PMC7182470.
84. Yau WY, Sullivan R, O'Connor E, Pellerin D, Parkinson MH, Giunti P, et al. Diagnostic yield and limitations of whole-genome sequencing for hereditary cerebellar ataxia. Brain Commun. 2025;7(3):fcaf188. doi:10.1093/braincomms/fcaf188. PMID: 40488180. PMC12142459.
85. Rafehi H, Fearnley LG, Read J, Snell P, Davies KC, Scott L, et al. A prospective trial comparing programmable targeted long-read sequencing and short-read genome sequencing for genetic diagnosis of cerebellar ataxia. Genome Res. 2025;35(4):769-785. doi:10.1101/gr.279634.124. PMID: 40015980. PMC12047251.
86. Beijer D, Fogel BL, Beltran S, Danzi MC, Németh AH, Züchner S, et al. Standards of NGS Data Sharing and Analysis in Ataxias: Recommendations by the NGS Working Group of the Ataxia Global Initiative. Cerebellum. 2024;23(2):391-400. doi:10.1007/s12311-023-01537-1. PMID: 36869969. PMC10951009.
87. Mirum Pharmaceuticals. CTEXLI (chenodiol) tablets, for oral use: prescribing information. NDA 219488. Initial U.S. approval of CTEXLI: 2025 (Drugs@FDA ORIG-1 approved 21 Feb 2025). Label effective 16 Dec 2025. Available from: https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&ApplNo=219488
88. Klepper J, Akman C, Armeno M, Auvin S, Cervenka M, Cross HJ, et al. Glut1 Deficiency Syndrome (Glut1DS): State of the art in 2020 and recommendations of the international Glut1DS study group. Epilepsia Open. 2020;5(3):354-365. doi:10.1002/epi4.12414. PMID: 32913944. PMC7469861.
89. Schuelke M, Mayatepek E, Inter M, Becker M, Pfeiffer E, Speer A, et al. Treatment of ataxia in isolated vitamin E deficiency caused by alpha-tocopherol transfer protein deficiency. J Pediatr. 1999;134(2):240-4. doi:10.1016/s0022-3476(99)70424-5. PMID: 9931538.
90. Hahn W, Erffmeier K, Schulze M, Zahnert F, Knake S, Tsalouchidou PE. Intermittent ketogenic fasting with medium-chain triglycerides improves ataxia in COQ8A-related coenzyme Q10 deficiency: A case report. Mol Genet Metab Rep. 2026;46:101296. doi:10.1016/j.ymgmr.2026.101296. PMID: 41716779. PMC12914829.
91. Horovitz DDG, Pessoa A, França Junior MC, Giugliani R, Souza CFM, Embiruçu EK, et al. Practical recommendations for diagnosis, management, and follow-up of Niemann-Pick type-C disease patients: a Brazilian perspective. Arq Neuropsiquiatr. 2025;83(3):1-8. doi:10.1055/s-0045-1807714. PMID: 40345672. PMC12064313.
92. Strupp M, Kalla R, Dichgans M, Freilinger T, Glasauer S, Brandt T. Treatment of episodic ataxia type 2 with the potassium channel blocker 4-aminopyridine. Neurology. 2004;62(9):1623-5. doi:10.1212/01.wnl.0000125691.74109.53. PMID: 15136697.
93. Malamud E, Otallah SI. Use of Dalfampridine in a Young Child with Episodic Ataxia Type 2. Child Neurol Open. 2022;9:2329048X221075447. doi:10.1177/2329048X221075447. PMID: 35127965. PMC8811424.
94. Bausch Health US. MYSOLINE (primidone tablets, USP): prescribing information. DailyMed setid af593171-dabb-4ea3-b44c-89ed457b2c46, version 11, label effective 10 Sep 2026. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=af593171-dabb-4ea3-b44c-89ed457b2c46
95. Janssen Pharmaceuticals. TOPAMAX (topiramate) tablets and sprinkle capsules: prescribing information. Initial U.S. approval 1996. DailyMed setid 21628112-0c47-11df-95b3-498d55d89593, version 31, label effective 12 Mar 2026. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=21628112-0c47-11df-95b3-498d55d89593
96. Ajanta Pharma USA. Acetazolamide tablets, USP: prescribing information. DailyMed setid 240ab24b-3124-4805-80c2-36dd1945e3b3, version 7, label effective 20 Jan 2026. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=240ab24b-3124-4805-80c2-36dd1945e3b3
97. European Medicines Agency. Zavesca (miglustat): EPAR – product information. Authorised 20 Nov 2002; latest EC decision 2 Jul 2026. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/zavesca
98. Actelion Pharmaceuticals. ZAVESCA (miglustat) capsules: prescribing information. NDA 021348. Approved 31 Jul 2003; label effective 20 Apr 2023. Available from: https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&ApplNo=021348
99. European Medicines Agency. Aqneursa (L-acetylleucine): EPAR – product information. CHMP opinion 13 Nov 2025; authorised 19 Jan 2026. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/aqneursa
100. Zevra Denmark. MIPLYFFA (arimoclomol) capsules: prescribing information. NDA 214927. Approved 20 Sep 2024; label effective 18 Mar 2026. Available from: https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&ApplNo=214927
101. Romano S, Coarelli G, Marcotulli C, Leonardi L, Piccolo F, Spadaro M, et al. Riluzole in patients with hereditary cerebellar ataxia: a randomised, double-blind, placebo-controlled trial. Lancet Neurol. 2015;14(10):985-91. doi:10.1016/S1474-4422(15)00201-X. PMID: 26321318.
102. European Medicines Agency. Meplyffa (arimoclomol citrate): opinion – negative. CHMP opinion adopted 23 Jul 2026. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/meplyffa
103. Biogen. SKYCLARYS (omaveloxolone) capsules: prescribing information. NDA 216718. Approved 28 Feb 2023; label effective 6 Apr 2026. Available from: https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&ApplNo=216718
104. European Medicines Agency. Skyclarys (omaveloxolone): EPAR. Authorised 9 Feb 2024; indication adults and adolescents ≥16 years. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/skyclarys
105. ClinicalTrials.gov. NCT06193200: Neurological effects of EryDex on subjects with ataxia telangiectasia (NEAT). Quince Therapeutics. Results posted. Available from: https://clinicaltrials.gov/study/NCT06193200
106. ClinicalTrials.gov. NCT06664853: Open-label extension study of EryDex in patients with ataxia telangiectasia (terminated: eDSP development discontinued). Available from: https://clinicaltrials.gov/study/NCT06664853
107. Ristori G, Romano S, Visconti A, Cannoni S, Spadaro M, Frontali M, et al. Riluzole in cerebellar ataxia: a randomized, double-blind, placebo-controlled pilot trial. Neurology. 2010;74(10):839-45. doi:10.1212/WNL.0b013e3181d31e23. PMID: 20211908.
108. Zesiewicz TA, Wilmot G, Kuo SH, Perlman S, Greenstein PE, Ying SH, et al. Comprehensive systematic review summary: Treatment of cerebellar motor dysfunction and ataxia [RETIRED]: Report of the Guideline Development, Dissemination, and Implementation Subcommittee of the American Academy of Neurology. Neurology. 2018;90(10):464-471. doi:10.1212/WNL.0000000000005055. PMID: 29440566. PMC5863491.
109. Ariello LE, Rastall DPW, Rosenthal LS. Treatment of primary adult-onset neurodegenerative cerebellar ataxias. Neurotherapeutics. 2026;23(1):e00805. doi:10.1016/j.neurot.2025.e00805. PMID: 41387161. PMC12976491.
110. Brown H, Barañano K. Pediatric Ataxia and Emerging Treatments. Pediatr Neurol. 2026;183:141-145. doi:10.1016/j.pediatrneurol.2026.07.021. PMID: 42561491.
111. National Ataxia Foundation. Pharma news [Internet]. Minneapolis: NAF; 2025–2026 [cited 2026 Oct 9]. Available from: https://www.ataxia.org/pharmanews/
112. Neurology Advisor. FDA denies approval of troriluzole for spinocerebellar ataxia [Internet]. 2025 [cited 2026 Oct 9]. Available from: https://www.neurologyadvisor.com/news/fda-denies-approval-of-troriluzole-for-spinocerebellar-ataxia/
113. Friedreich's Ataxia Research Alliance. Larimar Therapeutics announces FDA Breakthrough Therapy designation for nomlabofusp in FA and reiterates planned BLA submission in June 2026 [Internet]. 2026 [cited 2026 Oct 9]. Available from: https://www.curefa.org/news/larimar-therapeutics-announces-fda-breakthrough-therapy-designation-for-nomlabofusp-in-fa-and-reiterates-planned-bla-submission-in-june-2026/
114. Öz G, Cocozza S, Rezende TJR, Henry PG, Faber J, Harding IH, et al. MRI end-points for clinical trials in ataxias: recommendations from the Ataxia Global Initiative MRI Biomarkers Working Group. Nat Rev Neurol. 2026;22(7):439-456. doi:10.1038/s41582-026-01218-7. PMID: 42236987.
115. NeurologyLive. FDA denies approval for Friedreich ataxia agent vatiquinone, citing more efficacy needed [Internet]. 2025 [cited 2026 Oct 9]. Available from: https://www.neurologylive.com/view/fda-denies-approval-friedreich-ataxia-agent-vatiquinone-citing-more-efficacy-needed
116. ClinicalTrials.gov. NCT06953583: Phase 3 study of omaveloxolone (BIIB141) in participants with Friedreich's ataxia aged 2 to <16 years (BRAVE). Biogen. Available from: https://clinicaltrials.gov/study/NCT06953583
117. ClinicalTrials.gov. NCT07221292: Effects of N-acetyl-L-leucine on CACNA1A disorders: a phase III, randomized, placebo-controlled, double-blind, crossover study. IntraBio. Available from: https://clinicaltrials.gov/study/NCT07221292
118. Arpa Gutiérrez FJ, Abenza Abildúa MJ, Rouco Axpe I, Adarmes Gómez AD, Serrano Munuera C. Practical recommendations for the clinical evaluation of patients with hereditary ataxia and hereditary spastic paraplegia. Neurologia (Engl Ed). 2024;39(6):515-522. doi:10.1016/j.nrleng.2022.02.002. PMID: 36396094.
119. Ruano L, Melo C, Silva MC, Coutinho P. The global epidemiology of hereditary ataxia and spastic paraplegia: a systematic review of prevalence studies. Neuroepidemiology. 2014;42(3):174-83. doi:10.1159/000358801. PMID: 24603320.
120. Anheim M, Fleury M, Monga B, Laugel V, Chaigne D, Rodier G, et al. Epidemiological, clinical, paraclinical and molecular study of a cohort of 102 patients affected with autosomal recessive progressive cerebellar ataxia from Alsace, Eastern France: implications for clinical management. Neurogenetics. 2010;11(1):1-12. doi:10.1007/s10048-009-0196-y. PMID: 19440741.
121. Erichsen AK, Koht J, Stray-Pedersen A, Abdelnoor M, Tallaksen CM. Prevalence of hereditary ataxia and spastic paraplegia in southeast Norway: a population-based study. Brain. 2009;132(Pt 6):1577-88. doi:10.1093/brain/awp056. PMID: 19339254.
122. Orphanet. Ataxia with vitamin E deficiency (ORPHA:96). https://www.orpha.net/en/disease/detail/96
123. Pramparo T, Steiner RD, Rodems S, Jenkinson C. Allelic prevalence and geographic distribution of cerebrotendinous xanthomatosis. Orphanet J Rare Dis. 2023;18(1):13. doi:10.1186/s13023-022-02578-1. PMID: 36650582. PMC9843874.
124. de Jesus Araujo Dias A, Silveira C, Vinagre AM, Bonadia LC, Santos NBS, Rezende TJR, et al. SCA27B in Brazil: frequency, phenotype and genotype-phenotype correlations. J Neurol. 2026;273(6). doi:10.1007/s00415-026-13880-4. PMID: 42168446. PMC13194261.
125. Bonello M, Ray P. A Case of Ataxia with Isolated Vitamin E Deficiency Initially Diagnosed as Friedreich's Ataxia. Case Rep Neurol Med. 2016;2016:8342653. doi:10.1155/2016/8342653. PMID: 26989534. PMC4771890.
126. Corben LA, Collins V, Milne S, Farmer J, Musheno A, Lynch D, et al. Clinical management guidelines for Friedreich ataxia: best practice in rare diseases. Orphanet J Rare Dis. 2022;17(1):415. doi:10.1186/s13023-022-02568-3. PMID: 36371255. PMC9652828.