Spinocerebellar Ataxias due to Conventional Variants (including SCA42, CACNA1G)

Autosomal dominant cerebellar ataxias caused by sequence variants (missense, nonsense, frameshift, splicing) or by gene deletions and duplications rather than repeat expansions: SCA42 and the CACNA1G allelic spectrum, SCA5, SCA11, SCA13, SCA14, SCA15/16, SCA19/22, SCA20, SCA21, SCA23, SCA26, SCA27A, SCA28, SCA29, SCA34, SCA35, SCA38, SCA40 to SCA50 and allelic recessive forms.

Literature and regulatory information updated: 28 October 2026.

0. Initial Reference Table

The first table summarises the class and the entity that prompted this article (SCA42); the second lists each type in the class with its identifiers and the features that guide suspicion. The spinocerebellar ataxias (SCA) due to CAG triplet expansions (SCA1, SCA2, SCA3, SCA6, SCA7, SCA17 and dentatorubral-pallidoluysian atrophy, DRPLA) and the SCA due to non-coding repeat expansions (SCA8, SCA10, SCA12, SCA27B, SCA31, SCA36, SCA37 and SCA4) appear here only in the differential diagnosis [1, 2].

AD: autosomal dominant; AR: autosomal recessive; ID: intellectual disability; MSA-C: multiple system atrophy of cerebellar type; PRCA: PUM1-related cerebellar ataxia; PADDAS: PUM1-associated developmental delay, ataxia and seizures syndrome. The OMIM numbers, cytoband and inheritance of SCA11, SCA13, SCA14, SCA15/16, SCA19/22, SCA20, SCA27A, SCA28, SPAX5, SCA29, SCA34, SCA38, SCA42, SCA42ND, SCA48, SCAR16 and SCA50 match the respective OMIM entries [61–75]. The PUM1 gene has OMIM number 607204; the number of the SCA47 phenotypic entry was not found in the sources consulted, and the link between PUM1 and adult ataxia has been questioned (3.19) [55]. The Orphanet numbers of SCA48 (631103) and SCAR16 (412057) follow the Orphanet database, which corrects a swapped mapping in one of the secondary databases [19].

1. Introduction

1.1. Scope and definition of the class

Primary hereditary ataxias are, in the definition adopted by GeneReviews, disorders for which the adult or carer seeks diagnosis and follow-up in a neurology service with a multidisciplinary team; childhood-onset multisystem disorders, epileptic encephalopathies, cerebellar malformations defined by magnetic resonance imaging (MRI), leukodystrophies, inborn errors of metabolism and primary mitochondrial disorders are excluded [1]. Within this group, the autosomal dominant cerebellar ataxias (ADCA) are the most likely diagnostic group in an adult with cerebellar ataxia and vertical transmission; they are genetically heterogeneous, share a slowly progressive gait ataxia and overlap clinically with one another [76]. Most ADCA result from CAG triplet expansions encoding polyglutamine tracts; the remainder are due to repeat expansions in non-coding regions or to conventional mutations, that is, nucleotide substitutions, small insertions and deletions, and gene copy-number changes [13, 76]. This last group is the subject of this article.

The order of frequency in European cohorts has been stable since 2015: first the coding CAG expansions (seven genes), then the non-coding expansions and, finally, conventional variants distributed across more than 20 genes; when SCA42 was described, about 40% of ADCA had no molecular diagnosis [13]. In adults with late-onset cerebellar ataxia and no family history, up to 40% remain without an identified genetic cause even after a complete evaluation [1].

1.2. Harding clinical classification

Before the genes were identified, ADCA were grouped by the Harding clinical classification into four types, which Orphanet retains as clinical groups (ADCA type I, ORPHA:94145; type II, ORPHA:208508; type III, ORPHA:94148; type IV, ORPHA:94149) [19]. Type III designates late-onset pure cerebellar ataxia, which includes SCA14 and SCA11 [35, 36]; the SCA11 GeneReviews entry describes type III as the most frequent group of hereditary ataxias and SCA11 as responsible for about 2% of cases of this type [35]. The term olivopontocerebellar atrophy (OPCA), used in the past for some of these families, has fallen out of use [36]. The clinical classification does not predict the gene: the SCA38 GeneReviews entry stresses that SCA cannot be distinguished from one another on clinical criteria alone [47, 77].

1.3. Locus numbering and MDS nomenclature

SCA numbering followed the chronological order in which the loci were mapped. This system accumulated problems: wrongly assigned loci, the same gene with two numbers, unconfirmed loci and genes, a mixture of causal and risk genes, and discordance between phenotype and list [3]. Examples within the class are SCA16, subsumed into SCA15 after the original family was shown to carry an ITPR1 deletion, the empty designations SCA9 and SCA24, the merging of SCA19 and SCA22 into a single gene (KCND3) and the splitting of the former SCA27 into SCA27A (conventional FGF14 variants) and SCA27B (intronic GAA expansion in the same gene) [2, 21, 78]. The Movement Disorder Society Task Force on nomenclature proposed in 2016 a phenotypic prefix followed by the gene name (for example, ATX-CACNA1G for SCA42) and a list of disorders confirmed by explicit criteria; the 2022 update added 89 genes, 38 of them for dominant and recessive ataxias, and created the category of mixed movement disorders [3, 4]. The two nomenclatures coexist in the literature and in laboratory reports, and this article uses both.

1.4. Milestones in molecular discovery

SCA14 was the first non-episodic dominant ataxia attributed to a mechanism other than expansion: in 2003, Chen and colleagues identified three missense variants in the C1 regulatory domain of protein kinase C gamma (PRKCG) in a family linked to 19q13.4-qter, in another familial case and in a sporadic case, with reduced PKCγ immunoreactivity in Purkinje cells [79]. In 2006, Ikeda and colleagues linked SCA5 to in-frame deletions and a missense variant in SPTBN2 (β-III spectrin), in an 11-generation American family descended from the grandparents of President Abraham Lincoln and in two other families [34]. These were followed by TTBK2 in SCA11 (2007), ITPR1 deletions in SCA15 (2007), AFG3L2 in SCA28 (2010), PDYN in SCA23 (2010), KCND3 in SCA19 and SCA22 (2012), and ELOVL5 in SCA38 and TMEM240 in SCA21 (2014) [78, 80–86]. In 2015, Morino and colleagues counted 36 dominant SCA loci and 31 genes, with single-nucleotide variants or small indels in SCA5, 11, 13 to 15, 18 to 21, 23, 26 to 28, 34, 35, 38, 40 and 41 [25].

SCA42 was described simultaneously in 2015 by two independent groups. Coutelier and colleagues, from the French SPATAX network, studied a large family with ADCA and mild pyramidal signs; linkage analysis and exome sequencing identified the CACNA1G variant c.5144G>A p.(Arg1715His), in the voltage sensor of domain IV of the Cav3.1 calcium channel, and screening of 479 index cases found two further families with the same variant [13]. Morino and colleagues, in Hiroshima, arrived at the same variant in two Japanese families with pure adult-onset ataxia [25]. In 2018, Chemin and colleagues described de novo gain-of-function variants in the same gene in children with early cerebellar atrophy and severe developmental impairment, an entity now designated SCA42ND [15].

In the following years, the combination of exome sequencing and network analysis added FAT2 (SCA45) and PLD3 (SCA46) [87], GRM1 (SCA44) [51], PUM1 (SCA47) [54], heterozygous STUB1 (SCA48) [56], SAMD9L (SCA49) [57] and NPTX1 (SCA50) [58]. The review by Coarelli, Coutelier and Durr (2023) lists the genes added in the previous five years and advocates genome sequencing as the diagnostic tool of choice [88].

1.5. Definition of the core entities

SCA42 is the adult-onset, slowly progressive autosomal dominant cerebellar ataxia caused by heterozygous CACNA1G variants that reduce neuronal excitability, p.(Arg1715His) being the recurrent variant [9, 13]. SCA42ND is the allelic form with onset in infancy or early childhood, caused by de novo gain-of-function variants in the intracellular gate of the channel (p.(Ala961Thr), p.(Met1531Val) and others), with cerebellar atrophy or hypoplasia and severe motor and cognitive impairment [15, 89]. Between the two, rare CACNA1G-related phenotypes have been described: developmental and epileptic encephalopathy, episodic vestibulocerebellar ataxia, myoclonus-ataxia with intellectual disability and atypical parkinsonism [5, 31, 90, 91]. The practical rule, formulated in the gene curation, is that the variant, not the gene, determines which disorder is present [9].

The remaining entities in the class are defined by the gene and the effect of the variant; some have allelic recessive forms with distinct phenotypes (SPAX5 in AFG3L2, SCAR14 in SPTBN2, SCAR16 in STUB1, SCAR13 in GRM1, neuro-ichthyosis in ELOVL4) [11, 12, 33, 92, 93]. Several numbered designations rest on a single family or on disputed evidence (SCA26, SCA35, SCA40, SCA41, SCA43, SCA45, SCA46 and SCA50); for each one, this article indicates the level of evidence available [42, 45, 48, 49, 52, 53, 59, 94].

2. Molecular Genetics and Pathophysiology

2.1. Principles common to the class

The common target of SCA due to conventional variants is the Purkinje cell, the sole efferent neuron of the cerebellar cortex, and the circuits linking it to the deep cerebellar nuclei and the inferior olive. The genes in the class fall into functional groups: ion channels and excitability (CACNA1G, KCNC3, KCND3, TRPC3, CACNA1A), calcium and second-messenger signalling (ITPR1, PRKCG, GRM1), membrane cytoskeleton and primary cilium (SPTBN2, TTBK2), mitochondria and protein quality control (AFG3L2, STUB1), very-long-chain fatty acid metabolism (ELOVL4, ELOVL5), translation and RNA regulation (EEF2, PUM1) and neuropeptides or synaptic proteins (PDYN, NPTX1) [77, 87]. The review by Klockgether, Mariotti and Paulson describes three interrelated mechanisms in SCA, proteotoxicity, RNA toxicity and ion channel dysfunction, the third being the most represented in this class [77].

The effect of the variant on the protein is heterogeneous and not always known: haploinsufficiency (ITPR1 deletions in SCA15), dominant-negative effect (most AFG3L2 variants in SCA28, TTBK2 frameshifts in SCA11, ITPR1 variants in Gillespie syndrome), gain of function (GRM1 in SCA44, SAMD9L in SCA49, CACNA1G in SCA42ND) or altered channel gating with a net loss of excitability (CACNA1G in SCA42) [11, 13, 15, 21, 35, 38, 51, 57]. The mechanism has direct diagnostic consequences: when the disease results from haploinsufficiency due to deletion, a test that only sequences misses most cases (SCA15); when it results from a dominant-negative or gain-of-function effect restricted to precise domains, a truncating variant or a missense variant outside those domains does not confirm the diagnosis (SCA11, SCA42) [9, 21, 95].

2.2. CACNA1G and the Cav3.1 channel (SCA42 and the allelic spectrum)

CACNA1G lies at 17q21.33 (OMIM 604065) and encodes the pore-forming α1G subunit of the low-voltage-activated T-type calcium channel Cav3.1 [9]. The channel opens at relatively negative potentials and inactivates rapidly; it underlies low-threshold calcium potentials, neuronal pacemaking, oscillations and rebound burst firing, and is highly expressed in Purkinje cells, the deep cerebellar nuclei, the inferior olive and the thalamus, where it takes part in the slow oscillations of non-rapid eye movement (NREM) sleep [9, 13]. Cacna1g-null mice have no motor phenotype, are resistant to induced absence seizures and have bradycardia and slow atrioventricular conduction, whereas overexpression of the channel causes spike-wave discharges [13]. In the same mice, thalamocortical neurons lose burst firing and retain tonic firing, and loss of the channel in the thalamus is associated with visceral hyperalgesia, which points to a role for Cav3.1 in thalamic sensory filtering [17]. The MANE Select transcript is NM_018896.5 (NP_061496.2); articles also use NM_018896.4, NM_198384.3, the short isoform NM_198382.2 and the cDNA NM_198387.2 [10].

SCA42: loss of excitability. The recurrent variant c.5144G>A p.(Arg1715His) affects a residue of the S4 segment (voltage sensor) of domain IV [13]. In HEK293T cells, it shifts the activation curve towards more positive potentials (V½ −47.20 ± 0.65 mV in the wild type, n=18, versus −43.27 ± 0.73 mV in the mutant, n=16; p<0.001), does not alter the inactivation V½ but increases the slope of the curve (5.39 versus 4.29 mV) and shifts the window current towards more positive potentials, without altering current density or kinetics [13]. Morino and colleagues obtained a shift of activation in the same direction (−57.5 ± 9.1 mV versus −44.7 ± 6.0 mV; p=0.004) and also of the inactivation V½ (−79.1 versus −72.5 mV) [25]. In a computational model of a deep cerebellar nuclear neuron, the variant reduced the number of spikes per burst (4 versus 5) and delayed burst onset, that is, it decreased neuronal excitability [13]. Purkinje cells derived from a patient's induced pluripotent stem cells (iPSC) differentiated without defect, and the subcellular distribution of the channel was normal [25]. The variant was not associated with epilepsy, unlike gain-of-function variants described in other contexts [13].

SCA42ND: gain of function in the intracellular gate. The de novo variants c.2881G>A p.(Ala961Thr) and c.4591A>G p.(Met1531Val) make inactivation about five times slower, shift the inactivation V½ by more than 10 mV towards negative potentials and increase the window current; in a cerebellar nuclear neuron model they increase firing, and the window current is completely inhibited in vitro by the selective T-type channel blocker TTA-P2 [15]. In the original series, p.(Ala961Thr), in the pore-lining S6 segment, was found in three of the four patients, suggesting a mutational hotspot [17, 18]. The series by Qebibo and colleagues (19 patients, 8 novel variants) showed that variants in the S5–S6 segments, which form the intracellular gate, are associated with the most severe phenotypes and behave as gain of function, whereas p.(Met197Arg), in the IS4–S5 linker, reduces function [89]. The electrophysiological classification of 18 variants by Davakan and colleagues confirmed the gain of function of p.(Ala961Thr) and p.(Met1531Val), grouped the intracellular gate variants (p.(Leu208Pro), p.(Ala961Thr), p.(Ile962Asn), p.(Met1531Val)) by slow inactivation and deactivation, showed that p.(Arg1715His) and the de novo variant p.(Arg1718Gly), both in S4, slow activation, and concluded that the net effect depends on the electrical context of the neuron (Purkinje or thalamic firing pattern) [96]. In human thalamocortical assembloids, p.(Met1531Val) alters T-type currents in thalamic neurons and generates correlated thalamocortical hyperactivity [97].

Allelic spectrum. The CACNA1G spectrum includes: (1) inherited adult-onset SCA42 with loss of excitability; (2) early-onset de novo SCA42ND with gain of function; (3) infantile-onset developmental and epileptic encephalopathy without evident cerebellar atrophy, with p.(Leu208Pro) (mixed effects) and p.(Leu909Phe) (gain of function) [90]; (4) rare phenotypes described in isolated reports, namely episodic vestibulocerebellar ataxia (p.(Gly2320Cys)), myoclonus-ataxia with intellectual disability (p.(Asp1279Asn)), rapidly progressive atypical parkinsonism (p.(Arg2043Ter)) and paroxysmal tonic upgaze with congenital ataxia (de novo variant p.(Arg1718Gly)) [5, 31, 91, 96, 98]; (5) homozygosity for a frameshift in three siblings with intellectual disability and cataracts, without ataxia, mentioned by Coutelier and colleagues [13]. The parkinsonism associated with a truncation does not correspond to the missense mechanism established for SCA42 and should be read as an isolated report [5].

Penetrance, expressivity and modifiers. In the French families, all affected individuals tested were heterozygous and no unaffected individual was [13]. In North American family B, a carrier sister reported being asymptomatic without having been examined, which suggests variable penetrance or age at onset [22]. One French index patient also had a variant of uncertain significance (VUS) in SPTBN2 [13], and the Turkish patient with de novo p.(Arg1715His) had a heterozygous SACS variant inherited from the mother [14], which illustrates the co-occurrence of variants in ataxia genes, put forward as a possible explanation for variable penetrance or more severe phenotypes [88].

Weight of variants of uncertain significance. In ClinVar, CACNA1G had 14 pathogenic, 29 likely pathogenic and 817 VUS variants at the date of the record consulted; variants in this gene are frequently reported and rarely causal [9]. The gene is missense-intolerant (missense Z 4.97; observed/expected 598:903.6) [13]. An automated patch-clamp framework calibrated with 25 benign and 16 pathogenic variants, published as a preprint in 2026 and not yet peer reviewed, proposes functional criteria for the gene according to the American College of Medical Genetics and Genomics and Association for Molecular Pathology (ACMG/AMP) rules: more than 80% of pathogenic variants reduced current density, and two of five VUS tested showed an abnormal profile [99].

2.3. Normalisation of CACNA1G variants

The table uses Human Genome Variation Society (HGVS) nomenclature. The conversions below were obtained in VariantValidator (GRCh38) and, for the first three, confirmed in ClinVar [6, 10].

The notation c.4721T>A p.(Met1574Lys) for the Chinese family of Li and colleagues is valid only in isoforms NM_198377.2, NM_198378.2 and NM_198380.2, which contain an alternative exon; in the MANE transcript the same genomic position is intronic, so the variant is recorded as isoform-dependent, without a definitive conversion [10, 100]. Hara and colleagues number p.(Arg1715His) as c.5075G>A in the short isoform NM_198382.2 [30]. In the classification by Davakan and colleagues, protein numbering follows UniProt O43497 [96].

2.4. PRKCG and protein kinase C gamma (SCA14)

PRKCG (19q13.42; gene OMIM 176980) encodes PKCγ, a calcium-, phospholipid- and diacylglycerol-dependent serine/threonine kinase expressed only in neurons of the brain and spinal cord, which phosphorylates glutamate receptor subunits and takes part in synaptic plasticity [36, 101]. Most pathogenic variants cluster in exon 4, which encodes the diacylglycerol-binding C1 regulatory domain [36]. In vitro, the mutant proteins form cytotoxic aggregates in the cytoplasm of Purkinje cells, have altered kinase activity, impair degradation via the ubiquitin-proteasome pathway and lose the ability to inhibit calcium influx through TRPC channels; in the transgenic mouse the Purkinje cell dendritic tree is markedly reduced without evident degeneration [36, 101]. The mechanism is debated: null heterozygotes are neurologically normal, there is evidence of both gain and loss of function, and the nonsense variant p.(Arg76Ter) inhibits PKC phosphorylation and increases apoptosis, suggesting an at least partial dominant-negative effect [36]. Homozygous forms (the intragenic deletion c.530_919del and the stop-codon deletion c.2091_*98del) cause ataxia with intellectual disability, so the gene has two inheritance patterns [36].

In the British series by Chelban and colleagues, the more complex phenotype was associated with missense variants in the catalytic domain (exon 11) [37]. Penetrance is incomplete or age-dependent: there are clinically unaffected carriers older than 60 years in at least three families [36].

2.5. AFG3L2 and the mitochondrial m-AAA protease (SCA28, SPAX5, OPA12)

AFG3L2 (18p11.21; gene OMIM 604581) encodes the catalytic subunit of the m-AAA protease of the inner mitochondrial membrane, which forms homo-oligomers or hetero-oligomers with paraplegin (SPG7), unfolds and degrades mistranslated or misfolded proteins and processes substrates such as MRPL32, PINK1 and paraplegin itself [11, 82, 102]. The gene is associated with four phenotypes: SCA28 (autosomal dominant, the most frequent), autosomal recessive spastic ataxia 5 (SPAX5, OMIM 614487; 14 individuals described), AFG3L2 recessive cerebellar ataxia (two individuals) and optic atrophy 12 (OPA12, OMIM 618977; autosomal dominant) [11]. In SCA28 most variants act through a dominant-negative effect and lie in the central protease loops (exons 15 and 16), destabilising the structure required for substrate unfolding and degradation; in SPAX5 there is biallelic loss of function near the active site; in OPA12 the variants affect the ATPase domain (exons 9 to 12) and disturb OPA1 processing, with mitochondrial fragmentation [11]. The heterozygous variants p.(Met666Arg), p.(Gly671Arg) and p.(Glu700Lys) are associated with childhood onset, and the remaining missense variants with onset between the 2nd and 4th decades [11]. In the yeast studies at the time of gene discovery, one of the variants (E691K, in the OMIM entry notation; = NM_006796.3:c.2071G>A p.(Glu691Lys)) behaved as dominant-negative and others (for example S674L; = NM_006796.3:c.2021_2022delinsTA p.(Ser674Leu)) as loss of function, both with impaired mitochondrial respiration and proteolysis [68, 103]. In SPAX5, homozygosity for the hypomorphic Y616C allele reduces assembly of the protease with paraplegin, which explains a picture combining features of SCA28 and SPG7 spastic paraplegia with mitochondrial signs; in fibroblasts from another patient, reduced AFG3L2 lowered the mitochondrial membrane potential and activated the integrated stress pathway mediated by OMA1 and DELE1 [69].

GeneReviews gives two notations for p.(Val723Met) (c.2167G>A and c.2176G>A); validation on transcript NM_006796.3 confirms c.2167G>A p.(Val723Met) (NC_000018.10:g.12337349C>T) as the correct notation [10, 11]. This variant occurs in compound heterozygosity in the two recessive cases described, with p.(Tyr616Cys) and with p.(Arg632Ter) [11]. Deletions explain less than 1% of cases: deletion of exons 14 to 16 in Belgian families and a 5.3 Mb deletion that includes the whole gene; larger contiguous deletions, such as 18p deletion syndrome, give a more complex phenotype [11]. Penetrance of SCA28 appears complete; in OPA12 reduced penetrance or subclinical phenotypes are suggested [11].

2.6. ITPR1 and the inositol trisphosphate receptor (SCA15/16, SCA29, Gillespie syndrome)

ITPR1 (3p26.1; gene OMIM 147265) has 58 exons and encodes the type 1 inositol 1,4,5-trisphosphate receptor, a homotetramer whose C-terminal domain forms the channel that releases calcium from the endoplasmic reticulum after IP3 binding; it is the dominant IP3 receptor in Purkinje cells, the human cell type most sensitive to its dysfunction [21, 38]. In SCA15, about 96% of variants are deletions of exons or of the whole gene, often with partial deletion of the adjacent SUMF1 (exons 1 to 10, 1 to 40, 1 to 44 or 1 to 48 of ITPR1), and the presumed mechanism is haploinsufficiency; the missense variant c.3176C>T p.(Pro1059Leu) (NM_002222.4) was associated with a Japanese family [21, 81]. A microdeletion including ITPR1 and SUMF1 confers a risk of SCA15, but causes multiple sulfatase deficiency (recessive) only if there is a pathogenic variant on the other SUMF1 allele [21]. SCA29 results from missense variants and Gillespie syndrome from de novo variants with a dominant-negative effect which, in the largest series (13 individuals), affected only three residues (Glu2094, Gly2539 and the in-frame deletion of Lys2596); iris hypoplasia requires complete loss of activity or disruption of the tetramer, because null heterozygous mice have no iris defect [38]. The apparent anticipation in SCA15 (22 parent-child pairs with onset 9 years earlier in the children, with wide scatter) is considered insufficient to draw conclusions and unlikely in a deletion mechanism [21]. The designation SCA16 disappeared when the Japanese family that defined it, initially linked to 8q, showed linkage to 3p26 and a deletion of exons 1 to 48 of ITPR1 [64]. In SCA29, the missense variants described mainly affect the coupling and regulatory domain, with loss of function, but p.(Arg36Cys) in the N-terminal suppressor domain, de novo in a mother and transmitted to her two children, increases IP3 affinity and alters the calcium signal, indicating gain of function [70].

2.7. KCNC3 and the Kv3.3 channel (SCA13)

KCNC3 (19q13.33; gene OMIM 176264) encodes the voltage-gated potassium channel Kv3.3, of the Shaw subfamily, which forms heterotetramers with Kv3.1, opens late during depolarisation, deactivates rapidly and regulates the frequency, shape and duration of action potentials in fast-spiking neurons, including Purkinje cells; it also interacts with the actin cytoskeleton in growth cones, which is consistent with congenital hypoplasia [24, 104]. The variants act through a dominant-negative effect or gain of function, which is why deletions or duplications are not expected [24]. The genotype-phenotype correlation rests on a few families: p.(Arg420His) (c.1259G>A, validated on NM_004977.3) produces a channel that is expressed but non-functional and is associated with progressive adult ataxia; p.(Arg423His) slows activation and is associated with congenital cerebellar hypoplasia with epilepsy; p.(Phe448Leu) slows closure and is associated with childhood onset with intellectual disability; p.(Val535Met) is associated with childhood onset with mild intellectual disability; and the in-frame deletion p.(Pro583_Pro585del) with a rapidly progressive adult phenotype [10, 24]. Penetrance appears complete [24].

2.8. TTBK2 and the primary cilium (SCA11)

TTBK2 (15q15.2; gene OMIM 611695) encodes tau tubulin kinase 2, which removes CP110 from the mother centriole to initiate ciliary axoneme assembly, phosphorylates tau and TDP-43 and regulates synaptic vesicle trafficking [35, 95]. All families with confirmed SCA11 have small frameshifting deletions or insertions clustered in one exon (exon 11 in GeneReviews, exon 12 in Bauer and colleagues, a discrepancy in exon numbering), which truncate the protein downstream of the kinase domain; part of the messenger RNA escapes nonsense-mediated decay and the truncated protein interferes dominantly with ciliogenesis [35, 95]. The variant c.1306_1307delGA, described as p.Asp435TyrfsTer14 in GeneReviews, validates on NM_173500.4 as c.1306_1307del p.(Asp436TyrfsTer14), because nucleotide c.1306 is the first of codon 436 [10, 35]. The published TTBK2 missense variants (including p.(Val1097Ala) in a Chinese family) were benign or lack functional validation, so a missense variant should not be read as a diagnosis of SCA11 [35, 95]. Penetrance appears complete [35].

2.9. ELOVL5 and ELOVL4: fatty acid elongation (SCA38 and SCA34)

ELOVL5 (6p12.1; gene OMIM 611805) encodes an endoplasmic reticulum elongase that catalyses the rate-limiting step in the elongation of ω3 and ω6 polyunsaturated fatty acids, including arachidonic acid and DHA, and is highly expressed in Purkinje cells [29, 47, 105]. In SCA38 serum arachidonic acid and DHA levels are reduced to 30–40% of controls, the mutant protein p.(Gly230Val) accumulates in the perinuclear region and the mechanism combines loss and gain of function [47, 106]. Only two pathogenic variants have been described: c.689G>T p.(Gly230Val) in three Italian families (at least two with a shared haplotype), validated on NM_021814.5, and c.214C>G p.(Leu72Val) in a French family [10, 85]. The variant NM_001242828.1:c.304C>T p.(Gln102Ter) (designation published in GeneReviews, on an alternative transcript of which a more recent version exists) lies in an unvalidated alternative exon, is intronic in the main transcript and is benign, which is a pitfall in laboratory interpretation [47]. Penetrance is 100% by 50 years of age, without anticipation [47].

ELOVL4 (6q14.1; gene OMIM 605512) catalyses the condensation step in the elongation of saturated and polyunsaturated very-long-chain fatty acids, used in the membrane lipids of the retina, epidermis and central nervous system [93]. It has three presentations: Stargardt macular dystrophy type 3 (autosomal dominant, due to last-exon frameshifts with a dominant-negative effect), SCA34 (autosomal dominant) and severe congenital neuro-ichthyosis (biallelic) [93, 106]. In SCA34, the variant NM_022726.4:c.698C>T p.(Thr233Met) was reclassified from VUS to likely pathogenic on the basis of segregation in two Australian families [44].

2.10. KCND3 (SCA19/22)

KCND3 (1p13.2; gene OMIM 605411) encodes the potassium channel Kv4.3, responsible for the A-type current in cerebellar neurons [39]. Ataxia variants such as p.(Thr352Pro), p.(Met373Ile) and p.(Ser390Asn) are retained in the endoplasmic reticulum with reduced stability, and the in-frame deletion p.(Phe227del), described in Chinese and French families under the SCA22 label, does not conduct A-type current, which led to the merging of the two designations [78, 84]. GeneReviews still lists Brugada syndrome as an allelic disorder, but ClinGen classified this gene-disease relationship as disputed and the proposed Brugada variants were gain-of-function; the article makes no cardiac claims about carriers of KCND3 ataxia variants [1, 39].

2.11. STUB1 (SCA48 and SCAR16)

STUB1 (16p13.3; gene OMIM 607207) encodes CHIP, an E3 ubiquitin ligase and co-chaperone [12]. Biallelic variants cause SCAR16 (OMIM 615768), sometimes with hypogonadotropic hypogonadism (Gordon Holmes syndrome); heterozygous variants cause SCA48 (OMIM 618093) [12, 19]. SCA48 variants impair ligase activity and oligomerisation as much as recessive variants, the same variant can occur with a dominant or recessive presentation, and neuropathology shows ubiquitin- and p62-positive neuronal inclusions in the cerebellum, neocortex and brainstem; the authors describe an SCA48/SCAR16 continuum [12]. A third configuration is digenic: TBP alleles with an intermediate CAG/CAA repeat (about 41 to 49) are only symptomatic together with a heterozygous STUB1 variant (SCA17-DI), with a Huntington-like phenotype [12]. Among 31 index cases with intermediate TBP alleles, 27 (87%) also had a heterozygous STUB1 variant, all with 41 to 46 repeats; in the families, only carriers of both alterations had disease, and the 8 carriers of only one of them were asymptomatic [74]. These authors concluded that an isolated heterozygous STUB1 variant does not cause disease and that digenic SCA17 and SCA48 may be the same entity, noting that the range of TBP alleles considered pathogenic differs between laboratories and that patients diagnosed with SCA48 may have intermediate TBP alleles that were undetected or disregarded [74]. Other SCA48 series excluded TBP expansions in most families, and some cases had additional variants (for example, a PRKCG missense variant in a woman with onset at 27 years), so the question remains open [74].

2.12. SPTBN2 (SCA5, SCAR14 and dominant congenital ataxia)

SPTBN2 (11q13.2; gene OMIM 604985) encodes β-III spectrin of the Purkinje cell membrane skeleton, required to stabilise the glutamate transporter EAAT4 in the membrane [33, 34]. SCA5 results from in-frame deletions and missense variants in the actin- and ARP1-binding domain [34]. Biallelic loss of function causes SCAR14 (also called SPARCA1), with psychomotor delay and early-onset ataxia, and heterozygous de novo missense variants cause non-progressive congenital ataxia with delay, sometimes diagnosed as ataxic cerebral palsy [33]. For the missense variant p.(Arg480Trp), the literature is divided between an adult SCA5 phenotype and a childhood ataxia; the position of the variant and segregation count for more than the label [33].

2.13. Other genes in the class

FGF14 (13q33.1) encodes an intracellular factor that binds the α subunits of voltage-gated sodium channels at the axon initial segment and modulates the sodium current; SCA27A results from heterozygous loss-of-function variants (point, frameshift or structural, including contiguous 13q33 deletions with ITGBL1), whereas SCA27B results from the GAA expansion in intron 1 [2, 43]. TMEM240 (1p36.33) has an unknown normal function; the mutant protein impairs lysosomal autophagy and dendritic development of Purkinje cells in culture, and p.(Pro170Leu) is recurrent [40, 86]. PDYN (20p13) variants in SCA23 cluster in the dynorphin A domain and produce excess, more neurotoxic peptides; the p.R212W knock-in mouse develops glutamate-mediated Purkinje cell loss [41, 83]. In SCA26, the EEF2 missense variant p.(Pro596His) increases, in yeast, −1 ribosomal frameshifting and sensitivity to proteostatic stress [107]. SCA40 (CCDC88C p.(Arg464His)) activates the JNK pathway and apoptosis; the recessive form of the same gene causes congenital hydrocephalus [48, 108]. In SCA41, a TRPC3 variant in the TRP domain was proposed as a toxic gain of function, similar to the moonwalker mouse [49, 109].

SCA44 results from GRM1 missense variants that increase the constitutive activity of the metabotropic glutamate receptor mGluR1 (gain of function) or, in one case, from a de novo truncation with a dominant-negative effect; biallelic loss of function causes SCAR13, described in a Roma founder sub-isolate [51, 92]. In PUM1 (1p35.2), severity follows the reduction in protein: about 25% in adult ataxia with incomplete penetrance (SCA47) and about 50% in the developmental syndrome PADDAS, with derepression of targets such as ATXN1 [54, 110]. In 2025, the four affected members of the French-Canadian family in which SCA47 had been proposed were found to have a pathogenic FGF14 GAA expansion, which makes it unlikely that the PUM1 variant causes adult ataxia in this family; the association with PADDAS is not affected by this reassignment [55]. In SCA49, SAMD9L (7q21.2) variants produce a hyperactive protein with anticodon nuclease activity that blocks protein synthesis and has an antiproliferative effect; somatic genetic rescue in the bone marrow explains the discordance between haematological and neurological penetrance [57, 111]. NPTX1 variants in SCA50 (p.(Gly389Arg), recurrent in five families without a common founder, and p.(Glu327Gly)) induce ATF6-mediated endoplasmic reticulum stress [58]. SCA20 is due to a 260 kb duplication at 11q12.2–q12.3, in direct orientation, with ten known genes according to GeneReviews (at least 12 according to the OMIM entry) [66], absent in 1129 controls; the suspected mechanism is dosage, DAGLA is the most cited candidate, and it has not been excluded that the duplication is merely a marker in linkage disequilibrium [8].

2.14. Level of gene-disease evidence

Not all numbered designations are equally robust. SCA35 (TGM6) is disputed: the cumulative frequency of the variants described as pathogenic in gnomAD is at least 111 times higher than the prevalence of all dominant SCA, the gene tolerates loss-of-function variants, there are Chinese carriers without ataxia, and p.(Leu517Trp) did not segregate in a family explained by an ATN1 expansion [45, 46]. SCA26 (EEF2), SCA40 (CCDC88C), SCA41 (TRPC3), SCA43 (MME) and SCA46 (PLD3) rest on a single family or a single case and retain the provisional «?» sign in OMIM or a limited classification in GenCC [42, 48, 49, 53, 94]. SCA45 (FAT2) has a limited classification from most curators, and SCA50 (NPTX1) was classified as limited evidence by ClinGen in September 2024 [52, 59]. For SCA14, several GenCC sources assign definitive or strong evidence [101].

3. Clinical Spectrum and System-Specific Manifestations

3.1. Picture common to the dominant cerebellar ataxias

Many primary ataxias present between 30 and 50 years of age [1]. The first symptom is, in most cases, a slowly progressive gait disturbance with unexpected falls; other patients begin with imbalance described as «dizziness», which may lead to peripheral vestibular investigation, with clumsiness or tremor of the hands, which can be mistaken for essential tremor or parkinsonism, with dysarthria or choking, or, more rarely, with diplopia [1]. Symptoms are initially intermittent, at the end of the day, with fatigue or after alcohol, and later become constant [1]. Cerebellar examination shows a wide-based gait with difficulty in tandem walking, truncal instability when sitting, dysmetria and terminal tremor, dysdiadochokinesia, slow scanning dysarthria and oculomotor abnormalities (saccadic intrusions in primary position, horizontal or vertical nystagmus, saccadic hypermetria) [1]. MRI confirms cerebellar atrophy or hypoplasia [1].

In hereditary ataxias in general, walking aids become necessary 5 to 10 years after onset, followed by a wheelchair; choking, falls with head injury or hip fracture and, in the late stages, infections (aspiration pneumonia, urinary tract infection, pressure ulcers) are the main causes of morbidity and mortality [1]. The GeneReviews statement that survival rarely exceeds 25 years after onset refers to the most frequent hereditary ataxias and does not apply to most SCA due to conventional variants, in which life expectancy is normal or near normal (SCA11, SCA14, SCA15, SCA28, SCA38) [1, 11, 21, 35, 36, 47]. The available natural history data come from families and small series, without disease-modifying treatment; treatment reports concern symptomatic drugs in individual patients.

3.2. SCA42 (CACNA1G, adult onset)

French cohort. In the ten affected individuals of the three French families, age at onset ranged from 9 to 78 years (20, 68, 41, 9, 19, 18, 37, 40, 78 and 30 years) [13]. The initial symptom was gait instability in 9/10, with vertigo in two. Disability remained mild to moderate even after decades: Spinocerebellar Degeneration Functional Score (SDFS, 0 to 7) between 1 and 5, with SDFS 4 after 53 years of disease and 5 after 32 to 37 years; the documented Scale for the Assessment and Rating of Ataxia (SARA, 0 to 40) ranged from 4 to 21 [13]. There were oculomotor signs in 7/10 (saccadic pursuit, horizontal nystagmus, transient diplopia, hypometric saccades, limited upgaze, square-wave jerks), pyramidal signs in 5/10 (from brisk reflexes to spastic gait with a Babinski sign), depression in 3/10 and cognitive impairment in 2/10; other findings were orbicularis myokymia, postural tremor of the upper limbs and head tremor, dysphagia, scoliosis, erectile dysfunction and reduced vibration sense at the ankles. None had epilepsy [13]. MRI in five patients showed predominantly vermian cerebellar atrophy with a normal pons; in one case there was hypoplasia or atrophy of the cerebellum and brainstem [13].

Neuropathology. In a woman who died at 83 years there was severe atrophy of the vermis and mild atrophy of the hemispheres, Purkinje cell loss with «empty baskets» and Bergmann gliosis more marked in the vermis, fewer glomeruli in the granular layer, abundant polyglucosan bodies, normal dentate nucleus and pontine nuclei, neuronal loss and gliosis of the inferior olive and a normal substantia nigra; Alzheimer disease coexisted (Thal phase 3, Braak stage V), whose relationship with the variant remained undetermined [13].

Japanese cohorts. In the two Hiroshima families, onset ranged from 20 to 70 years (family 1) and from 18 to 57 years (family 2), with slow progression in all; the proband had pure ataxia beginning at 45 years, with gait ataxia, dysarthria and horizontal gaze-evoked nystagmus, without cognitive impairment, epilepsy, amyotrophy, involuntary movements or parkinsonism [25]. Two patients in family 2 had prominent rest tremor, and the pyramidal signs frequent in the French families were not prominent; MRI showed cerebellar atrophy with a normal cerebrum and brainstem [25]. In a Hokkaido series, 2 of 84 families had p.(Arg1715His), with slowly progressive pure cerebellar ataxia [23].

Other populations. In a family of Italian origin studied in the USA, onset ranged from 22 to 58 years (mean 37.5), with a subjective sensation of leg weakness and imbalance followed by gait and limb ataxia with dysarthria, without nystagmus, pyramidal or extrapyramidal signs, sensory changes or cognitive decline; all had cerebellar atrophy [22]. In an Eastern European family the proband's onset was at 67 years and a second cousin's at 39 years [22]. A man from Yemen assessed in Israel began at 25 years with imbalance and neck jerks, progressing to tremor with cervical dystonia (right laterocollis) and mild ataxia, with mild vermian atrophy [22]. In South Korea, two probands from unrelated families began at 25 and 48 years, with SARA scores of 9 and 12.5 [112]. In a Chinese family with an isoform-dependent variant, the three affected individuals began between 43 and 60 years, with pure ataxia without nystagmus, hypotonia and mild hyporeflexia, and MRI of the proband and her sister showed the hot cross bun sign in the pons, a finding more associated with multiple system atrophy [100].

Psychiatry, cognition and developmental onset. A Turkish boy with de novo p.(Arg1715His) had macrocephaly in infancy, walked at 24 months, said his first words at 2.5–3 years, had enuresis until 9–10 years, received special education from 4 years, had a speech fluency disorder, social anxiety and phobias from 6 years and a full-scale intelligence quotient (IQ) of 50 at 8 years; at 14 years falls, a wide-based gait and dysmetria appeared, and MRI showed inferior vermian hypoplasia [14]. The authors frame the symptoms within the cerebellar cognitive affective syndrome (CCAS) described by Schmahmann [14]. Other reports add restless legs syndrome in a Korean patient [112] and describe, within the ATX-CACNA1G spectrum, peripheral neuropathy, chorea, dystonia, cognitive impairment, depression, aggressiveness and delusions [5].

3.3. Early-onset forms and rare CACNA1G phenotypes

SCA42ND. The four patients in the series by Chemin and colleagues had severe motor and cognitive impairment and cerebellar atrophy, with variable facial dysmorphism, digital anomalies, microcephaly and epilepsy [15]. They were four unrelated girls, aged 3 to 13 years at the time of the report, who presented soon after birth or in early infancy with severe hypotonia; this was followed by psychomotor delay, severe to profound intellectual disability, absent speech or speech limited to a few words, poor eye contact and head control, axial hypotonia with variable spasticity, hypertonia or dystonia of the limbs, and inability to walk independently in all [18]. All had cerebellar ataxia and two dysmetria; three had oculomotor apraxia with strabismus and the fourth hypermetropia; two had mild microcephaly (−2 and −2.5 standard deviations) and two early-onset seizures compatible with epileptic encephalopathy; dysmorphic features included small or upslanting palpebral fissures, hypertelorism, deep-set eyes, a short nose, prognathism, anteverted ears and abnormalities of the hair and posterior hairline [18]. In Italy, four unrelated children with p.(Ala961Thr) or p.(Met1531Val) shared vermian atrophy or hypoplasia, hypertonia, psychomotor and speech delay, severe intellectual disability, ophthalmological findings and dysmorphic or ectodermal findings, constituting a homogeneous neurodevelopmental syndrome [28]. In a Japanese study of 348 children with neurodevelopmental disorders, one with p.(Ala961Thr) had a previous clinical diagnosis of Rett syndrome and another with p.(Met1531Val) of West syndrome, both without independent walking [113]. The series of 19 patients by Qebibo and colleagues consolidated the description of these two variants and broadened the spectrum to novel variants, mostly de novo [89]. A Turkish child with de novo p.(Arg1553Gln) had global delay, autism spectrum disorder, progressive gait instability, macrocephaly and cerebellar and vermian hypoplasia, without seizures [114].

Developmental and epileptic encephalopathy. The variants p.(Leu208Pro) and p.(Leu909Phe) were associated with infantile-onset encephalopathy, with axial hypotonia, delay and severe to profound cognitive impairment, without evident cerebellar atrophy; seizures were transient with the former and initially refractory, with ataxia, with the latter [90].

Congenital ataxia with tonic gaze deviation. A 2-month-old infant with a pathogenic de novo variant had paroxysmal tonic upgaze, paroxysmal torticollis and congenital ataxia, documented on video [98].

Episodic vestibulocerebellar ataxia. In a Spanish family with four affected individuals across three generations and the variant p.(Gly2320Cys), the proband had, from 28 years of age, episodes of dizziness and unsteadiness (lasting 3 minutes to 2 hours, weekly, triggered by head rotation, anxiety, exhaustion, stress or sleep deprivation) and three prolonged episodes with headache, facial numbness and dysarthria; the video head impulse test (vHIT) showed reduced gain in all six semicircular canals and caloric responses were markedly reduced; the mother had non-progressive episodes for 27 years [91].

Myoclonus-ataxia with intellectual disability. A 53-year-old Italian man, without family history, with learning difficulties since childhood and involuntary movements of the upper limbs, had ataxic gait, moderate dysarthria, square-wave saccadic intrusions, action tremor and cortical myoclonus, SARA 9 and a full-scale IQ of 52, with MRI without cerebellar atrophy; he had the likely pathogenic variant p.(Asp1279Asn), in segment S1 of domain III [31].

Atypical parkinsonism. A 45-year-old Argentinian man, without family history, developed at 41 years unsteadiness and falls, symmetrical bradykinesia with micrographia, tremor and segmental dystonia, with dream enactment behaviour, constipation and urinary incontinence; after 4 years of disease he had severe freezing of gait, severe dysarthria and dysphagia, camptocormia and pyramidal signs, and was in a wheelchair, with normal MRI and a poor response to levodopa; he had the nonsense variant p.(Arg2043Ter) [5].

Sporadic adult ataxia. In the German SPORTAX cohort, a patient with the likely pathogenic variant p.(Gly1619Ser) had ataxia with absent ankle reflexes, distal muscle atrophy, saccadic pursuit, gaze-evoked nystagmus, dysphagia, urinary dysfunction and mild cognitive impairment, with a second probable variant in PEX26 [115]. In the series by Davakan and colleagues, novel variants accompanied pictures such as adult neuromuscular disease with ptosis and ataxia, congenital ataxia co-inherited with a CACNA1A variant, or progressive ataxia with ophthalmoplegia and olivopontocerebellar atrophy [96].

3.4. SCA14 (PRKCG)

More than 60 individuals or families have been described [36]. Onset is usually in the 4th decade, with a range of 3 to 70 years [36]. The frequencies compiled by GeneReviews are: gait ataxia in 100% (initial symptom in most), dysarthria in 70%, oculomotor abnormalities in 60% (nystagmus, saccadic pursuit), pyramidal signs in 60%, mild to moderate cognitive impairment in 20%, reduced sensation, mainly vibration, in 17%, tremor in 15%, myoclonus in 10%, depression in seven families or individuals and hearing impairment in four individuals; mild to moderate, predominantly midline cerebellar atrophy is present in all [36]. In the British series of 25 individuals from 13 families, progression was slow, rarely with severe disability, and about one-third had complex ataxia with focal or task-induced dystonia, peripheral neuropathy, parkinsonism, myoclonus or a pyramidal syndrome [37].

Episodic axial myoclonus (irregular jerky movements of the trunk and head lasting minutes to hours) was described in an early-onset Japanese family, and PRKCG variants have been found in progressive myoclonic ataxia and myoclonus-dystonia [36]. Cognition varies between families: intellectual, attentional and executive deficits in 13/18 (72%) in a French family, mild deficits in a few in other series and no significant deficit in ten Norwegian patients compared with controls from the same family [36]. Almost all remain ambulant, many with falls and using a stick, and life expectancy is not reduced; many live beyond 70 years [36]. Neuropathology shows Purkinje cell loss with PKCγ mislocalised in large cytoplasmic aggregates [36]. In the six-generation Dutch family with p.(Gly118Asp), mean onset was at 40 years (21 to 59), and the two patients with onset in the third decade had task-induced focal dystonia; two unaffected members were carriers [63]. In the Japanese family in which the locus was mapped, patients with onset up to 27 years began with intermittent axial myoclonus, and a 76-year-old obligate carrier and a 44-year-old relative had no symptoms [63]. In a five-generation British family, mean onset was at 37 years (15 to 68), some patients had ataxia only on examination and the overall course was benign; the diagnosis was made by exome sequencing after 13 negative SCA genes [63]. Among French families with dominant ataxia, SCA14 explained 7 of 454 (1.5%) [63].

3.5. SCA28 and the AFG3L2 spectrum

SCA28 begins at a mean of 26.5 ± 17.2 years, with a range from birth to 74 years (GeneReviews gives an upper limit of 76 years in another section) [11]. Most begin with lower limb incoordination; less often the first sign is ptosis, ophthalmoplegia, dysarthria, upper limb incoordination or spasticity [11]. Over the course of the disease, nystagmus appears in about 50% (sometimes with oscillopsia), horizontal and vertical ophthalmoparesis in about 50%, ptosis in about 30%, reduced vibration sense at the ankles in about 20%, other movement disorders (parkinsonism, dystonia, chorea, myoclonus) in about 15%, cognitive decline in about 25% and psychiatric manifestations (depression, behavioural disorders) in about 20%; the pyramidal syndrome becomes frequent with disease duration [11]. MRI shows predominant atrophy of the superior vermis with a spared brainstem and, sometimes, abnormal signal in the globus pallidus [11]. Most remain ambulant and independent throughout life, and life expectancy is not limited [11]. In the Italian family in which the locus was mapped (at least 14 affected across 4 generations), mean onset was at 19.5 years (12 to 36), and the oculomotor pattern changed with disease duration: gaze-evoked nystagmus in the first years and, later, dysmetric and slow saccades, ophthalmoparesis and ptosis, without sensory or cognitive involvement [68]. In a European series of 366 probands with dominant ataxia without common expansions, 9 (2.6%) had missense variants in exons 15 and 16 of AFG3L2, without pathogenic CNVs; in these patients (8 French families and 1 Italian) mean onset was at 30.7 years (6 to 60), with dysarthria in 68%, nystagmus in 54%, ophthalmoplegia in 48% and ptosis in 42%, six with a complete pyramidal syndrome, and the slow progression did not cause major functional disability [68].

SPAX5 ranges from severe infantile neurodegeneration (microcephaly, regression at around 9 months, death at 2.5 years) to early-onset complex ataxia with myoclonic epilepsy, spasticity, dystonia, chorea, neuropathy or optic atrophy [11]. In the original description, two Colombian siblings born to consanguineous parents had spastic gait from 2 years of age (the younger never walked unaided and died of pneumonia at 13 years), progressive myoclonic epilepsy with tonic-clonic seizures from 8 years, dysarthria, dysphagia, ptosis, oculomotor apraxia, dystonia and sensorimotor axonal neuropathy, with normal cognition; muscle showed mispositioned mitochondria with lipid droplets and reduced mitochondrial DNA copy number, and the carrier mother had mild cerebellar atrophy on MRI without symptoms [69]. Two unrelated Italian patients, homozygous for M625I (= NM_006796.3:c.1875G>A p.(Met625Ile), with conflicting classifications in ClinVar [103]) and identified in a cohort of 84 patients with progressive myoclonic epilepsy, had severe myoclonus from 10 years after normal development, ataxia and mild cognitive decline, and became wheelchair users at 22 and 48 years, without early spasticity or neuropathy [69]. The two non-spastic recessive cases had ataxia, ptosis and dysarthria from 20 years and ataxia with ptosis and ophthalmoparesis from 55 years [11]. OPA12 manifests as progressive optic atrophy, usually before 20 years, with acuity between 0.2/10 and 2/10, photophobia and dyschromatopsia, indistinguishable from OPA1 dominant optic atrophy, without ataxia; it accounts for about 4% of OPA1-negative optic atrophies [11].

3.6. SCA15/16, SCA29 and Gillespie syndrome (ITPR1)

The GeneReviews entry on SCA15, withdrawn in 2024 as out of date, gathered 67 affected individuals from 20 families (Australia, Japan, France, Germany, Italy, United Kingdom) with onset between 7 and 72 years [21]. Gait ataxia is very slowly progressive: nine Australians were independent after a mean of 31 years of disease, 13 French patients after 22 years, and 8/10 Germans were ambulant after 9 years; one person became a wheelchair user after 40 years [21]. Postural and kinetic tremor is frequent and may be the presenting sign, head or truncal tremor occurs in less than 30%, gaze-evoked nystagmus in about 80% (60% in another section of the same entry), and vestibulo-ocular reflex gain is reduced in about half the patients of two families; mild dysphagia appears after two decades and cognition is usually preserved [21]. MRI shows atrophy of the rostral and dorsal vermis with sparing of the hemispheres and brainstem [21].

In the series published after the GeneReviews withdrawal, five German families had slowly progressive ataxia with vermian atrophy and needed a walking frame or wheelchair 15 to 17 years after onset, with action and postural tremor in 7/10, intention tremor in all and mild pyramidal or posterior column signs; in six European families, onset was usually with gait ataxia, followed by oculomotor abnormalities and dysarthria, with pyramidal signs in one family and executive decline in another [64]. The OMIM entry considers SCA15 the most frequent non-repeat SCA in Central Europe, based on 5 of 56 German families with dominant SCA without expansions [64].

SCA29 manifests from birth with motor delay, hypotonia and mild cognitive delay, evolving into static or very slowly progressive gait and limb ataxia, with cerebellar atrophy and variable nystagmus, dysarthria and tremor [38]. In an Australian family with at least 20 affected individuals across four generations, all had congenital ataxia or late, wide-based walking, cognitive impairment was more disabling than the ataxia, five reported improved coordination with age and only 3 of 7 studied had vermian atrophy; in a Canadian family, the proband had psychomotor delay at 8 months, was not walking at 28 months and developed complex focal seizures at 5 years, controlled with medication [70]. In the family with p.(Arg36Cys), none of the three had nystagmus, all had pinpoint miotic pupils, and MRI of the mother and son showed no cerebellar atrophy; two children with a clinical diagnosis of ataxic cerebral palsy had de novo ITPR1 variants with normal MRI [70]. Dominant non-progressive congenital ataxia is genetically heterogeneous: a family with upbeat nystagmus and late episodes of vertigo and ataxia excluded linkage to 3p [70]. Gillespie syndrome combines bilateral partial iris hypoplasia, with a scalloped pupillary margin, congenital hypotonia, non-progressive ataxia, progressive cerebellar atrophy and intellectual disability [38].

3.7. SCA13 (KCNC3)

There are three phenotypes [24]. The congenital phenotype has cerebellar hypoplasia with non-progressive limb, truncal and gait ataxia, dysarthria, tremor, motor delay and mild to moderate cognitive impairment relatively specific to language, sometimes epilepsy or psychiatric manifestations, with motor and cognitive improvement over the lifespan. The childhood-onset phenotype has slowly progressive cerebellar atrophy and ataxia, with moderate intellectual disability (IQ 62 to 76 in the family described in 2000) and a life expectancy that is not shortened. The adult phenotype, associated with p.(Arg420His), has slowly progressive ataxia, mild cognitive impairment, few oculomotor signs and, in some, myoclonus; cerebellar atrophy may precede symptoms [24]. A form with onset at 30 years and progression to severe disease within 10 to 15 years accompanies the deletion p.(Pro583_Pro585del) [24]. In the Filipino family with p.(Arg420His), 21 people were carriers and 9 did not yet have symptoms; in the 12 symptomatic individuals, mean onset was at 37.3 years (25 to 60), all had gait and limb ataxia (limited to the lower limbs in two), most had dysarthria, six reduced vibration sense in the toes, five brisk reflexes, none oculomotor or extrapyramidal abnormalities, three used a stick after many years and only a 73-year-old woman was in a wheelchair [62]. Two European sisters with the same variant also had hearing loss or optic atrophy, and a 12-year-old boy with a clinical diagnosis of ataxic cerebral palsy had the de novo variant p.(Thr428Ile), which renders the channel non-functional with a dominant-negative effect [62]. In pregnancy, weight gain may worsen gait ataxia [24].

3.8. SCA11 (TTBK2)

Twenty-eight individuals from six families (United Kingdom, Pakistan, France, Germany, Denmark and China) have been described [35]. All had cerebellar ataxia; 21/28 limb ataxia, 22/28 dysarthria, 20/28 nystagmus, 18/28 saccadic pursuit, 18/28 hyperreflexia, 4/28 diplopia, 2/28 ophthalmoplegia and 1/28 extrapyramidal signs [35]. Onset ranged from 9 years (Danish family, with a sibling affected at 4 years) to 40–50 years; half of the British patients with vertical nystagmus had upbeat nystagmus [35]. Mean disease duration in the British family was 26.8 years; 8/17 British and Pakistani patients were wheelchair users 20 to 30 years after onset, whereas the French and German patients remained active for many years; many lived beyond 75 years and the nine deaths occurred between 55 and 88 years [35]. Neuropathy is uncommon, except in the Chinese family, and neuropathology showed abnormal tau deposits in the brainstem and cortex [35, 80]. The British family in which the gene was identified, from Devon, spanned eight generations, and the second family, of Pakistani origin, had a different small TTBK2 deletion [61].

3.9. SCA38 (ELOVL5)

In the 21 affected individuals of three Italian families, onset was between 26 and 50 years, with nystagmus in 21/21 and gait ataxia in 20/21 as the presentation, followed by limb ataxia, dysarthria, dysphagia, horizontal diplopia and, late, ophthalmoparesis and ophthalmoplegia and sensorimotor axonal polyneuropathy [47]. The distinctive features are pes cavus without weakness in 14/17 (82%), hyposmia in 13/17 (76%), hearing loss in 7/21 (33%), abnormal auditory evoked potentials in 11/12 and anxiety disorder in 7/21 [47]. All were independent in the first decade of symptoms; in the second decade (n=13), 70% used a stick; in the third (n=8), 87% were wheelchair users and 50% had dysphagia; there is no dementia and there are no extrapyramidal signs [47]. MRI shows mainly vermian atrophy with sparing of the cortex and brainstem, and fluorodeoxyglucose positron emission tomography (FDG-PET) shows cerebellar hypometabolism [47]. In three deceased patients, the mean time from onset to death was 41 years [47]. In the patients of the original description, serum arachidonic acid and DHA were reduced, and the mutant proteins accumulated in the Golgi apparatus and increased CHOP, a marker of the unfolded protein response; screening of 456 European SCA probands revealed three additional families [72].

3.10. SCA34 (ELOVL4)

In a review of 60 cases, gait ataxia was present in 88.3%, limb ataxia in 76.7%, dysarthria in 63.3% and nystagmus in 58.3%; erythrokeratodermia variabilis, classically associated, was present in only 33.3%; there were also pyramidal signs, dysautonomia, retinitis pigmentosa and cognitive impairment; cerebellar atrophy was universal and the hot cross bun sign appeared in 32.4% [44]. SCA34 should be considered regardless of skin changes [44]. In the five-generation French-Canadian family (25 affected) described in 1972, the ichthyosiform erythematous plaques appeared shortly after birth, improved in summer and almost always disappeared by 25 years, and ataxia with hyporeflexia, nystagmus and dysarthria began after 40 years, with late wheelchair use and no cognitive or sensory deficit; four carrier relatives were unaffected (incomplete penetrance) [71]. In two Japanese families without skin lesions, mean onset was at 33.9 years (13 to 56), almost all managed without a stick until 60 years, 8 of 9 had a Babinski sign and 4 the hot cross bun sign in the pons [71].

3.11. SCA20 (11q12 duplication)

The only family described, Australian of Anglo-Celtic ancestry, had 16 affected individuals with onset between 19 and 64 years (mean 47) [8]. In 10/16 the disease began with dysarthria without ataxia, often resembling adductor spasmodic dysphonia, and dysarthria preceded ataxia in about two-thirds [8]. Palatal tremor without ear click was present in 10/16, downward hypermetric saccades in 10/16 and minor pyramidal signs in 5/16; dentate nucleus calcification on computed tomography (CT) was present in five affected individuals with 5 years or less of symptoms, without basal ganglia calcification in 9/11 [8]. Progression was slow: one patient used a wheelchair after 40 years of symptoms and another had a gastrostomy after 15 years [8]. In the initial description of 14 affected individuals, the mean and median age at onset was 46.5 years, about 10 years earlier in children than in affected parents, and dentate calcification was present in the 9 patients studied by CT [66]. In Portugal, six families (19 affected) were described with dominant ataxia preceded by daily spasmodic coughing fits lasting several minutes, beginning between 25 and 55 years, with ataxia between 40 and 65 years, downbeat nystagmus in 10 patients with long disease duration and dentate calcification in one; the authors acknowledged a similarity to SCA20, but the OMIM entry does not mention the 11q12 duplication in these families [66].

3.12. SCA48 and SCAR16 (STUB1)

SCA48 usually begins from the 4th decade, often with the cerebellar cognitive affective syndrome (a dysexecutive and behavioural picture that progresses to cognitive decline and psychiatric disorder), followed by gait ataxia and dysarthria, with frequent parkinsonism, chorea or dystonia [12, 56]. Cerebellar atrophy is vermian and hemispheric; in presymptomatic carriers there is already focal atrophy of the vermis, paravermis and medial portions of lobules VI and VII before ataxia [56]. In the published series, onset ranged from 22 to 74 years: mean 39.9 years in two families from southern Italy (with tonic-clonic seizures in childhood in two patients and epilepsy in three), 49 years in five French families and 65.3 years in a Dutch family whose nine affected members died after a mean duration of 13.2 years [74]. About half of the patients in the Catalan family began with cognitive or affective symptoms (anxiety, agoraphobia, executive decline) and the other half with ataxia; two had only cognitive symptoms and the patients were anosognosic [74]. One Turkish woman began at 51 years with a frontal dysexecutive syndrome and another presented at 48 years as behavioural variant frontotemporal dementia, which broadens the differential to frontotemporal dementias and Huntington disease [74]. MRI shows posterior cerebellar atrophy and, in almost all patients of a series of ten, bilateral T2 hyperintensity of the dentate nuclei («crab sign»); positron emission tomography showed cerebellar, striatal and cortical hypometabolism, and DaTSCAN presynaptic dopaminergic reduction in some [74]. At autopsy in the Dutch family there were p62-positive intranuclear inclusions in the cerebellum, substantia nigra, hippocampus, pontine nuclei and cortex [74]. Penetrance is incomplete or age-dependent (asymptomatic carriers at 75 and 64 years; a carrier father asymptomatic at 65 years), and in 23 families transmission was mainly maternal, suggesting sex-dependent penetrance [74].

SCAR16 includes progressive ataxia from childhood to adulthood, dysarthria, nystagmus, lower limb spasticity, mild sensory neuropathy and, classically, hypogonadotropic hypogonadism, with frequent cognitive impairment [12]. In the original Chinese family, four siblings began between 14 and 19 years, with normal hormones; two Chinese sisters with onset in late adolescence had amenorrhoea and poor sexual development with low FSH, LH, oestradiol and progesterone, compatible with Gordon Holmes syndrome [73]. In a German series of 167 patients with recessive ataxia, 3 (1.8%) had biallelic STUB1 variants, with onset between 2 and 48 years, pyramidal involvement in all and no gonadal dysfunction; in a consanguineous Arab family onset was between birth and 2 years, with cerebellar hypoplasia, a thin pons and corpus callosum and a prematurely aged appearance in young adulthood [73]. The Stub1-null mouse has severe cerebellar dysfunction, Purkinje cell loss and gonadal dysfunction [73].

3.13. SCA19/22 (KCND3)

Orphanet describes mild ataxia with predominantly executive cognitive impairment, myoclonus and postural tremor; the discovery articles add pyramidal signs, tremor and peripheral neuropathy in some patients, mild atrophy of the hemispheres and vermis and Purkinje degeneration at autopsy [39, 84]. Later reports describe paroxysmal exacerbations of ataxia and a phenotype with parkinsonism, cognitive dysfunction and brain iron accumulation [39]. The original Dutch family had mild ataxia with impaired performance on the Wisconsin Card Sorting Test, myoclonus and irregular low-frequency postural tremor, and autopsy showed Purkinje cell loss mainly in the anterior vermis; the Chinese family that defined SCA22 had pure ataxia with onset between 13 and 46 years, with apparent anticipation across generations (means of 40.5, 20.7 and 12.5 years) [65]. In a French family, onset ranged from 24 to 51 years, with urinary urgency or incontinence in five and a single patient in a wheelchair after 43 years of disease [65]. In 16 patients from two other French families with the same deletion, mean onset of ataxia was at 23.1 years (2 to 66), 8 had mild parkinsonism, 5 epilepsy with median onset at 5.3 years (3 to 12), 12 of 15 assessed had mild cognitive impairment and most had had learning difficulties [65]. De novo variants produce more severe pictures: a Belgian boy with an in-frame 9 bp duplication had motor delay from 3 years, valproate-responsive epilepsy, a full-scale IQ of 54 and severe ataxia at 10 years with normal MRI, and a Japanese man with p.(Gly384Ser) had developmental delay, an IQ of 59 and truncal myoclonus from 18 years [65]. In Taiwan, the variants p.(Cys317Tyr) (de novo) and p.(Pro375Ser) showed loss of function with a dominant-negative effect on the normal channel [65].

3.14. SCA5 and the SPTBN2 spectrum

SCA5 is a slowly progressive, pure adult cerebellar ataxia with onset between 10 and 50 years, incoordination, dysarthria and oculomotor abnormalities [33, 34]. SCAR14 manifests as psychomotor delay, severe early-onset gait ataxia, oculomotor abnormalities, cerebellar atrophy and intellectual disability, and heterozygous de novo missense variants as non-progressive congenital ataxia with delay, sometimes labelled ataxic cerebral palsy at 6 months [33].

3.15. SCA21 (TMEM240)

In a systematic review of 42 patients plus 5 new ones, ataxia began relatively early (median 18 years; 1 to 61), with hyporeflexia in 69%, bradykinesia in 65%, slow saccades in 38% and pyramidal signs in 17%, and with hypokinetic and hyperkinetic disorders (polyminimyoclonus, proximal myoclonus); cognitive impairment, severe in the original family, is not obligatory, and one childhood-onset case improved markedly [40, 86]. There are sporadic presentations that mimic ataxic cerebral palsy; the absence of perinatal events, progression or cerebellar atrophy should prompt genetic testing [40].

3.16. SCA27A (FGF14)

In a review of 33 cases, early-onset tremor was present in 95.8% (mean onset at 12.1 years) and gait ataxia began later (mean 23.7 years), with limb ataxia, dysarthria and nystagmus; some had episodic ataxia, often triggered by fever, orofacial dyskinesia, psychiatric disorders, cognitive impairment or parkinsonism [2, 43]. Cerebellar atrophy is less frequent than in SCA27B [2]. In the three-generation Dutch family in which p.(Phe145Ser) (historical notation) was described, all 14 affected individuals had hand tremor from childhood, worsened by stress and exertion, with mild ataxia from 15–20 years; 6 did not complete primary school, 5 had outbursts of aggression, 3 depression, 8 subtle orofacial dyskinesias and 6 head tremor, and only the three oldest had severe ataxia [67]. An English family followed for decades, with 17 affected individuals and a 161 kb deletion involving FGF14 and ITGBL1, mainly had an oculomotor disorder from the first 2 years (poor smooth pursuit, gaze-evoked and upbeat nystagmus, strabismus), non-progressive and with mild ataxia in a few [67]. In French children, presentation was with acute episodic ataxia triggered by fever or stress between 2 and 8 years, or with early tremor, with subsequent permanent symptoms; in a French-Canadian family, episodes of incoordination, vertical oscillopsia, dysarthria and headache preceded a progressive ataxia [67]. In a Swedish family of nine affected individuals with a 599 kb deletion, all needed educational support, outbursts of anger and attention deficit hyperactivity disorder were frequent, one woman had psychosis and PET showed prefrontal, temporal and cerebellar hypometabolism [67]. Onset may be late: a Japanese man with a nonsense variant began at 47 years [67]. Haploinsufficiency is the proposed mechanism, supported by a translocation disrupting isoform 1b and by the Fgf14-deficient mouse, which has ataxia and a paroxysmal hyperkinetic disorder [67].

3.17. SCA23 (PDYN)

Onset is usually in adulthood (mean of about 43 ± 15 years; 10 to 73), with slowly progressive gait and limb ataxia and dysarthria, variable peripheral neuropathy, dysphagia, tremor and cerebellar atrophy; it was defined in four Dutch families, with intrafamilial variability [41, 83]. A Brazilian report from 2026, cited in the gene curation, describes a woman with onset at 19 years, no affected relatives and a negative first-line panel before a PDYN variant was identified [41].

3.18. SCA49 and SAMD9L syndrome

SAMD9L-related syndrome forms a continuum that includes ataxia-pancytopenia syndrome (OMIM 159550), predisposition to monosomy 7, SAMD9L-associated autoinflammatory disease (SAAD) and SCA49 [57, 111]. Cytopenias of all lineages with a hypocellular marrow usually appear in childhood, with a risk of bone marrow failure, myelodysplasia and myeloid leukaemia, often with monosomy 7, and with B-lymphocyte and natural killer (NK) cell immunodeficiency; between 5% and 12% of childhood myelodysplasia is attributable to germline SAMD9L variants [57]. Neurological manifestations appear later, in adolescence or adulthood: slowly progressive cerebellar ataxia, sensorimotor axonal polyneuropathy, dysautonomia and paracentral retinal dysfunction, with white matter changes and cysts that may precede symptoms [111].

3.19. Rare or provisional entities

3.20. Summary by system

4. Clinical Recognition and Red Flags

4.1. When to suspect

An SCA due to a conventional variant should be considered in four situations. The first is slowly progressive adult cerebellar ataxia with a dominant family history in which testing for the common expansions was negative; laboratories generally test the most frequent expansions first, and conventional variants explain part of the remaining families [13, 76]. The second is sporadic adult-onset ataxia, because in the SPORTAX cohort 6.4% of 377 patients with sporadic ataxia had a pathogenic or likely pathogenic single-nucleotide variant, including in CACNA1G, KCNC3, STUB1, TRPC3 and TMEM240, and six of these patients had been classified as clinically probable multiple system atrophy of cerebellar type [115]. The third is the child with congenital ataxia, developmental delay or the label of ataxic cerebral palsy, especially if there are no convincing perinatal events, if symptoms progress or if there is cerebellar atrophy (SCA13, SCA29, SCA42ND, SCA21, SPTBN2) [33, 40, 117]. The fourth is the presence of a characteristic extracerebellar feature (following table).

The history should include a three-generation pedigree, consanguinity, geographical origin, onset, stage of disability and non-neurological signs [76]. Anticipation points to a repeat disorder; absence of male-to-male transmission suggests X-linked inheritance; affected siblings or consanguinity suggest recessive inheritance; a grandfather with late-onset ataxia and a grandson with intellectual disability suggest fragile X-associated tremor/ataxia syndrome (FXTAS) [1]. An apparently negative family history may result from non-recognition of the disease, early death of the parent, late onset or reduced penetrance (SCA14, SCA47) [1, 36, 54]. Multifactorial and acquired cerebellar disorders are four to five times more frequent than hereditary ones and may confound the family history [1].

4.2. Clinical clues to the gene

4.3. Warning and emergency signs

The Ataxia UK recommendations classify as a red flag progression over weeks or months, which suggests a paraneoplastic or prion cause or multiple system atrophy and requires urgent investigation; a child with ataxia should be referred urgently to paediatrics, and any progressive ataxia should be referred to a neurologist and, ideally, to a specialist ataxia centre [27]. No SCA due to a conventional variant has an acute course, so a subacute ataxia should not be attributed to a sequence variant without excluding acquired causes [26, 27]. In SCA49, a new cytopenia requires immediate haematological assessment because of the risk of myelodysplasia [57]. In SCA28 and SCA38, progressive dysphagia signals a risk of aspiration [11, 47]. In SCA27A and in the episodic form associated with CACNA1G, fever and other triggers may precipitate episodes [43, 91].

4.4. Newborn screening

There is no newborn screening for any of the entities in the class in the sources consulted.

5. Diagnosis: Criteria, Tests and Algorithms

5.1. Criteria

There are no formal consensus clinical criteria for any SCA in this class; the diagnosis is established in a proband with a suggestive picture by identifying a heterozygous pathogenic or likely pathogenic variant in the corresponding gene (or biallelic variants in the allelic recessive forms), and a VUS neither confirms nor excludes the diagnosis [11, 24, 36, 47]. The diagnosis of SCA20 requires the 11q12.2–q12.3 duplication or, in its absence, the combination of a suggestive picture, marked dentate calcification on CT and mild to moderate pancerebellar atrophy on MRI, with normal calcium metabolism [8].

5.2. Clinical assessment and investigations before genetic testing

The North American recommendations of the Clinical Research Consortium for the Study of Cerebellar Ataxia and the National Ataxia Foundation (CRC-SCA/NAF, 2025) result from a systematic review (7362 articles identified, 68 included) and contain 20 recommendations [26]. All symptomatic patients should have a complete neurological assessment before testing (recommendation 1); without a family history, non-genetic causes are first excluded by imaging, serology and cerebrospinal fluid as indicated, unless there is high clinical suspicion of a specific syndrome (recommendation 2); with a known family history and a compatible picture, genetic testing is the first investigation (recommendation 3); and testing should be expedited when an ataxia with an approved treatment is suspected, which in 2025 existed only for Friedreich ataxia (recommendation 4) [26].

The Ataxia UK guidelines (3rd edition, 2019, endorsed by ERN-RND) propose a stepwise investigation in adults [27]. In primary care: urea and electrolytes, creatinine, full blood count, erythrocyte sedimentation rate or C-reactive protein, liver enzymes, gamma-glutamyl transferase, thyroid function, vitamin B12, folate, glucose and chest radiography. First line in secondary care: alpha-fetoprotein, blood film, caeruloplasmin and copper, coeliac disease serology, creatine kinase, genetic tests for Friedreich ataxia, SCA1, 2, 3, 6 and 7 (and 12 and 17) and FXTAS, lactate, lipid-adjusted vitamin E, lumbar puncture, MRI of the brain and cervical spine, paraneoplastic, anti-GAD and anti-voltage-gated calcium channel antibodies, CT of the chest, abdomen and pelvis and 14-3-3 protein. Second line: cholestanol, oxysterols, bile acids, coenzyme Q10, electroencephalogram, very-long-chain fatty acids, muscle biopsy, ophthalmological assessment with optical coherence tomography (OCT), nerve conduction studies, phytanic acid, leucocyte enzymes, whole-body PET and the remaining genetic tests by next-generation sequencing (NGS) [27]. The SARA and ICARS (International Cooperative Ataxia Rating Scale) scales record severity and progression [27].

In the assessment after diagnosis, GeneReviews includes neurological examination with SARA, brain MRI, nerve conduction studies and electromyography when neuropathy is suspected, assessment of swallowing and nutritional status, pulmonary function tests and a sleep study when there are symptoms, and cognitive assessment with the CCAS scale [1]. For specific entities: CT for dentate calcification and calcium biochemistry in SCA20 [8]; ophthalmological assessment with OCT and acuity in SCA28 and OPA12 [11]; dermatological and retinal assessment in SCA34 [44]; smell and audiology in SCA38 [47]; full blood count, bone marrow and immunology in SCA49 [57]; electroencephalogram in SPAX5 and in the neurodevelopmental forms of CACNA1G [9, 11].

5.3. Choice of genetic test: divergence between guidelines

The two main guidelines diverge on the order of testing. Ataxia UK places the common expansions first line and NGS second line [27]. CRC-SCA/NAF recommends that the initial test be as comprehensive as possible and reliably detect repeat expansions; ideally, short-read genome sequencing with tandem repeat sizing and reflex to long-read sequencing or repeat-primed polymerase chain reaction (RP-PCR); if this is not feasible, an expansion panel with reflex to genome sequencing; single-gene testing is reserved for high clinical suspicion (recommendation 18, partly level 1) [26]. The phenotype should be sent to the laboratory, a negative exome or genome sequencing result should be reanalysed after 2 years or more or when symptoms change (recommendation 19), and DNA banking should be offered to symptomatic individuals with a negative test or who decline testing (recommendation 20) [26]. The divergence mainly reflects the date of the recommendations and access to technologies; in both, investigation of conventional variants presupposes that expansions have been or are being studied [26, 27].

For the disorders in this class, GeneReviews recommends a multigene ataxia panel that includes deletion/duplication analysis or a genomic test (exome or genome), and advises against single-gene testing, except for a known familial variant [1, 11, 36]. Panels vary in genes and sensitivity; some do not include rare genes such as TTBK2 [35]. Trio testing (proband and parents) is the method of the paediatric series and allows de novo variants to be recognised [15, 28].

5.4. Test limitations by entity

A negative panel or exome result does not exclude an SCA: repeat expansions, deep intronic variants, variants in poorly covered regions, low-level mosaicism and structural variants may be missed, and the proportion of variants detectable by a method is not equivalent to the clinical sensitivity of the test, which has not been estimated for SCA42 in the sources consulted [1, 10, 11]. The CACNA1G variants described are all single-nucleotide variants or small indels detectable by sequencing [10].

5.5. Diagnostic yield in cohorts

These values measure proportions in cohorts selected by referral centres and are not population prevalences. Yield is higher when there is a positive family history and early onset [119] and lower in populations in which expansions predominate [7].

5.6. Variant interpretation and report content

The report should state the gene, the reference transcript with version (for CACNA1G, the MANE NM_018896.5), the HGVS nomenclature at DNA and protein level, zygosity, the classification according to the American College of Medical Genetics and Genomics and Association for Molecular Pathology (ACMG/AMP) criteria and the correspondence with historical nomenclature when it differs [10, 26]. There are several interpretation points specific to the class. In CACNA1G, the recurrent variant p.(Arg1715His) is classified as pathogenic or likely pathogenic by several submitters without conflict; outside the recurrent variants, the domain (voltage sensor, intracellular gate), de novo origin or segregation and electrophysiological studies weigh in the classification, and calibrated functional frameworks are still at the preprint stage [6, 96, 99]. In TTBK2, only truncating frameshifts in the clustered exon are established [95]; in TGM6, the variants have a frequency incompatible with a rare dominant disease [46]; in ELOVL5, NM_001242828.1:c.304C>T is benign [47]; in PRKCG, c.285C>G p.(Asp95=) is benign [36]; and in KCND3, the classification of an ataxia variant does not imply a risk of Brugada syndrome [39].

Segregation studies in affected and unaffected relatives and parental testing in apparently sporadic cases are the most accessible tools for reclassifying a VUS [1, 13]. In the Japanese series, all novel variants in conventional genes remained VUS under ACMG criteria [7]. Dual diagnosis (1.3% in the SPORTAX cohort) and the co-occurrence of variants in several ataxia genes should be taken into account [88, 115].

5.7. Algorithm

5.8. Prenatal and preimplantation diagnosis

When the familial pathogenic variant has been identified, prenatal diagnosis (chorionic villus sampling or amniocentesis) and preimplantation genetic testing (PGT) are possible; CRC-SCA/NAF recommends that anyone wishing to test a pregnancy be referred for genetic counselling before or early in pregnancy [26, 36]. A VUS should not be used for prenatal decisions, and GeneReviews recommends extreme caution with variants without confirmed pathogenicity [36]. Requests for prenatal diagnosis of adult-onset disorders are uncommon and require careful counselling [47].

6. Differential Diagnosis

6.1. Principles

The differential diagnosis of an SCA due to a conventional variant has three layers: the other dominant ataxias, especially those caused by expansions, which are more frequent and are not detected by sequencing; recessive, X-linked and mitochondrial ataxias that can mimic vertical transmission or appear as isolated cases; and acquired and non-genetic degenerative ataxias, four to five times more frequent than hereditary ones [1]. Within the class, clinical distinction between SCA5, 11, 15, 19/22, 21, 23, 26 and 27A may be impossible, and pure cerebellar ataxia on MRI is shared by many SCA [21, 35]. The polyglutamine expansion SCA (SCA1, 2, 3, 7, 17, DRPLA) typically begin before 40 years, progress more rapidly and involve the brainstem on MRI [11, 47].

6.2. Comparative table

6.3. Main differentials in prose

Expansions and SCA6. A normal sequencing panel or exome does not exclude SCA6 or other expansions [1, 60]. Morino and colleagues noted the clinical similarity between SCA42 (pure ataxia, adult onset, slow progression) and SCA6 [25]. Downbeat nystagmus is more typical of SCA6 and episodic ataxia type 2, whereas half of the British patients with SCA11 and vertical nystagmus had upbeat nystagmus [35]. The SCA due to non-coding repeats (SCA8, SCA10, SCA12, SCA27B, SCA31, SCA36, SCA37 and SCA4) are included here only as differentials; they require the same repeat detection methods as the polyglutamine disorders [1, 2].

SCA27B and SCA27A. The FGF14 GAA expansion is the most frequent genetic cause in some late-onset ataxia cohorts: 61% of French Canadians, 18% of Germans, 15% of Australians and 10% of Indians with late-onset cerebellar ataxia, 48% of patients with idiopathic downbeat nystagmus, and 11.9% of the SPORTAX cohort [43, 115]. SCA27A, due to conventional variants in the same gene, is distinguished by onset with tremor in adolescence and by the diagnostic method [2, 43]. Penetrance thresholds vary between studies: 250 to 300 repeats with incomplete penetrance and more than 300 with complete penetrance in the series of 128 patients, versus complete penetrance from 335 repeats in an Australian and German series; maternal transmission tends to expand the allele and paternal transmission to contract it, with less paternal inheritance observed (30%) [55]. In the series of 128 patients, ataxia was episodic at onset in 46%, mean onset of episodes was at 55 years and of progressive ataxia at 59 years, 42% had downbeat nystagmus, 58% eventually needed a walking aid and 13% a wheelchair [55]. Long-read genome sequencing was the most reliable method in a 2025 series, and it was this series that reassigned the French-Canadian SCA47 family to SCA27B [55].

SCA28 and SPG7. Paraplegin (SPG7) forms the m-AAA complex with AFG3L2, and adult SPG7 (about 5% of recessive hereditary spastic paraplegias) shares ophthalmoparesis, ptosis, parkinsonism, cognitive decline and optic atrophy; the distinction can only be made by molecular testing [11].

Sporadic ataxia and MSA-C. In the SPORTAX cohort, all six patients with a clinical diagnosis of probable MSA-C and a single-nucleotide variant had genitourinary dysfunction but not orthostatic hypotension, and the variants found included KCNC3; a patient with a TRPC3 variant was initially suspected of having MSA-C, and the hot cross bun sign was described in a Chinese family with SCA42 and in 32.4% of SCA34 cases [44, 49, 100, 115]. These findings indicate that the hot cross bun sign and autonomic dysfunction do not exclude a genetic SCA.

SCA20 and calcifications. Dentate calcification occurs in 0.7% of the population over 65 years, rarely without basal ganglia calcification; hyperparathyroidism and pseudohypoparathyroidism should be excluded biochemically; palatal tremor without calcification occurs in sporadic progressive ataxia with palatal tremor, rarely in MSA-C, in adult Alexander disease and in dominant branchial myoclonus with ataxia [8].

Paediatric forms. In children, congenital cerebellar hypoplasia (SCA13, SCA29, SCA42ND, SCAR14) must be distinguished from cerebellar malformations and metabolic disorders; a child with cerebellar atrophy needs metabolic and structural investigation and not just a panel [9, 117]. The variant p.(Ala961Thr) was found in children with a previous clinical diagnosis of Rett syndrome, and p.(Met1531Val) in a child with West syndrome [113].

Episodic phenotypes. CACNA1G-associated episodic vestibulocerebellar ataxia was diagnosed after exclusion of KCNA1, CACNA1A, hemiplegic migraine genes and SCA6 [91]. Episodic ataxia type 2 and CACNA1A forms with fixed interictal ataxia may be indistinguishable from SCA6 [60].

7. Treatment and Risk-Management Strategies

7.1. Principles

No SCA due to a conventional variant has a disease-modifying treatment approved by any regulatory authority; in 2025, the only therapies approved by the US Food and Drug Administration (FDA) for any ataxia were for Friedreich ataxia [9, 26]. Omaveloxolone, approved in the USA and Europe for Friedreich ataxia from 16 years of age, has no data or rationale for extrapolation to this class, and vitamin E and coenzyme Q10 apply only to recessive ataxias with deficiency of these compounds [1, 9]. Management is multidisciplinary and symptomatic, with rehabilitation, treatment of associated manifestations, prevention of complications and counselling [1, 27]. The only exception with a genotype-specific randomised trial is DHA supplementation in SCA38 [29]. All therapeutic evidence on SCA42 comes from reports of one or two patients and from cellular studies [30, 31, 91, 112]. Most symptomatic recommendations in the Ataxia UK guidelines are good practice points (GPP), based on experience, with few randomised trials [27].

7.2. Rehabilitation and non-pharmacological measures

Physiotherapy (balance exercises, gait training, strengthening) and occupational therapy maintain mobility and independence in activities of daily living; inpatient rehabilitation with physiotherapy and occupational therapy may improve ataxia and function in degenerative ataxias [1, 120, 121]. Neither exercise nor physiotherapy halts progression, but patients should remain active [8, 11]. Adaptive devices (sticks, walking frames, powered wheelchairs, weighted cutlery, dressing hooks), home adaptations (grab rails, raised toilets, ramps) and weight control are part of the plan [1]. Speech and language therapy treats dysarthria and guides augmentative and alternative communication, which does not impair speech [1, 24]. Dysphagia is managed with modified food consistency, defined by videofluoroscopy or fibreoptic endoscopic evaluation of swallowing (FEES), anti-aspiration strategies and gastrostomy when the risk is high or caloric intake insufficient [11, 24, 27]. Transcranial magnetic stimulation over the cerebellum, tested in spinocerebellar degenerations of various causes, may improve cerebellar motor signs after 21 daily sessions [1, 120]. In SCA38, a study of cerebellar transcranial direct current stimulation with 7 participants was registered and completed in 2023 (NCT05951010), with no results read; the review by Coarelli and colleagues describes contradictory results of this technique in phase 2/3 studies [88].

Fall prevention includes walking aids, adaptations and a personal alarm; in SCA15 and SCA20 GeneReviews recommends bone densitometry and treatment of osteoporosis to limit fractures [8, 21]. For cognition, cognitive rehabilitation is grade C in the Ataxia UK guidelines and cognitive behavioural therapy, including Goal Management Training®, is mentioned by GeneReviews [1, 27]. In congenital SCA13, early intervention, special education and an individual education plan apply, together with physiotherapy to minimise contractures, scoliosis and hip dislocation [24].

7.3. Symptomatic treatment by manifestation

The Ataxia UK guidelines propose an order of drugs by symptom, applicable to the class [27]. The doses below are those of the labels for the approved indications, because the ataxia guidelines do not give doses; all use for ataxia symptoms is off-label, except where the label indication coincides (spasticity).

Cerebellar tremor responds poorly to antitremor drugs, as the GeneReviews entries on SCA13, SCA14 and SCA20 emphasise [8, 24, 36]. Clonazepam and other benzodiazepines may worsen ataxia and coordination, so the benefit on spasticity, tremor or myoclonus must be weighed against this risk [11]. Botulinum toxin is indicated in focal spasticity and dystonia, with no specific dose in the ataxia sources [11, 27]. The Portuguese SmPCs differ in some maxima: propranolol in tremor is limited to 160 mg/day, primidone to 1.5 g/day and gabapentin reaches 3600 mg/day in peripheral neuropathic pain; baclofen allows 100–120 mg/day only under hospital monitoring, and the clonazepam SmPC asks for special care in patients with ataxia (table in 7.12) [128–132].

7.4. Riluzole

Riluzole (Rilutek® and generics) inhibits glutamate release and is approved for amyotrophic lateral sclerosis (ALS); in ataxias, use is off-label [133, 134]. GeneReviews mentions the possibility of considering it at a dose of 100 mg/day, as the only drug shown to improve ataxia in patients with ataxias of mixed causes, with monitoring of liver enzymes [1]. The 2014 EFNS/ENS consensus gave it a level B recommendation, and the 2018 American Academy of Neurology (AAN) systematic review considered the benefit probable; this review is flagged as retracted in PubMed [120, 121].

In the table, OR denotes the odds ratio and CI the confidence interval.

These trials have limitations. The samples are small and the confidence intervals wide. The populations are mixed, and the abstract of the trial by Romano and colleagues does not reveal whether it included patients with SCA due to conventional variants, so the benefit cannot be attributed to this class. The only single-genotype trial (ATRIL, SCA2) was negative. Response outcomes defined by thresholds measure short-term symptomatic effect and not disease modification, and results cannot be compared across trials with different scales and populations [135–137]. The SAMD9L GeneReviews entry suggests considering a therapeutic trial of riluzole for the ataxia of SCA49, also off-label [111].

Dosage and safety (label). The recommended dose in ALS is 50 mg orally twice daily, at least 1 hour before or 2 hours after meals, matching the dose in the 12-month trials [133, 136]. Aminotransferases (alanine aminotransferase, ALT; aspartate aminotransferase, AST) should be measured before and during treatment; the drug is not recommended with baseline ALT or AST above 5 times the upper limit of normal (ULN) and should be discontinued if there is hepatic dysfunction; in the ALS trials, about 50% had ALT above the ULN, 8% above 3 times and 2% above 5 times, with peaks in the first 3 months [133]. The label also mentions neutropenia (request a full blood count in the presence of fever) and interstitial lung disease (discontinue) [133]. In the trial by Romano and colleagues there were no serious adverse events, and two patients had liver enzyme rises below 2 times the ULN [136].

7.5. Troriluzole

Troriluzole (BHV-4157) is a prodrug of riluzole that acts on glutamate [138]. According to the annual report (Form 10-K) of the sponsor, Biohaven, filed with the US Securities and Exchange Commission, trial BHV4157-201 (phase 2b/3, 8 weeks, 140 mg/day) did not separate from placebo on SARA; trial BHV4157-206 (phase 3, 48 weeks, 200 mg/day, outcome on the f-SARA, a modified functional scale) was stratified by genotype into SCA1/2, SCA3 and SCA6/7/8/10, that is, expansion SCA only; the real-world evidence study BHV4157-206-RWE (3 years, comparison with natural history controls from the CRC-SCA consortium) met its endpoint [138]. The same document reports submission of the new drug application (NDA) in the 4th quarter of 2024, its acceptance by the FDA in February 2025 with priority review, receipt of a Complete Response Letter (CRL) in November 2025 and a request for a type A meeting; the quarterly report for the 2nd quarter of 2026 does not mention approval and refers to a class action concerning statements about troriluzole in SCA [138]. These dates come from the sponsor and not from an official FDA source consulted. There is an expanded access programme (NCT06034886) [138]. No patient with SCA due to a conventional variant was included in the main trials, so even if the drug is approved, extrapolation to this class has no basis in data [138].

7.6. Docosahexaenoic acid in SCA38

DHA is the only treatment tested in a randomised trial specific to an SCA due to a conventional variant [29]. The rationale is the reduced serum DHA in patients, due to deficiency of the elongase ELOVL5, which is highly expressed in Purkinje cells [29, 106].

The trial is very small, with a single variant and five patients per arm in the controlled phase; the open-label phase has no comparator, and there are no data in asymptomatic carriers [29, 47]. Serum DHA levels did not differ before and after 40 weeks [106]. A two-year follow-up was published in Spain, but only the title was consulted [140]. DHA is supplied as a food supplement and has no FDA or European Medicines Agency (EMA) authorisation for this indication; the reference dose is that of the trial, 600 mg/day orally in symptomatic adults over 18 years, with no maximum dose defined in the source [29, 105].

7.7. SCA42 and the CACNA1G spectrum: drugs used in reports

Because Cav3.1 is a T-type channel, several authors have tested drugs that act on these channels [30, 112]. The evidence is limited to case reports and one cellular study, and there is no registered clinical trial for SCA42 or CACNA1G [9]. All these uses are off-label.

In the study by Hara and colleagues, in HEK293T cells, zonisamide at concentrations of 10 to 50 µM, equivalent to the patient's plasma levels, shifted the activation curve of the mutant channel towards more negative potentials in a concentration-dependent manner, bringing it closer to that of the wild type, with no effect on the normal channel or on current density; efonidipine (5 µM) had no effect [30]. The dose of 25 mg/day is lower than antiepileptic doses; according to the authors, in Japan zonisamide is also approved for Parkinson disease [30]. The opposite results in the Korean patients, with doses up to 200–300 mg/day, indicate that the tremor response in one patient does not generalise to ataxia [112]. Acetazolamide and topiramate, first-line drugs in other episodic ataxias, worsened the only patient with the episodic form associated with CACNA1G [27, 91].

Zonisamide: comparison of labels. In the USA (Zonegran® and generics), zonisamide is indicated as adjunctive therapy for partial seizures in adults; it is started at 100 mg/day, may be increased to 200 mg/day after two weeks and then to 300 and 400 mg/day at intervals of at least two weeks; efficacy was shown between 100 and 600 mg/day, adverse effects are more frequent from 300 mg/day and safety below 16 years has not been established [142]. In the European Union (Zonegran®, authorised on 10 March 2005), the indication includes monotherapy of partial seizures in adults with newly diagnosed epilepsy and adjunctive therapy in adults, adolescents and children from 6 years; in adult monotherapy, 100 mg/day is used in weeks 1–2, 200 mg/day in weeks 3–4 and 300 mg/day in weeks 5–6 (maintenance 300 mg/day, maximum 500 mg/day); in adjunctive therapy with CYP3A4 inducers, treatment starts at 50 mg/day in two doses, and in children from 6 years at 1 mg/kg/day up to 6–8 mg/kg/day (maximum 500 mg/day); withdrawal should be gradual (reduction of 100 mg per week in adults) [143]. The drug is contraindicated in sulfonamide hypersensitivity, may cause metabolic acidosis and renal stones, has a long half-life (steady state up to two weeks) and was teratogenic in animals [142, 143].

Other comparison labels. Ethosuximide (Zarontin®, USA) is indicated in absence epilepsy: 500 mg/day from 6 years, with increases of 250 mg every 4 to 7 days; doses above 1.5 g/day require strict supervision [144]. Clonazepam (USA) is started in adults at up to 1.5 mg/day in three doses, with increases of 0.5–1 mg every 3 days up to a maximum of 20 mg/day, and in children up to 10 years or 30 kg at 0.01–0.03 mg/kg/day [145]. Carbamazepine is indicated in epilepsy and trigeminal neuralgia, with a low starting dose, gradual increase and level monitoring [146].

7.8. Myoclonus and epilepsy: SCA14, SPAX5 and neurodevelopmental forms

In SCA14, GeneReviews states that clonazepam or valproic acid may be effective for axial myoclonus, based on a Japanese family, without giving a dose [36]. In SPAX5, GeneReviews refers to the guidance for progressive myoclonic epilepsies (Cameron 2023): levetiracetam and valproate as first line for seizures and myoclonus, avoiding valproate in women of childbearing potential because of teratogenicity; as second line or for intolerance, topiramate, zonisamide, lamotrigine, brivaracetam, piracetam or perampanel; benzodiazepines (clonazepam, clobazam) as adjuncts for myoclonus; carbamazepine and phenytoin may worsen myoclonus and should be avoided [11]. In SCA42ND and the CACNA1G encephalopathies, the usual antiepileptic drugs are used, with developmental and educational support [9].

The reference doses are those of the labels for the approved indications, because the sources for the class do not give doses. Levetiracetam (Keppra XR®, extended release, USA) is started at 1000 mg once daily, with increases of 1000 mg every 2 weeks up to a maximum of 3000 mg/day, from 12 years and 50 kg, with adjustment to renal function [147]. Valproic acid (USA) in complex partial seizures, from 10 years, is started at 10–15 mg/kg/day and increased by 5–10 mg/kg per week until response; the optimal response is usually obtained below 60 mg/kg/day, the usual therapeutic range is 50–100 µg/mL and the likelihood of thrombocytopenia increases with levels above 110 µg/mL in women and 135 µg/mL in men [148]. Clonazepam follows the regimen described in 7.7 [145].

7.9. Episodic ataxias and the boundary with CACNA1A and FGF14

In episodic ataxias, the Ataxia UK guidelines recommend avoiding triggers (stress, caffeine, alcohol, excessive exertion; GPP), acetazolamide as first line in episodic ataxia types 1 and 2 (GPP) with hydration and annual renal ultrasound (D), 4-aminopyridine as second line in type 2 on a named-patient basis (C) and carbamazepine, phenytoin or lamotrigine as second line in type 1 (GPP) [27]. On the US label, acetazolamide has indications for oedema, centrencephalic epilepsies, glaucoma and acute mountain sickness, with no indication for ataxia; in epilepsy 8–30 mg/kg/day is used in divided doses [60, 149].

Fampridine (prolonged-release 4-aminopyridine; Fampyra® in the EU, dalfampridine Ampyra® in the USA) is approved only to improve walking in adult multiple sclerosis: in the EU since 20 July 2011, with an expanded disability status scale (EDSS) score of 4 to 7, and in the USA with a maximum recommended dose of 10 mg twice daily, about 12 hours apart, prior determination of creatinine clearance and a risk of seizures that increases with dose and renal impairment [150, 151]. In SCA27B (expansion), 86% of 50 patients reported a response to off-label 4-aminopyridine, and a randomised phase 3 trial of fampridine 10 mg versus placebo in 70 patients is under way in France (NCT07185347, started October 2025, recruiting) [43]. For SCA27A (conventional FGF14 variants) there are no data; extrapolation would be purely mechanistic [43]. A study of dalfampridine on gait in 20 patients with SCA (NCT01811706, completed in 2013) appears in the trial registry, with no results read.

7.10. SCA49: haematological and immunological management

In SCA49, the haematological component determines management. GeneReviews recommends follow-up by a haematologist and a full blood count with differential every 6–12 months, immediate assessment in the presence of signs of cytopenia, consideration of annual bone marrow examination and, in autoinflammatory disease, consideration of corticosteroids or Janus kinase (JAK) inhibitors [57, 111]. Non-steroidal anti-inflammatory drugs, anticoagulants and thrombolytics are contraindicated if there is severe thrombocytopenia and require caution in other situations [57]. The sources consulted do not detail the indications for haematopoietic cell transplantation.

7.11. Investigational therapies and recent developments

The developments published between November 2025 and October 2026 are: de novo SCA42 with neurodevelopmental and psychiatric onset (Saricicek 2026), the de novo variant p.(Arg1553Gln) with cerebellar hypoplasia (Selek 2026), the 2.9% yield of conventional genes in Japan, all VUS (Watanabe 2026), the preprint of the functional framework for CACNA1G and the troriluzole CRL announced by the sponsor [7, 14, 99, 114, 138]. In 2025 the electrophysiological classification of 18 CACNA1G variants, the series of 19 congenital patients, the SPORTAX cohort and the CRC-SCA/NAF recommendations were published [26, 89, 96, 115]. In the ClinicalTrials.gov search of 28 October 2026 there were no registered trials for SCA42 or CACNA1G; for SCA14, SCA28, SCA13, SCA48, SCA15 and SCA5 only a generic rare disease registry appeared (NCT01793168). A search without results does not demonstrate the absence of studies.

7.12. Regulatory status

In Portugal, the summaries of product characteristics (SmPC) approved by INFARMED for the symptomatic drugs used in the class confirm that none of them has an indication for cerebellar ataxia; the doses below serve for comparison with those of the US labels cited in 7.3, 7.7 and 7.8, which remain the reference in this article in the absence of doses in ataxia guidelines [27].

The package inserts of the Brazilian Health Regulatory Agency (ANVISA) for these drugs were not consulted, so the situation in Brazil could not be compared.

7.13. Agents to avoid, pregnancy, lactation, perioperative care and transition

Alcohol and sedative drugs, including benzodiazepines, may worsen gait and coordination in all SCA in the class; in SCA15 there is abnormal sensitivity to the motor effects of alcohol, and in SCA20 GeneReviews advises avoiding neurotoxic drugs such as isoniazid [8, 11, 21, 36]. In SPAX5, carbamazepine and phenytoin may worsen myoclonus [11]. In SCA49, non-steroidal anti-inflammatory drugs, anticoagulants and thrombolytics are avoided in thrombocytopenia [57]. In the Korean patient with SCA42, zonisamide worsened restless legs syndrome [112].

In pregnancy, weight gain may worsen gait ataxia (SCA13) [24]. In women with SPAX5 and epilepsy, antiepileptic drugs should be reviewed to avoid drugs that are teratogenic in pregnancy and harmful during breastfeeding, and valproate should be avoided in women of childbearing potential [11]. Zonisamide was teratogenic in animal studies [142]. The sources consulted do not describe specific data on pregnancy, delivery, anaesthesia or surgery in SCA42 or in the other SCA in the class; the general precautions regarding aspiration risk in dysphagia and falls apply [1]. Transition from paediatric to adult care is relevant in the congenital forms (SCA13, SCA29, SCA42ND), in which GeneReviews recommends reassessing social and educational needs with age [24, 36].

7.14. Guideline recommendations and divergences

8. Reproductive Options and Family Counselling

8.1. Recurrence risks by scenario

In the autosomal dominant SCA of the class, most probands have an affected parent, and each child of a carrier has a 50% risk of inheriting the variant [1]. Counselling depends on zygosity and the gene, because the same gene can cause dominant and recessive disease [102].

De novo variants have been described in SCA11 (Danish family), SCA13, SCA14, SCA28 (one case), SCA42 (one Turkish case), SCA49 (frequent) and the congenital forms (SCA29, Gillespie, SCA42ND); parental mosaicism has been described in SCA29 [1, 11, 14, 24, 35, 36, 57]. A parent with germline or somatic mosaicism may be mildly affected, and leucocyte testing does not detect all mosaicism [1, 11]. When the variant is not found in the parents, the possibility of paternity or maternity other than reported, including assisted reproduction, or of undisclosed adoption should also be considered [35].

The family history may appear negative because of non-recognition of the disease, early death of the parent, late onset or reduced penetrance [36]. In SCA14 there are unaffected carriers older than 60 years (up to 76 years in a Japanese family), in SCA34 there are unaffected carriers, in SCA48 there are carriers without symptoms at 64 and 75 years, in SCA13 the Filipino family had 9 still asymptomatic carriers among 21, and in SCA47 the penetrance described as incomplete lost its significance after the original family was reassigned to SCA27B [55, 62, 63, 71, 74]; in SCA38 penetrance is 100% by 50 years and in SCA28 it appears complete [11, 36, 47, 54]. Heterozygotes for the recessive AFG3L2 forms are generally asymptomatic, but isolated cerebellar atrophy on MRI and isolated optic atrophy have been described in some [11]. In SCA49, if the variant is not detected in a relative's blood, DNA from non-haematopoietic tissue should be tested, because somatic rescue in the bone marrow can mask it [57].

8.2. Cascade and predictive testing

Testing of at-risk relatives is only possible once the pathogenic variant has been identified in an affected individual [76]. Predictive testing of asymptomatic adults follows general guidance: pre-test genetic counselling, ideally by a genetic counsellor or, failing that, by a neurologist trained in neurogenetics or a clinical geneticist; assessment of psychological state at the clinician's discretion; informed consent; absence of coercion; and post-test counselling for any result, with a support person present at disclosure [26]. The result does not predict age at onset, severity, symptoms or progression, and intrafamilial variability is the rule [36, 76]. In families with SCA48, predictive testing of STUB1 alone has uncertain value, because digenic segregation data indicate that the disease may also require an intermediate TBP allele; sizing TBP in the affected individual and in tested relatives avoids attributing risk to a single variant [74]. The US recommendations include discussion of legal protections against genetic discrimination (the US Genetic Information Nondiscrimination Act, GINA, of 2008) and of insurance implications [26]. In SCA49, testing relatives has direct clinical utility, because it identifies those who benefit from haematological surveillance [57]. A VUS cannot be used for predictive testing [36]. In SCA46, the association of PLD3 with Alzheimer disease has not been replicated and does not justify predictive testing for dementia [53].

8.3. Minors

Predictive testing of asymptomatic minors for untreatable adult-onset disorders is considered inappropriate by the professional associations cited by GeneReviews; symptomatic minors can be tested at any age [36]. CRC-SCA/NAF allows testing of asymptomatic minors only when there is an approved treatment in childhood and the familial variant has been identified, and testing of symptomatic minors after confirmation by a paediatric neurologist experienced in ataxia [26]. Ataxia UK generally does not recommend testing asymptomatic minors, allowing case-by-case assessment [27]. In SCA49, early haematological surveillance may justify testing at-risk children, according to the GeneReviews entry on SAMD9L syndrome [57].

8.4. Reproductive options

With the familial variant known, the options include prenatal diagnosis by chorionic villus sampling or amniocentesis and PGT, and family planning should be discussed before pregnancy [26, 36]. Requests for prenatal diagnosis of adult-onset disorders are uncommon and require careful counselling [47]. In families with SCA28 or OPA12, if the carrier's partner also has an AFG3L2 variant, the offspring are at risk of severe recessive disease, such as SPAX5 [11]. For de novo forms (SCA42ND), the recurrence risk for the parents is low, but slightly higher than in the population because of the possibility of germline mosaicism [1]. DNA banking of undiagnosed symptomatic patients allows relatives to be tested in the future [26, 76].

8.5. Psychosocial dimension

GeneReviews recommends family support through social services, care coordination, home nursing, respite care for carers and palliative care in the advanced stages, and renewed genetic counselling at each visit when family planning questions arise [1]. Depression was described in 3/10 patients with SCA42 in the French cohort, and anxiety and phobias marked the onset of the Turkish case [13, 14].

9. Special Considerations: Population Diversity and Epidemiology

9.1. Global frequency and what each study measures

There is no population prevalence estimate for SCA42 or for most SCA in the class; Orphanet classifies each as <1/1,000,000 and records 25 cases of SCA42 worldwide [9]. The available data are proportions in referral-centre cohorts, which measure the fraction of already selected patients (with ataxia, with dominant transmission and, in general, with expansions excluded) explained by each gene; they are not prevalences and cannot be transposed to the population [13, 22]. For the dominant ataxias as a whole, the CACNA1A curation cites a global prevalence of about 1/100,000 from the SCA6 GeneReviews entry, and Manes and colleagues cite 5.5/100,000 for SCA, both by indirect attribution [29, 60]. SCA28 has an estimated incidence of 0.045/100,000 [11].

9.2. United States

The Lincoln family, spanning 11 generations and descended from the grandparents of President Abraham Lincoln, is the largest SCA5 family and one of the three used to identify the gene [33, 34]. In the UCLA cohort, SCA42 explained 1 in 225 undiagnosed families, and the authors recommend considering it in any undiagnosed ADCA regardless of origin [22]. SCA22 was described in a North American Ashkenazi Jewish family (p.(Gly345Val)) [78].

9.3. Europe and founder populations

In Europe, notable examples are the Dutch founder variant PRKCG p.(Gly118Asp), whose carriers descend from a common ancestor from North Brabant born in 1722, and the possible French founder p.(Phe643Leu) [36, 63], ELOVL5 p.(Gly230Val) in Italian families with a shared haplotype [85], SCA23 in Dutch families [83], SCA26 in a family of Norwegian ancestry [42], SCA43 in a Belgian family [50], SCA46 in a Dutch family [53] and SCA20 in an Australian family of Anglo-Celtic ancestry [8]. CACNA1G p.(Arg1715His) has no common founder effect in the three French families (different haplotypes), indicating mutational recurrence [13]; the two Japanese families share a 360 kb haplotype, compatible with an unproven common ancestor [25], and the Korean authors suggest a possible founder variant in East Asia, a hypothesis not tested by haplotypes [112]. In SCA27B, haplotype analysis showed a founder effect in part, but not all, of the French-Canadian families [55]. SCAR13 was described in a Roma founder sub-isolate [92], and ARSACS, a recessive differential, has a carrier frequency of 1:21 in the Saguenay–Lac-Saint-Jean region of Quebec [1]. SCA35 and TGM6 variants are enriched in East Asian controls, which weighs against causality [45, 46].

9.4. Portugal and Brazil

The sources consulted do not include frequency data for SCA due to conventional variants in Portugal; the epidemiology of SCA in Portugal is dominated by Machado-Joseph disease. The OMIM entry for SCA20 mentions six Portuguese families (19 affected) with dominant ataxia preceded by spasmodic coughing fits, clinically likened to SCA20 without confirmation of the 11q12 duplication in the same source [66]. For Brazil, the only information found is a 2026 report, cited in the PDYN gene curation, of a woman with SCA23 and onset at 19 years without a family history, whose primary article was not identified [41]. In Argentina, the case of atypical parkinsonism associated with CACNA1G was described [5]. The absence of data in the sources consulted does not demonstrate the absence of cases.

9.5. Special populations

In children, a diagnosis of ataxic cerebral palsy should be reviewed when there is progression, cerebellar atrophy or absence of perinatal events (SCA21, SPTBN2, SCA42ND) [33, 40, 117]. In older people, coincident dementia and age-related dentate calcification (0.7% over 65 years) may confound interpretation [8, 36]. In pregnancy, weight gain may worsen ataxia [24]. In women of childbearing potential with epilepsy (SPAX5, neurodevelopmental CACNA1G forms), the choice of antiepileptic drug should take teratogenicity into account [11].

9.6. European network and patient associations

The relevant European reference network is ERN-RND (European Reference Network on Rare Neurological Diseases), with a group for hereditary cerebellar ataxias and spastic paraplegias and a knowledge portal on the diseases; ERN-RND endorsed the Ataxia UK guidelines [27, 32]. Patient associations include, in Portugal, the Associação Portuguesa de Ataxias Hereditárias (APAHE, www.apahe.pt); in Brazil, the Associação Brasileira de Ataxias Hereditárias e Adquiridas (ABAHE, www.abahe.org.br); in Europe, the Euro-ATAXIA federation (www.euroataxia.org), Ataxia UK (www.ataxia.org.uk), the French Connaître les Syndromes Cérébelleux (www.csc.asso.fr), the Italian AISA (www.atassia.it) and the Spanish FEDAES (fedaes.org); in the USA, the National Ataxia Foundation (www.ataxia.org), which runs support groups, research funding and a public list of therapies in development (Research Pipeline), and the CoRDS registry [1, 36, 158]. The Ataxia Global Initiative brings together research consortia for outcome standards and biomarkers [158].

10. Illustrative Clinical Case

10.1. SCA42 in a Korean family (Son and colleagues, 2023)

A previously healthy 31-year-old Korean man was referred to the neurology clinic for a progressive gait disturbance beginning at 25 years, which started as difficulty running and progressed insidiously to difficulty walking [112]. He had no developmental delay in childhood and cognition was normal. Examination showed gait ataxia, scanning speech, dysmetria and dysdiadochokinesia, with normal extraocular movements and saccades, no upper motor neuron signs and normal sensation and strength; SARA was 9 [112]. Since the age of 20 he had also had complaints meeting the criteria for restless legs syndrome (RLS-6 scale 33/60; international IRLS scale 20/40) [112].

His father was a wheelchair user because of a gait disturbance and imbalance beginning in adolescence; the other relatives were healthy, and the three-generation pedigree is in Figure 1A of the article [112]. Brain MRI showed atrophy of the vermis and cerebellar cortex, without brainstem atrophy (Figure 1C) [112].

Given a suspected autosomal dominant familial aetiology, the authors tested for the SCA1, 2, 3, 6, 7 and 17 and DRPLA expansions and a 27-gene ataxia panel, all negative. Exome sequencing identified the CACNA1G variant c.5144G>A p.(Arg1715His) (annotated on NM_198384.3, corresponding to NM_018896.5:c.5144G>A), confirmed by Sanger sequencing in the proband and his father [10, 112]. The chromatograms and MRI are in Figure 1 of the article, which was consulted [112].

On the rationale of blocking T-type channels, the patient received ethosuximide up to 1000 mg/day orally, without effect; then zonisamide up to 200 mg/day, without improvement in ataxia and with worsening of restless legs, which returned to baseline after discontinuation [112]. A second unrelated family (ancestry and absence of relatedness confirmed by exome analysis) had the same variant: a 55-year-old woman with onset at 48 years, scanning speech, gait ataxia, hypermetria and a SARA of 12.5, and a brother affected since 40 years; MRI showed severe cerebellar atrophy, and ethosuximide up to 500 mg/day and zonisamide up to 300 mg/day did not improve the ataxia [112].

10.2. Correlation with the neuropathology of the French family

Figure 1 of Coutelier and colleagues, consulted, shows in the cerebellum of a woman who died at 83 years the severe vermian atrophy, Purkinje cell loss with empty baskets and Bergmann gliosis, and the inferior olive with neuronal loss (figure 1); Figure 2 shows the pedigrees of the three families and the position of the variant in segment S4 of domain IV (figure 2) [13].

10.3. Lessons from the case

The case brings together the features that point to SCA42: relatively pure cerebellar ataxia with onset in young adulthood, slow progression, atrophy of the vermis and cerebellar cortex without brainstem involvement, vertical transmission and wide variability in age at onset within the family (adolescence in the father, 25 years in the son) [13, 112]. It also shows the usual diagnostic sequence, in which expansion testing and a restricted panel were negative and only the exome gave the diagnosis, consistent with the recommendation of a comprehensive test from the outset [26]. The absence of benefit from ethosuximide and zonisamide, with worsening of one symptom, contrasts with the report of tremor suppression with low-dose zonisamide in another patient; the use of these drugs in the reports was experimental and does not constitute a routine recommendation [30, 112]. Restless legs syndrome had not previously been described in SCA42; the authors note its higher frequency in SCA1, SCA2 and SCA3 [112].

11. Prognosis and Long-Term Follow-Up

11.1. Prognosis

Prognosis differs by entity, but most SCA due to conventional variants progress slowly and have a normal or near-normal life expectancy [11, 21, 35, 36, 47]. The data come from families and small series, without disease-modifying treatment, and annual SARA progression measures have not been published for these entities in the sources consulted.

The common risks in the advanced stages are falls, aspiration and their complications [1]. In SCA42, progression measures and lifespan have not been studied systematically.

11.2. Surveillance

There are no published surveillance guidelines for hereditary ataxias in general; the intervals below derive from the recommendations of the GeneReviews entries, based on the authors' experience [1, 11].

The surveillance intervals match those described in section 7: annual neurological and rehabilitation assessment in most entities, every 2 or 3 years in SCA20 and SCA15, six-monthly or annual electroencephalogram in SPAX5 and a full blood count every 6–12 months in SCA49 [8, 11, 21, 57].

11.3. Networked follow-up

Follow-up by a multidisciplinary team and referral to a specialist ataxia centre are recommended by Ataxia UK, and ERN-RND enables discussion of rare cases between European centres [27, 32]. Periodic reanalysis of the genomic data of undiagnosed patients (after 2 years or more) is part of follow-up [26].

12. Summary and Concluding Remarks

SCA due to conventional variants form a class of more than 25 genes and one duplication which, together, explain a minority fraction of dominant cerebellar ataxias, after polyglutamine expansions and non-coding expansions [7, 13]. The common target is the Purkinje cell, and the mechanisms range from altered ion channel gating to mitochondrial dysfunction, lipid metabolism and protein quality control [77]. SCA42, caused mainly by the recurrent variant CACNA1G p.(Arg1715His), is a slowly progressive adult cerebellar ataxia, with variable pyramidal signs and tremor, described in France, Japan, the USA, Korea, China and Turkey; through de novo gain-of-function variants in the intracellular gate, the same gene causes SCA42ND, a severe neurodevelopmental syndrome, and rare episodic, myoclonic and parkinsonian phenotypes [13, 15, 25, 89].

Diagnosis requires a comprehensive genetic test that is not limited to sequencing: expansions must be excluded or studied in parallel, copy-number analysis is mandatory for SCA15 and SCA20, and interpretation of variants in genes with many VUS, such as CACNA1G, depends on the domain, segregation and functional data [8, 9, 21, 26]. The most useful clinical clues are myoclonus and dystonia in SCA14, ophthalmoparesis and ptosis in SCA28, very slowly progressive tremor in SCA15, pes cavus and hyposmia in SCA38, erythrokeratoderma in SCA34, dentate calcification and palatal tremor in SCA20, cognitive-affective changes in SCA48 and cytopenias in SCA49 [8, 11, 12, 21, 36, 44, 47, 57].

There is no approved disease-modifying treatment. Rehabilitation and guideline-based symptomatic treatment are the basis of management; riluzole has evidence from trials in mixed populations and one negative trial in SCA2; DHA improved SARA in a trial of 10 patients with SCA38; and experience with zonisamide, ethosuximide, clonazepam and carbamazepine in SCA42 is limited to reports with divergent results [27, 29, 30, 112, 136, 137]. Troriluzole has not been studied in this class, and the November 2025 CRL is known from the sponsor's report [138]. Genetic counselling follows dominant inheritance with a 50% risk to offspring, but must incorporate the incomplete penetrance of some entities, de novo forms, mosaicism and allelic recessive forms [1, 11]. Surveillance is annual in most entities and gains a haematological component in SCA49 [57].

13. References

1. 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. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1138/
2. Pellerin D, Danzi M, Renaud M, Houlden H, Synofzik M, Zuchner S, Brais B. GAA-FGF14-Related Ataxia. 2024 Jan 25. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK599589/
3. Marras C, Lang A, van de Warrenburg BP, Sue CM, Tabrizi SJ, Bertram L, et al. Nomenclature of genetic movement disorders: Recommendations of the international Parkinson and movement disorder society task force. Mov Disord. 2016;31(4):436-57. doi:10.1002/mds.26527. PMID: 27079681.
4. Lange LM, Gonzalez-Latapi P, Rajalingam R, Tijssen MAJ, Ebrahimi-Fakhari D, Gabbert C, et al. Nomenclature of Genetic Movement Disorders: Recommendations of the International Parkinson and Movement Disorder Society Task Force - An Update. Mov Disord. 2022;37(5):905-935. doi:10.1002/mds.28982. PMID: 35481685.
5. Martínez-Villota VA, Castillo-Torres SA, Rossi M, Merello M. Rapidly Progressive Atypical Parkinsonism as a Presenting Feature of ATX-CACNA1G (SCA42). Mov Disord Clin Pract. 2023;10(4):713-715. doi:10.1002/mdc3.13704. PMID: 37070041; PMC10105115.
6. National Center for Biotechnology Information. ClinVar: VCV000221981, VCV000280269, VCV000559581 (CACNA1G) [Internet]. Bethesda (MD): NLM [cited 2026 Oct 28]. Available from: https://www.ncbi.nlm.nih.gov/clinvar/variation/221981/
7. Watanabe T, Kume K, Inoue K, Nakamura M, Yamamoto S, Kurashige T, et al. Whole exome sequencing in Japanese spinocerebellar ataxia identifies novel variants. J Hum Genet. 2026;71(1):35-39. doi:10.1038/s10038-025-01405-2. PMID: 40940404; PMC12689424.
8. Storey E, Gardner RJM. Spinocerebellar Ataxia Type 20. 2007 Feb 27 [updated 2019 Apr 18]. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1471/
9. GeneGuidance. CACNA1G — clinical curation (SCA42, SCA42ND). Version of 2026-09-26.
10. Freeman PJ, Hart RK, Gretton LJ, Brookes AJ, Dalgleish R. VariantValidator: Accurate validation, mapping, and formatting of sequence variation descriptions. Hum Mutat. 2018;39(1):61-68. doi:10.1002/humu.23348. PMID: 28967166. Accessed via https://rest.variantvalidator.org on 2026-10-28.
11. Coarelli G, Mouraux C. AFG3L2-Related Neurologic Disorders. 2011 May 17 [updated 2025 Sep 18]. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK54582/
12. GeneGuidance. STUB1 — clinical curation (SCA48, SCAR16). Version of 2026-09-26.
13. Coutelier M, Blesneac I, Monteil A, Monin ML, Ando K, Mundwiller E, et al. A Recurrent Mutation in CACNA1G Alters Cav3.1 T-Type Calcium-Channel Conduction and Causes Autosomal-Dominant Cerebellar Ataxia. Am J Hum Genet. 2015;97(5):726-37. doi:10.1016/j.ajhg.2015.09.007. PMID: 26456284; PMC4667105.
14. Saricicek G, Teke H. From early psychiatric symptoms to an adolescent diagnosis: Clues from childhood in spinocerebellar ataxia Type-42. Dusunen Adam. 2026;39(1):59-61. doi:10.14744/DAJPNS.2026.00315. PMID: 42459783; PMC13289307.
15. Chemin J, Siquier-Pernet K, Nicouleau M, Barcia G, Ahmad A, Medina-Cano D, et al. De novo mutation screening in childhood-onset cerebellar atrophy identifies gain-of-function mutations in the CACNA1G calcium channel gene. Brain. 2018;141(7):1998-2013. doi:10.1093/brain/awy145. PMID: 29878067.
16. OMIM® Online Mendelian Inheritance in Man. #616795 SPINOCEREBELLAR ATAXIA 42; SCA42. Johns Hopkins University; last edited on 11/07/2023. https://omim.org/entry/616795. Accessed 29 October 2026.
17. OMIM® Online Mendelian Inheritance in Man. *604065 CALCIUM CHANNEL, VOLTAGE-DEPENDENT, T TYPE, ALPHA-1G SUBUNIT; CACNA1G. Johns Hopkins University; last edited on 14/08/2018. https://omim.org/entry/604065. Accessed 29 October 2026.
18. OMIM® Online Mendelian Inheritance in Man. #618087 SPINOCEREBELLAR ATAXIA 42, EARLY-ONSET, SEVERE, WITH NEURODEVELOPMENTAL DEFICITS; SCA42ND. Johns Hopkins University; last edited on 02/01/2023. https://omim.org/entry/618087. Accessed 29 October 2026.
19. Orphanet. Autosomal dominant cerebellar ataxia (ORPHA:99) and SCA entries listed in fontes/orphanet_sca_local.txt. Available from: https://www.orpha.net/en/disease/detail/99
20. Orphanet. Spinocerebellar ataxia type 42 (ORPHA:458803). Available from: https://www.orpha.net/en/disease/detail/458803
21. Storey E. Spinocerebellar Ataxia Type 15 [entry withdrawn on 25 Apr 2024; historical reference]. 2006 May 30 [updated 2014 Jun 12]. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1116/
22. Ngo K, Aker M, Petty LE, Chen J, Cavalcanti F, Nelson AB, et al. Expanding the global prevalence of spinocerebellar ataxia type 42. Neurol Genet. 2018;4(3):e232. doi:10.1212/NXG.0000000000000232. PMID: 29629410; PMC5886831.
23. Kimura M, Yabe I, Hama Y, Eguchi K, Ura S, Tsuzaka K, et al. SCA42 mutation analysis in a case series of Japanese patients with spinocerebellar ataxia. J Hum Genet. 2017;62(9):857-859. doi:10.1038/jhg.2017.51. PMID: 28490766.
24. Waters MF. Spinocerebellar Ataxia Type 13. 2006 Nov 9 [updated 2020 Jun 4]. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1225/
25. Morino H, Matsuda Y, Muguruma K, Miyamoto R, Ohsawa R, Ohtake T, et al. A mutation in the low voltage-gated calcium channel CACNA1G alters the physiological properties of the channel, causing spinocerebellar ataxia. Mol Brain. 2015;8:89. doi:10.1186/s13041-015-0180-4. PMID: 26715324; PMC4693440.
26. Srinivasan SR, Mook AD, Rochman M, Chen JYH, Mu W, Wilmot GR, et al. Practice Recommendations for Genetic Testing of Ataxias. Ann Clin Transl Neurol. 2025;12(12):2398-2409. doi:10.1002/acn3.70171. PMID: 40883955; PMC12698944.
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. Barresi S, Dentici ML, Manzoni F, Bellacchio E, Agolini E, Pizzi S, et al. Infantile-Onset Syndromic Cerebellar Ataxia and CACNA1G Mutations. Pediatr Neurol. 2020;104:40-45. doi:10.1016/j.pediatrneurol.2019.09.005. PMID: 31836334.
29. Manes M, Alberici A, Di Gregorio E, Boccone L, Premi E, Mitro N, et al. Docosahexaenoic acid is a beneficial replacement treatment for spinocerebellar ataxia 38. Ann Neurol. 2017;82(4):615-621. doi:10.1002/ana.25059. PMID: 28976605; PMC5698802.
30. Hara N, Morino H, Matsuda Y, Satoh K, Hashimoto K, Maruyama H, et al. Zonisamide can ameliorate the voltage-dependence alteration of the T-type calcium channel Ca(V)3.1 caused by a mutation responsible for spinocerebellar ataxia. Mol Brain. 2020;13(1):163. doi:10.1186/s13041-020-00700-7. PMID: 33243296; PMC7690142.
31. De Riggi M, De Giorgi A, Pollini L, Angelini L, Paparella G, Cannavacciuolo A, et al. CACNA1G Causes Dominantly Inherited Myoclonus-Ataxia with Intellectual Disability: A Case Report. Cerebellum. 2024;23(6):2679-2683. doi:10.1007/s12311-024-01734-6. PMID: 39287920; PMC11585495.
32. European Reference Network on Rare Neurological Diseases (ERN-RND). Disease Knowledge Hub — Cerebellar Ataxia & Hereditary Spastic Paraplegias [Internet]. Available from: https://www.ern-rnd.eu/disease-knowledge-hub/ (accessed 28 Oct 2026).
33. GeneGuidance. SPTBN2 — clinical curation (SCA5, SCAR14). Version of 2026-09-26.
34. Ikeda Y, Dick KA, Weatherspoon MR, Gincel D, Armbrust KR, Dalton JC, et al. Spectrin mutations cause spinocerebellar ataxia type 5. Nat Genet. 2006;38(2):184-90. doi:10.1038/ng1728. PMID: 16429157.
35. Chen Z, Puzriakova A, Houlden H. Spinocerebellar Ataxia Type 11. 2008 Jul 22 [updated 2019 Oct 31]. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1757/
36. Chen DH, Bird TD, Raskind WH. Spinocerebellar Ataxia Type 14. 2005 Jan 28 [updated 2020 Feb 20]. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1399/
37. Chelban V, Wiethoff S, Fabian-Jessing BK, Haridy NA, Khan A, Efthymiou S, et al. Genotype-phenotype correlations, dystonia and disease progression in spinocerebellar ataxia type 14. Mov Disord. 2018;33(7):1119-1129. doi:10.1002/mds.27334. PMID: 29603387; PMC6175136.
38. GeneGuidance. ITPR1 — clinical curation (SCA15/16, SCA29, Gillespie). Version of 2026-09-26.
39. GeneGuidance. KCND3 — clinical curation (SCA19/22). Version of 2026-09-26.
40. GeneGuidance. TMEM240 — clinical curation (SCA21). Version of 2026-09-26.
41. GeneGuidance. PDYN — clinical curation (SCA23). Version of 2026-09-26.
42. GeneGuidance. EEF2 — clinical curation (SCA26). Version of 2026-09-26.
43. GeneGuidance. FGF14 — clinical curation (SCA27A, SCA27B). Version of 2026-09-26.
44. Nishide M, Le Marquand K, Davis MR, Halmágyi GM, Fellner A, Narayanan RK, et al. Two New Families and a Literature Review of ELOVL4-Associated Spinocerebellar Ataxia Type 34. Cerebellum. 2024;23(1):268-277. doi:10.1007/s12311-023-01522-8. PMID: 36696030; PMC10864522.
45. GeneGuidance. TGM6 — clinical curation (SCA35). Version of 2026-09-26.
46. Fung JLF, Tsang MHY, Leung GKC, Yeung KS, Mak CCY, Fung CW, et al. A significant inflation in TGM6 genetic risk casts doubt in its causation in spinocerebellar ataxia type 35. Parkinsonism Relat Disord. 2019;63:42-45. doi:10.1016/j.parkreldis.2019.01.013. PMID: 30670339.
47. Brusco A, Di Gregorio E, Borroni B. Spinocerebellar Ataxia Type 38. 2019 Jul 11. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK543515/
48. GeneGuidance. CCDC88C — clinical curation (SCA40). Version of 2026-09-26.
49. GeneGuidance. TRPC3 — clinical curation (SCA41). Version of 2026-09-26.
50. Depondt C, Donatello S, Rai M, Wang FC, Manto M, Simonis N, et al. MME mutation in dominant spinocerebellar ataxia with neuropathy (SCA43). Neurol Genet. 2016;2(5):e94. doi:10.1212/NXG.0000000000000094. PMID: 27583304; PMC4991603.
51. Watson LM, Bamber E, Schnekenberg RP, Williams J, Bettencourt C, Lickiss J, et al. Dominant Mutations in GRM1 Cause Spinocerebellar Ataxia Type 44. Am J Hum Genet. 2017;101(3):451-458. doi:10.1016/j.ajhg.2017.08.005. PMID: 28886343; PMC5591020.
52. GeneGuidance. FAT2 — clinical curation (SCA45). Version of 2026-09-26.
53. GeneGuidance. PLD3 — clinical curation (SCA46). Version of 2026-09-26.
54. Gennarino VA, Palmer EE, McDonell LM, Wang L, Adamski CJ, Koire A, et al. A Mild PUM1 Mutation Is Associated with Adult-Onset Ataxia, whereas Haploinsufficiency Causes Developmental Delay and Seizures. Cell. 2018;172(5):924-936.e11. doi:10.1016/j.cell.2018.02.006. PMID: 29474920; PMC5832058.
55. OMIM® Online Mendelian Inheritance in Man. #620174 SPINOCEREBELLAR ATAXIA 27B, LATE-ONSET; SCA27B. Johns Hopkins University; last edited on 02/07/2025. https://omim.org/entry/620174. Accessed 29 October 2026.
56. Genis D, Ortega-Cubero S, San Nicolás H, Corral J, Gardenyes J, de Jorge L, et al. Heterozygous STUB1 mutation causes familial ataxia with cognitive affective syndrome (SCA48). Neurology. 2018;91(21):e1988-e1998. doi:10.1212/WNL.0000000000006550. PMID: 30381368.
57. Raskind WH, Keel S, Allenspach EJ. SAMD9L-Related Syndrome. 2017 Jun 1 [updated 2026 Aug 20]. In: Adam MP, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK435692/
58. Coutelier M, Jacoupy M, Janer A, Renaud F, Auger N, Saripella GV, et al. NPTX1 mutations trigger endoplasmic reticulum stress and cause autosomal dominant cerebellar ataxia. Brain. 2022;145(4):1519-1534. doi:10.1093/brain/awab407. PMID: 34788392.
59. GeneGuidance. NPTX1 — clinical curation (SCA50). Version of 2026-09-26.
60. GeneGuidance. CACNA1A — clinical curation (FHM1, EA2, SCA6, DEE42). Version of 2026-09-26.
61. OMIM® Online Mendelian Inheritance in Man. #604432 SPINOCEREBELLAR ATAXIA 11; SCA11. Johns Hopkins University; last edited on 03/04/2024. https://omim.org/entry/604432. Accessed 29 October 2026.
62. OMIM® Online Mendelian Inheritance in Man. #605259 SPINOCEREBELLAR ATAXIA 13; SCA13. Johns Hopkins University; last edited on 27/11/2023. https://omim.org/entry/605259. Accessed 29 October 2026.
63. OMIM® Online Mendelian Inheritance in Man. #605361 SPINOCEREBELLAR ATAXIA 14; SCA14. Johns Hopkins University; last edited on 30/08/2016. https://omim.org/entry/605361. Accessed 29 October 2026.
64. OMIM® Online Mendelian Inheritance in Man. #606658 SPINOCEREBELLAR ATAXIA 15; SCA15. Johns Hopkins University; last edited on 06/09/2023. https://omim.org/entry/606658. Accessed 29 October 2026.
65. OMIM® Online Mendelian Inheritance in Man. #607346 SPINOCEREBELLAR ATAXIA 19; SCA19. Johns Hopkins University; last edited on 07/11/2023. https://omim.org/entry/607346. Accessed 29 October 2026.
66. OMIM® Online Mendelian Inheritance in Man. #608687 SPINOCEREBELLAR ATAXIA 20; SCA20. Johns Hopkins University; last edited on 04/12/2023. https://omim.org/entry/608687. Accessed 29 October 2026.
67. OMIM® Online Mendelian Inheritance in Man. #193003 SPINOCEREBELLAR ATAXIA 27A; SCA27A. Johns Hopkins University; last edited on 17/03/2023. https://omim.org/entry/193003. Accessed 29 October 2026.
68. OMIM® Online Mendelian Inheritance in Man. #610246 SPINOCEREBELLAR ATAXIA 28; SCA28. Johns Hopkins University; last edited on 08/05/2026. https://omim.org/entry/610246. Accessed 29 October 2026.
69. OMIM® Online Mendelian Inheritance in Man. #614487 SPASTIC ATAXIA 5, AUTOSOMAL RECESSIVE; SPAX5. Johns Hopkins University; last edited on 07/08/2024. https://omim.org/entry/614487. Accessed 29 October 2026.
70. OMIM® Online Mendelian Inheritance in Man. #117360 SPINOCEREBELLAR ATAXIA 29; SCA29. Johns Hopkins University; last edited on 21/12/2022. https://omim.org/entry/117360. Accessed 29 October 2026.
71. OMIM® Online Mendelian Inheritance in Man. #133190 SPINOCEREBELLAR ATAXIA 34; SCA34. Johns Hopkins University; last edited on 16/05/2024. https://omim.org/entry/133190. Accessed 29 October 2026.
72. OMIM® Online Mendelian Inheritance in Man. #615957 SPINOCEREBELLAR ATAXIA 38; SCA38. Johns Hopkins University; last edited on 15/08/2024. https://omim.org/entry/615957. Accessed 29 October 2026.
73. OMIM® Online Mendelian Inheritance in Man. #615768 SPINOCEREBELLAR ATAXIA, AUTOSOMAL RECESSIVE 16; SCAR16. Johns Hopkins University; last edited on 24/06/2021. https://omim.org/entry/615768. Accessed 29 October 2026.
74. OMIM® Online Mendelian Inheritance in Man. #618093 SPINOCEREBELLAR ATAXIA 48; SCA48. Johns Hopkins University; last edited on 24/03/2026. https://omim.org/entry/618093. Accessed 29 October 2026.
75. OMIM® Online Mendelian Inheritance in Man. #620158 SPINOCEREBELLAR ATAXIA 50; SCA50. Johns Hopkins University; last edited on 24/03/2026. https://omim.org/entry/620158. Accessed 29 October 2026.
76. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 70–72 (Ataxic adult).
77. Klockgether T, Mariotti C, Paulson HL. Spinocerebellar ataxia. Nat Rev Dis Primers. 2019;5(1):24. doi:10.1038/s41572-019-0074-3. PMID: 30975995.
78. Lee YC, Durr A, Majczenko K, Huang YH, Liu YC, Lien CC, et al. Mutations in KCND3 cause spinocerebellar ataxia type 22. Ann Neurol. 2012;72(6):859-69. doi:10.1002/ana.23701. PMID: 23280837; PMC4085146.
79. Chen DH, Brkanac Z, Verlinde CL, Tan XJ, Bylenok L, Nochlin D, et al. Missense mutations in the regulatory domain of PKC gamma: a new mechanism for dominant nonepisodic cerebellar ataxia. Am J Hum Genet. 2003;72(4):839-49. doi:10.1086/373883. PMID: 12644968; PMC1180348.
80. Houlden H, Johnson J, Gardner-Thorpe C, Lashley T, Hernandez D, Worth P, et al. Mutations in TTBK2, encoding a kinase implicated in tau phosphorylation, segregate with spinocerebellar ataxia type 11. Nat Genet. 2007;39(12):1434-6. doi:10.1038/ng.2007.43. PMID: 18037885.
81. van de Leemput J, Chandran J, Knight MA, Holtzclaw LA, Scholz S, Cookson MR, et al. Deletion at ITPR1 underlies ataxia in mice and spinocerebellar ataxia 15 in humans. PLoS Genet. 2007;3(6):e108. doi:10.1371/journal.pgen.0030108. PMID: 17590087; PMC1892049.
82. Di Bella D, Lazzaro F, Brusco A, Plumari M, Battaglia G, Pastore A, et al. Mutations in the mitochondrial protease gene AFG3L2 cause dominant hereditary ataxia SCA28. Nat Genet. 2010;42(4):313-21. doi:10.1038/ng.544. PMID: 20208537.
83. Bakalkin G, Watanabe H, Jezierska J, Depoorter C, Verschuuren-Bemelmans C, Bazov I, et al. Prodynorphin mutations cause the neurodegenerative disorder spinocerebellar ataxia type 23. Am J Hum Genet. 2010;87(5):593-603. doi:10.1016/j.ajhg.2010.10.001. PMID: 21035104; PMC2978951.
84. Duarri A, Jezierska J, Fokkens M, Meijer M, Schelhaas HJ, den Dunnen WF, et al. Mutations in potassium channel kcnd3 cause spinocerebellar ataxia type 19. Ann Neurol. 2012;72(6):870-80. doi:10.1002/ana.23700. PMID: 23280838.
85. Di Gregorio E, Borroni B, Giorgio E, Lacerenza D, Ferrero M, Lo Buono N, et al. ELOVL5 mutations cause spinocerebellar ataxia 38. Am J Hum Genet. 2014;95(2):209-17. doi:10.1016/j.ajhg.2014.07.001. PMID: 25065913; PMC4129408.
86. Delplanque J, Devos D, Huin V, Genet A, Sand O, Moreau C, et al. TMEM240 mutations cause spinocerebellar ataxia 21 with mental retardation and severe cognitive impairment. Brain. 2014;137(Pt 10):2657-63. doi:10.1093/brain/awu202. PMID: 25070513.
87. Nibbeling EAR, Duarri A, Verschuuren-Bemelmans CC, Fokkens MR, Karjalainen JM, Smeets CJLM, et al. Exome sequencing and network analysis identifies shared mechanisms underlying spinocerebellar ataxia. Brain. 2017;140(11):2860-2878. doi:10.1093/brain/awx251. PMID: 29053796.
88. Coarelli G, Coutelier M, Durr A. Autosomal dominant cerebellar ataxias: new genes and progress towards treatments. Lancet Neurol. 2023;22(8):735-749. doi:10.1016/S1474-4422(23)00068-6. PMID: 37479376.
89. Qebibo L, Davakan A, Nesson-Dauphin M, Boulali N, Siquier-Pernet K, Afenjar A, et al. The characterization of new de novo CACNA1G variants affecting the intracellular gate of Cav3.1 channel broadens the spectrum of neurodevelopmental phenotypes in SCA42ND. Genet Med. 2025;27(3):101337. doi:10.1016/j.gim.2024.101337. PMID: 39674904.
90. Berecki G, Helbig KL, Ware TL, Grinton B, Skraban CM, Marsh ED, et al. Novel Missense CACNA1G Mutations Associated with Infantile-Onset Developmental and Epileptic Encephalopathy. Int J Mol Sci. 2020;21(17):. doi:10.3390/ijms21176333. PMID: 32878331; PMC7503748.
91. Gazulla J, Izquierdo-Alvarez S, Ruiz-Fernández E, Lázaro-Romero A, Berciano J. Episodic Vestibulocerebellar Ataxia Associated with a CACNA1G Missense Variant. Case Rep Neurol. 2021;13(2):347-354. doi:10.1159/000515974. PMID: 34248568; PMC8255690.
92. GeneGuidance. GRM1 — clinical curation (SCA44, SCAR13). Version of 2026-09-26.
93. GeneGuidance. ELOVL4 — clinical curation (SCA34). Version of 2026-09-26.
94. GeneGuidance. MME — clinical curation (CMT2T, SCA43). Version of 2026-09-26.
95. GeneGuidance. TTBK2 — clinical curation (SCA11). Version of 2026-09-26.
96. Davakan A, Cmarko L, Ribeiro Oliveira-Mendes B, Bernat C, Boulali N, Montnach J, et al. Electrophysiological classification of CACNA1G gene variants associated with neurodevelopmental and neurological disorders. Front Pharmacol. 2025;16:1613072. doi:10.3389/fphar.2025.1613072. PMID: 41111510; PMC12528054.
97. Kim JI, Miura Y, Li MY, Revah O, Selvaraj S, Birey F, et al. Human assembloids reveal the consequences of CACNA1G gene variants in the thalamocortical pathway. Neuron. 2024;112(24):4048-4059.e7. doi:10.1016/j.neuron.2024.09.020. PMID: 39419023.
98. Riquet A, Cleuziou P, Floret V, Quesque F, Defoort S, Smol T. Paroxysmal Tonic Upgaze in a Patient With Congenital Ataxia due to a De Novo Missense Variant of CACNA1G. Pediatr Neurol. 2023;139:22-23. doi:10.1016/j.pediatrneurol.2022.11.003. PMID: 36508879.
99. Finol-Urdaneta RK, Tan C, Maksemous N, Ma JG, Lockhart PJ, Snell P, et al. Calibrated high-throughput electrophysiology enables clinical interpretation of CACNA1G missense variants [preprint]. bioRxiv. 2026. doi:10.64898/2026.05.10.724145.
100. Li X, Zhou C, Cui L, Zhu L, Du H, Liu J, et al. A case of a novel CACNA1G mutation from a Chinese family with SCA42: A case report and literature review. Medicine (Baltimore). 2018;97(36):e12148. doi:10.1097/MD.0000000000012148. PMID: 30200108; PMC6133555.
101. GeneGuidance. PRKCG — clinical curation (SCA14). Version of 2026-09-26.
102. GeneGuidance. AFG3L2 — clinical curation. Version of 2026-09-26.
103. National Center for Biotechnology Information. ClinVar: VCV000005470, VCV000005471, VCV000162524 (AFG3L2) [Internet]. Bethesda (MD): NLM [cited 2026 Nov 01]. Available from: https://www.ncbi.nlm.nih.gov/clinvar/variation/5470/
104. GeneGuidance. KCNC3 — clinical curation (SCA13). Version of 2026-09-26.
105. GeneGuidance. ELOVL5 — clinical curation (SCA38). Version of 2026-09-26.
106. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 42, 42.4.8 ELOVL4 Deficiency; 42.4.9 ELOVL5 Deficiency.
107. Hekman KE, Yu GY, Brown CD, Zhu H, Du X, Gervin K, et al. A conserved eEF2 coding variant in SCA26 leads to loss of translational fidelity and increased susceptibility to proteostatic insult. Hum Mol Genet. 2012;21(26):5472-83. doi:10.1093/hmg/dds392. PMID: 23001565; PMC3516132.
108. Tsoi H, Yu AC, Chen ZS, Ng NK, Chan AY, Yuen LY, et al. A novel missense mutation in CCDC88C activates the JNK pathway and causes a dominant form of spinocerebellar ataxia. J Med Genet. 2014;51(9):590-5. doi:10.1136/jmedgenet-2014-102333. PMID: 25062847; PMC4145425.
109. Fogel BL, Hanson SM, Becker EB. Do mutations in the murine ataxia gene TRPC3 cause cerebellar ataxia in humans?. Mov Disord. 2015;30(2):284-6. doi:10.1002/mds.26096. PMID: 25477146; PMC4318721.
110. GeneGuidance. PUM1 — clinical curation (SCA47). Version of 2026-09-26.
111. GeneGuidance. SAMD9L — clinical curation (SCA49). Version of 2026-09-26.
112. Son H, Yoon JG, Kim MJ, Moon J, Kim HJ. First Cases of Spinocerebellar Ataxia 42 in Two Korean Families. J Mov Disord. 2023;16(1):110-113. doi:10.14802/jmd.22150. PMID: 36628426; PMC9978254.
113. Kunii M, Doi H, Hashiguchi S, Matsuishi T, Sakai Y, Iai M, et al. De novo CACNA1G variants in developmental delay and early-onset epileptic encephalopathies. J Neurol Sci. 2020;416:117047. doi:10.1016/j.jns.2020.117047. PMID: 32736238.
114. Selek A, Orpay F, Unay B. A novel de novo CACNA1G variant p.(Arg1553Gln) associated with neurodevelopmental delay and cerebellar hypoplasia: Expanding the phenotypic spectrum. Eur J Med Genet. 2026;82:105088. doi:10.1016/j.ejmg.2026.105088. PMID: 42309484.
115. Beijer D, Mengel D, Önder D, Wilke C, Traschütz A, Faber J, et al. The genetic landscape of sporadic adult-onset degenerative ataxia: a multi-modal genetic study of 377 consecutive patients from the longitudinal multi-centre SPORTAX cohort. EBioMedicine. 2025;115:105715. doi:10.1016/j.ebiom.2025.105715. PMID: 40273470; PMC12051541.
116. Wang JL, Yang X, Xia K, Hu ZM, Weng L, Jin X, et al. TGM6 identified as a novel causative gene of spinocerebellar ataxias using exome sequencing. Brain. 2010;133(Pt 12):3510-8. doi:10.1093/brain/awq323. PMID: 21106500.
117. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 74–77 (Ataxic child).
118. Coutelier M, Hammer MB, Stevanin G, Monin ML, Davoine CS, Mochel F, et al. Efficacy of Exome-Targeted Capture Sequencing to Detect Mutations in Known Cerebellar Ataxia Genes. JAMA Neurol. 2018;75(5):591-599. doi:10.1001/jamaneurol.2017.5121. PMID: 29482223; PMC5885259.
119. Meli A, Montano V, Palermo G, Fogli A, Rocchi A, Gerfo AL, et al. Diagnosis of hereditary ataxias: a real-world single center experience. J Neurol. 2025;272(2):111. doi:10.1007/s00415-024-12772-9. PMID: 39812846.
120. 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 [retirado]: report of the Guideline Development, Dissemination, and Implementation Subcommittee of the American Academy of Neurology. Neurology. 2018;90(10):464–71. doi:10.1212/WNL.0000000000005055. PMID: 29440566.
121. 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.
122. Baclofen tablets [labelling; ANDA209102]. DailyMed set id 020632f4-5e8a-f4a0-e063-6294a90a6301; revised 2025-01-28. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=020632f4-5e8a-f4a0-e063-6294a90a6301
123. Covis Pharma. Zanaflex (tizanidine) capsules and tablets [labelling; NDA020397, NDA021447]. DailyMed set id 043d9e51-bfa2-4add-9058-5ece332f7e99; revised 2025-05-05. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=043d9e51-bfa2-4add-9058-5ece332f7e99
124. Par Health. Dantrolene sodium capsules [labelling; NDA017443]. DailyMed set id 11689f56-15e0-4350-b53f-ae96e5994fea; revised 2026-08-21. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=11689f56-15e0-4350-b53f-ae96e5994fea
125. Gabapentin capsules [labelling; ANDA207099]. DailyMed set id 030e656a-e384-49ac-830f-32140c18112d; revised 2019-01-07. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=030e656a-e384-49ac-830f-32140c18112d
126. Propranolol hydrochloride tablets [labelling; ANDA071972]. DailyMed set id 013a485a-4227-46ae-b19f-c550302fcb96; revised 2024-08-09. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=013a485a-4227-46ae-b19f-c550302fcb96
127. Primidone tablets [labelling; ANDA040866]. DailyMed set id 1c56d09c-4224-4875-bb22-e8bfd595d378; revised 2023-03-03. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=1c56d09c-4224-4875-bb22-e8bfd595d378
128. INFARMED, I.P. Inderal 10 mg and 40 mg coated tablets (propranolol), Atnahs Pharma Netherlands: summary of product characteristics (Infomed, med_id 10157); first authorisation on 27 Mar 1967; text approved on 8 Jul 2022. Infomed. Accessed 29 Oct 2026.
129. INFARMED, I.P. Mysoline 250 mg tablets (primidone), SERB: summary of product characteristics (Infomed, single result of the search for primidone); text revised in Apr 2026. Infomed. Accessed 29 Oct 2026.
130. INFARMED, I.P. Neurontin 300 mg capsules (gabapentin), Viatris Healthcare: summary of product characteristics (Infomed, med_id 6057); first authorisation on 27 Sep 1996. Infomed. Accessed 29 Oct 2026.
131. INFARMED, I.P. Lioresal 10 mg and 25 mg tablets (baclofen), Novartis Farma: summary of product characteristics (Infomed, med_id 5075). Infomed. Accessed 29 Oct 2026.
132. INFARMED, I.P. Rivotril 0.5 mg and 2 mg tablets (clonazepam), Cheplapharm Arzneimittel: summary of product characteristics (Infomed, med_id 7612); first authorisation on 16 Oct 1974; text approved on 20 Oct 2022. Infomed. Accessed 29 Oct 2026.
133. Sun Pharmaceutical Industries. Riluzole tablets, USP [labelling; ANDA091417; reference Rilutek®]. DailyMed set id 4c851be4-f0d0-4b24-baf9-427ddfc6f7b0; revised 2020-05-16. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=4c851be4-f0d0-4b24-baf9-427ddfc6f7b0
134. European Medicines Agency. Rilutek (riluzole): EPAR [Internet]. Amsterdam: EMA; authorisation 1996-06-10, updated 2024-08-21. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/rilutek
135. 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.
136. 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.
137. Coarelli G, Heinzmann A, Ewenczyk C, Fischer C, Chupin M, Monin ML, et al. Safety and efficacy of riluzole in spinocerebellar ataxia type 2 in France (ATRIL): a multicentre, randomised, double-blind, placebo-controlled trial. Lancet Neurol. 2022;21(3):225-233. doi:10.1016/S1474-4422(21)00457-9. PMID: 35063116.
138. Biohaven Ltd. Form 10-K, fiscal year ended December 31, 2025 [Internet]. Washington (DC): U.S. Securities and Exchange Commission; 2026 [cited 2026 Oct 28]. Available from: https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001935979&type=10-K .
139. Manes M, Alberici A, Di Gregorio E, Boccone L, Premi E, Mitro N, et al. Long-term efficacy of docosahexaenoic acid (DHA) for Spinocerebellar Ataxia 38 (SCA38) treatment: An open label extension study. Parkinsonism Relat Disord. 2019;63:191-194. doi:10.1016/j.parkreldis.2019.02.040. PMID: 30862453.
140. Gazulla J, Benavente I, García-González E, Berciano J. Two-year follow-up of docosahexaenoic acid supplementation in spinocerebellar ataxia type 38 (SCA38). J Neurol. 2022;269(10):5643-5646. doi:10.1007/s00415-022-11138-3. PMID: 35438348.
141. Hara N, Nezu T, Kobatake K, Morino H, Kawakami H, Maruyama H. Treatment of intractable resting tremor of spinocerebellar ataxia 42 with zonisamide. J Neurol Sci. 2019;396:119-120. doi:10.1016/j.jns.2018.11.013. PMID: 30448718.
142. Preferred Pharmaceuticals. Zonisamide capsules [labelling; ANDA077625; reference Zonegran®]. DailyMed set id 061fdccc-d08e-4e42-8a32-54afb6c89701; revised 2025-08-01. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=061fdccc-d08e-4e42-8a32-54afb6c89701
143. European Medicines Agency. Zonegran (zonisamide): EPAR — Product Information (SmPC section 4.2) [Internet]. Amsterdam: EMA; authorisation 2005-03-10, EC decision 2026-07-21. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/zonegran
144. Parke-Davis/Pfizer. Zarontin (ethosuximide) capsules [labelling; NDA012380]. DailyMed set id 0e008f33-70a1-4bc6-b3a0-d45214418ab6; revised 2026-05-25. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=0e008f33-70a1-4bc6-b3a0-d45214418ab6
145. Clonazepam tablets [labelling; ANDA077147; reference Klonopin®]. DailyMed set id 00f323c0-e007-40df-ae7f-e6b3c7ba0aa8; revised 2026-03-16. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=00f323c0-e007-40df-ae7f-e6b3c7ba0aa8
146. Carbamazepine [labelling; ANDA207798; reference Tegretol®]. DailyMed set id 047bc284-060a-4db9-bf37-75d10f95a0a6; revised 2025-02-18. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=047bc284-060a-4db9-bf37-75d10f95a0a6
147. UCB. Keppra XR (levetiracetam) extended-release tablets [labelling; NDA022285]. DailyMed set id 2919e43b-69a8-434c-a2d2-1f3ecd7554c0; revised 2025-06-24. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=2919e43b-69a8-434c-a2d2-1f3ecd7554c0
148. Valproic acid oral solution [labelling; ANDA073178]. DailyMed set id 1286ab49-fa34-488e-e063-6394a90ad582; revised 2024-05-27. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=1286ab49-fa34-488e-e063-6394a90ad582
149. Avet Pharmaceuticals. Acetazolamide tablets [labelling; ANDA205530]. DailyMed set id 0700a910-79aa-4431-aecf-637fe14fcae8; revised 2026-05-12. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=0700a910-79aa-4431-aecf-637fe14fcae8
150. European Medicines Agency. Fampyra (fampridine): EPAR [Internet]. Amsterdam: EMA; authorisation 2011-07-20, EC decision 2025-08-06. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/fampyra
151. Dalfampridine extended-release tablets [labelling; ANDA206765; reference Ampyra®]. DailyMed set id 1134ac0f-59c1-a20b-e063-6294a90affcc; revised 2026-07-28. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=1134ac0f-59c1-a20b-e063-6294a90affcc
152. INFARMED, I.P. Riluzol Aurovitas 50 mg film-coated tablets (riluzole), Generis Farmacêutica: summary of product characteristics (Infomed, med_id 45918); text approved on 9 Feb 2023. Infomed. Accessed 29 Oct 2026.
153. INFARMED, I.P. Tegretol 200 mg tablets (carbamazepine), Novartis Farma: summary of product characteristics (Infomed, med_id 8335); text revised on 9 Sep 2026. Infomed. Accessed 29 Oct 2026.
154. INFARMED, I.P. Sirdalud 2 mg tablets (tizanidine), Sandoz Farmacêutica: summary of product characteristics (Infomed, med_id 7909); first authorisation on 31 Jan 1992. Infomed. Accessed 29 Oct 2026.
155. INFARMED, I.P. Infomed — database of medicinal products for human use: searches for dantrolene, fampridine, zonisamide and levetiracetam (authorised/marketed filter) [Internet]. Available from: https://extranet.infarmed.pt/INFOMED-fo/. Accessed 29 Oct 2026.
156. European Medicines Agency. Keppra (levetiracetam): EPAR — Product information (Annex I, SmPC, sections 4.1–4.2) [Internet]. Amsterdam: EMA. Available from: https://www.ema.europa.eu/pt/documents/product-information/keppra-epar-product-information_pt.pdf. Accessed 29 Oct 2026.
157. INFARMED, I.P. Depakine Chrono 500 mg prolonged-release tablets (sodium valproate and valproic acid), Sanofi - Produtos Farmacêuticos: summary of product characteristics (Infomed, med_id 2415); first authorisation on 27 Nov 1989. Infomed. Accessed 29 Oct 2026.
158. Patient associations and ataxia research consortia [Internet]: APAHE (https://www.apahe.pt/), ABAHE (https://www.abahe.org.br/), Euro-ATAXIA (https://www.euroataxia.org/), Ataxia UK (https://www.ataxia.org.uk/), Connaître les Syndromes Cérébelleux (https://www.csc.asso.fr/), AISA (https://www.atassia.it/), National Ataxia Foundation (https://www.ataxia.org/; accessed 29 Oct 2026), Ataxia Global Initiative (https://ataxia-global-initiative.net/). Accessed 28 Oct 2026.