Hereditary Spastic Paraplegias

Literature and regulatory information updated: 6 October 2026.

Pure and complicated classification, inheritance patterns, diagnostic approach and differential diagnosis, with dedicated subsections on SPG4/SPAST, SPG7, SPG11, SPG3A/ATL1, SPG76/CAPN1 and AP-4 deficiency (SPG47, SPG50, SPG51, SPG52)

0. Initial Reference Table

1. Introduction

Hereditary spastic paraplegias constitute a group of monogenic disorders in which corticospinal tract axons fail to develop normally or, more frequently, undergo progressive degeneration after initially normal development [4]. The central manifestation, common to all forms, is bilateral, symmetrical, slowly progressive spastic paraparesis of the lower limbs; this paraparesis occurs relatively in isolation in pure forms and accompanied by other neurological or extraneurological signs in complicated forms [4, 24]. Histopathological studies of pure forms characterised neurodegeneration as length-dependent retrograde (dying back) degeneration of the terminal ends of corticospinal tract and dorsal column axons, explaining why the lower limbs are affected first and predominantly [4].

The entity's history spans two centuries. Charles-Prosper Ollivier d'Angers outlined a suggestive description in 1827; Wilhelm Heinrich Erb described the clinical picture of “spastic spinal paralysis” in 1875, and Jean-Martin Charcot began teaching it as a clinical entity that same year [1]. Adolf von Strümpell recognised its hereditary nature in 1880, and Maurice Lorrain associated his name with Strümpell's after his 1898 thesis, the first review of twenty-nine affected families, prepared under Fulgence Raymond's supervision [1, 25]. Modern clinical systematisation is attributable to Anita Harding: in 1981, in a study of 22 families with the “pure” form, she identified two AD-HSP types, one with predominantly pre-35-year onset (type I), very slow progression, spasticity more marked than weakness and not infrequently delayed walking, and another with onset usually after 35 years (type II), greater weakness, urinary symptoms and sensory loss, and more rapid progression [26]. In 1983, she proposed the classification of hereditary ataxias and paraplegias that established the distinction between pure and complicated forms still used in clinical practice [3].

The molecular era began with positional cloning of the first genes. In 1998, Casari and colleagues identified, in a family linked to chromosome 16q24.3, a homozygous 9.5 kb deletion in a gene encoding paraplegin, a mitochondrial metalloprotease, and documented signs of oxidative phosphorylation (OXPHOS) deficiency in muscle biopsies [27]. In 1999, Hazan and colleagues identified spastin, an AAA (ATPases associated with diverse cellular activities) family protein, as the cause of SPG4, linked to chromosome 2p22 and accounting for 40-50% of AD-HSP families [28]. In 2001, Zhao and colleagues described mutations in a new GTPase gene, atlastin, at the SPG3A locus, then implicated in approximately 9% of AD-HSP families [29]. In 2007, Stevanin and colleagues identified spatacsin (SPG11) as a major cause of AR-HSP with a thin corpus callosum [30], and in 2016 two independent teams associated biallelic CAPN1 variants, encoding calpain 1, with an AR form designated SPG76, with models in Caenorhabditis elegans, Drosophila melanogaster, zebrafish and mouse [12, 31].

The number of entities has grown continuously: the Oxford Desk Reference recorded at least 70 loci and more than 30 cloned genes [4], Saudubray's textbook refers to approximately 80 HSP forms with various inheritance modes [7], and the 2025 GeneReviews review lists more than 20 genes associated with pure forms alone [2]. This heterogeneity coexists with marked intra- and interfamilial phenotypic variability, such that knowing the causal gene does not predict phenotype in an individual or family [2]. For the geneticist, the central tasks are confirming diagnosis (excluding acquired causes and treatable inborn errors), establishing inheritance mode, interpreting age-dependent penetrance and guiding counselling for at-risk relatives, often in families where other members have received incorrect diagnoses such as multiple sclerosis [4].

AP-4 adaptor complex deficiency illustrates how entities initially described as cerebral palsy came to be classified among the complicated HSPs. In 2009, Verkerk and colleagues described, in a consanguineous sibship followed for more than 20 years, five patients with congenital spastic tetraplegia, intellectual disability, reduced cerebral white matter and cerebellar atrophy, and identified by homozygosity mapping the splice donor site variant c.1137+1G>T in intron 14 of AP4M1 [32]. In 2011, Moreno-De-Luca and colleagues found by chromosomal microarray a homozygous deletion involving AP4E1 in a consanguineous family with spastic tetraplegia, microcephaly, epilepsy and profound intellectual disability, and proposed the designation “AP-4 deficiency syndrome” [33]; in the same year, Abou Jamra and colleagues described eight patients from three consanguineous families with variants in AP4S1, AP4B1 and AP4E1, with severe intellectual disability, absent speech, stereotypic laughter, hypotonia progressing to spastic paraplegia, microcephaly and short stature [34]. The four forms were designated SPG47 (AP4B1), SPG50 (AP4M1), SPG51 (AP4E1) and SPG52 (AP4S1); the 2020 international cohort, with 156 patients from 101 families, showed that severity and the main phenotypes are similarly distributed across the four genes, so the four forms are now treated as a single entity, AP-4-associated HSP [14, 15].

2. Molecular Genetics and Pathophysiology

2.1 Inheritance patterns and relative distribution

AD, AR and XLR inheritance patterns have been described for both pure and complicated forms [4]. In pure forms, AD inheritance is the most frequent, accounting for 70-80% of families in Northern Europe and North America (75-80% of all pure forms in the GeneReviews review), AR inheritance for 10-20% and XLR inheritance for less than 5% [4, 2]. Most complicated forms have AR inheritance, although AD and XLR forms exist, the latter notably including L1CAM-associated disorders [4]. Across all HSP, AR forms represent 25-30%, with a continually growing gene list, many described in a single family or individual, and increased frequency in populations with higher consanguinity rates [2].

Among pure AD forms, SPG4 (SPAST) is most frequent (approximately 40%); SPG3A (ATL1) is second, with 10-15% overall and more than 75% of early-onset pure AD forms; SPG30 (KIF1A) and SPG31 (REEP1) each account for approximately 5%; the remaining predominantly adult-onset pure AD forms each represent 1% or less [2]. Among AR forms, the most frequent in the general population are SPG5 (CYP7B1, 7.3% of AR-HSP and 3% of non-syndromic simplex cases), SPG7 (SPG7, approximately 5% of AR-HSP, 5-12% in some series) and SPG11 (SPG11, 3-5% of AR-HSP and 75% of individuals with pure or complicated HSP and a thin or absent corpus callosum on imaging) [2]. The Oxford Desk Reference adds that CYP7B1 mutations (SPG5A) are found in 5-10% of pure AR-HSP and that SPG11 and SPG15 are important causes of the complex AR form “HSP with a thin corpus callosum” [4]. The proportion attributed to SPG3A differs between current GeneReviews chapters: the 2025 overview retains 10-15% of AD-HSP, while the SPG3A chapter, drawing on a meta-analysis of 13,570 patients, estimates approximately 5% [2, 11, 35].

The 2024 French PNDS, which combines the French series of 1550 probands published in 2022 with international reviews and cohorts, gives similar but not identical estimates: AD inheritance accounts for more than 80% of individuals with a pure form; SPG4 accounts for more than 40% of pure AD-HSP and about 15% of sporadic cases; SPG3A for 5–10% of AD-HSP, with a higher frequency in populations of European origin (8%) than in Asian and American populations (4% each), and with onset before the age of 10 years in 40% of patients; SPG30 and SPG31 for about 5% each; SPG10 for about 3%; and the remaining predominantly adult-onset AD forms (ATP2B4, CPT1C, ERLIN2, HSPD1, NIPA1, REEP2, RTN2, SLC33A1, WASHC5, ZFYVE27) for less than 2% each, some known in a single French family [21]. Among the AR forms, SPG5A represents 3–8% of sporadic cases depending on the population, SPG7 about 5% of AR-HSP (almost always with a complex phenotype) and the remainder (SPG14, SPG21, SPG28, SPG30, SPG35, SPG48, SPG58, SPG62, SPG65, SPG71, SPG72, SPG73, SPG77) less than 1% of pure forms [21]. The protocol stresses that KIF1A, ATL1 and SPG7 may be transmitted in an AD or AR manner depending on the variant, which affects counselling (section 8.6) [21].

Source: PNDS, tables 1 and 2 [21].

2.2 Shared cellular mechanisms

HSP proteins converge on a limited number of cellular pathways: intracellular membrane shaping and trafficking, endoplasmic reticulum (ER) morphogenesis, microtubule dynamics and axonal transport, mitochondrial function, lipid metabolism, and endolysosomal and autophagic function [4, 24]. Spastin and atlastin-1 interact with each other and several other HSP proteins (including REEP1/SPG31) and participate in intracellular membrane shaping and trafficking [4, 8]. Most HSP, probably including all pure AD forms, is associated with neurodegeneration rather than altered neurodevelopment [4].

Saudubray's textbook includes HSP in the chapter on cellular trafficking disorders, grouping genes by mechanism: autophagy (SPAST and ABCD1, involved in fatty acid trafficking from lipid droplets to peroxisomes; SPG11/spatacsin; ZFYVE26/spastizin, SPG15 or Kjellin syndrome; TECPR2, SPG49; AP5Z1, SPG48), cytoskeleton (REEP1, SPG31; KIF1C, SPG58; KIF5A, SPG10; KIF1A, SPG30), Golgi disorders (TANGO2, SLC33A1/SPG42 and ATL1), vesicular trafficking (AP4B1, AP4E1, WASHC5/SPG8, NIPA1/SPG6, SPG21/MAST syndrome, SPART/Troyer syndrome, SPG20) and organelle interactions (seipin/BSCL2, REEP1 and atlastin-1, implicated at the interface between lipid droplets and fatty acid incorporation) [36]. It also notes that most of these forms are complex and late-onset, and that the same gene may appear in other phenotypic categories: spatacsin is associated with SPG11, Charcot–Marie–Tooth disease type 2 (CMT2) and juvenile amyotrophic lateral sclerosis type 5 (ALS5) [36].

A growing HSP subgroup comprises inborn errors of complex lipid metabolism, particularly defects in phospholipid and sphingolipid synthesis and remodelling: DDHD1 (SPG28), DDHD2 (SPG54), CYP2U1 (SPG56), PNPLA6/NTE (SPG39), FA2H (SPG35), GBA2 (SPG46), B4GALNT1 (SPG26), as well as PLA2G6, ELOVL4 and PCYT2 [7, 37, 38]. The first identified AD metabolic HSP form was associated with ALDH18A1 variants, encoding Δ1-pyrroline-5-carboxylate synthetase (P5CS), with hypocitrullinaemia [7]. SPG9 is associated with monoallelic (SPG9A, AD) or biallelic (SPG9B, AR) ALDH18A1 variants affecting specific residues; at least 50 patients have been described, approximately two-thirds with monoallelic variants, and a severity continuum proportional to P5CS loss of function has been proposed (SPG9A < SPG9B < AD cutis laxa type 3 ≤ AR cutis laxa type 3A) [39].

2.3 SPG4 — SPAST (spastin)

The SPAST gene (2p22.3; OMIM 604277) encodes spastin, a 616-amino-acid AAA family protein synthesised as two isoforms from two different initiation codons: M1 (68 kDa), detected at appreciable levels only in the spinal cord, and M87 (60 kDa), present in spinal cord and cerebral cortex [8]. Spastin is an ATP-dependent microtubule-severing enzyme regulating microtubule length, number and mobility; it preferentially recognises and severs polyglutamylated microtubules, with activity increasing with the number of glutamates per tubulin up to approximately eight and decreasing above this threshold, making it essential for biogenesis and maintenance of complex microtubule arrangements in axons, spindles and cilia [8, 40]. Four domains support these functions: the N-terminal domain participates in lipid metabolism (regulating lipid droplet size and distribution) and ER morphogenesis through interaction with atlastin-1 (SPG3A) and REEP1 (SPG31); the MIT (microtubule-interacting and trafficking) domain binds CHMP1 and IST1 proteins of the ESCRT-III complex, involving spastin in cytokinesis, nuclear envelope sealing and endosomal tubule recycling; the microtubule-binding domain and AAA cassette ensure ATP binding and hydrolysis [8].

The predominant mechanism is haploinsufficiency: most pathogenic variants introduce premature termination codons, subject to messenger RNA degradation through nonsense-mediated decay, and approximately 20% of patients have large deletions [8]. ClinGen dosage curation assigns SPAST a haploinsufficiency score of 3 (“sufficient evidence for dosage pathogenicity”) and triplosensitivity of 0, a formal basis for considering deletion of one or more exons pathogenic [40]. In Parodi and colleagues' cohort of 842 patients, 77% of pathogenic variants (missense, frameshift, splicing, nonsense and deletions) were in the AAA cassette, compromising microtubule-severing activity [41]. Alternative mechanisms have been proposed for some missense variants: constitutive microtubule binding with a dominant negative effect and altered organelle transport, and a gain-of-function effect demonstrated in a murine model overexpressing mutant human spastin [8]. Neuropathology shows distal degeneration of long spinal tracts with microglial reaction and, in one case studied by Parodi and colleagues, degeneration of spinocerebellar and spinocortical tracts and posterior columns, with numerous small-diameter fibres among large myelinated fibres in the corticospinal tract, suggestive of regeneration; extramotor tau pathology was observed in some, but not all, cases [8, 41].

Two laboratory considerations directly affect diagnosis. First, single- or multiexon deletions and duplications account for 20-25% of pathogenic SPAST variants: sequencing detects 75-80% and targeted deletion/duplication analysis (MLPA, quantitative PCR, long-range PCR or exon microarray) the remaining 20-25% [8, 42]. The Oxford Desk Reference emphasises that these deletions may escape sequencing-only strategies and that MLPA (multiplex ligation-dependent probe amplification) is typically required [4]. Second, SPAST undergoes alternative splicing with variable exon 4 inclusion; no pathogenic variants have been described in this exon, suggesting the isoform without exon 4 (transcript NM_199436.1) is the predominant functional form in the adult nervous system [8]. The variants c.131C>T (p.Ser44Leu; frequency 0.4% in controls) and c.134C>A (p.Pro45Gln), in trans with a pathogenic variant, are associated with earlier onset and a more severe phenotype, acting as modifiers [8].

2.4 SPG3A — ATL1 (atlastin-1)

ATL1 (14q22.1; OMIM 606439) encodes atlastin-1, a GTPase of the dynamin subclass with two transmembrane domains and a GTP-binding domain with catalytic activity, expressed predominantly in the pyramidal neurons giving rise to corticospinal tracts [11]. It is located mainly in the ER and Golgi complex, and also in axonal growth cones; it promotes GTP-dependent ER membrane fusion and formation of three-way junctions in the tubular ER network, and interacts with spastin [11, 43]. The proposed mechanism is gain of function or dominant negative: most missense variants cluster around the GTPase domain, reducing catalytic activity; atlastin-1 forms tetramers, so heterocomplexes of normal and mutant protein may interfere with tetramer activity; pathogenic variant expression alters ER connectivity; and experimental expression reduction compromises axonal elongation [11]. Sequencing detects approximately 99% of variants; a single exon deletion has been described [11]. Notable variants include c.715C>T (p.Arg239Cys), recurrent and possibly associated with a methylated CpG dinucleotide, c.470T>G (p.Leu157Trp) and c.1243C>T (p.Arg415Trp), the latter with incomplete penetrance [11].

2.5 SPG7 — paraplegin

SPG7 (16q24.3; OMIM 602783) encodes paraplegin, a catalytic component of the inner mitochondrial membrane m-AAA protease, ubiquitously expressed in adult and fetal tissues; in addition to the AAA domain, it contains a divalent metal ion-binding site characteristic of metal-dependent proteases [9]. Paraplegin provides protein quality control in a hetero-oligomeric complex with AFG3L2; inactivation of the paraplegin–AFG3L2 complex reduces complex I activity, and resulting mitochondrial dysfunction causes neuronal injury and altered axonal transport [9]. The complex also processes regulatory substrates, including maturation of mitoribosomal protein MRPL32, and degrades MCU and SMDT1/EMRE, so its loss dysregulates mitochondrial calcium uptake [9]. Muscle biopsies from severely affected patients may show ragged-red fibres and cytochrome c oxidase-deficient fibres [9, 27], and Saudubray's textbook includes SPG7 and AFG3L2 among mitochondrial quality control genes whose variants cause multiple mitochondrial DNA (mtDNA) deletions, with AR and AD inheritance [44]. The disease mechanism is loss of function [9].

SPG7 exemplifies “missing heritability”: genome sequencing studies identified non-coding or deep intronic variants undetected by exome sequencing, potentially explaining some patients with only one identified pathogenic variant [9]. The c.1529C>T (p.Ala510Val) variant is present on at least one allele in 58% of patients in the 241-case European cohort and is associated with later onset and ataxia at presentation [45].

2.6 SPG11 — spatacsin

SPG11 (15q21.1; OMIM 610844) encodes spatacsin, a conserved protein without significant similarity to other known proteins, expressed throughout the nervous system and particularly in the cerebellum, cerebral cortex, hippocampus and pineal gland [10, 30]. Spatacsin interacts with the SPG15 (spastizin) and SPG48 (AP5Z1) gene products, is an accessory protein of adaptor complex 5 (AP-5) and is required for lysosomal tubulation and recycling (autophagic lysosome reformation) [10]. Loss of function reduces lipid clearance from lysosomes, with ganglioside and cholesterol accumulation, altered membrane cholesterol and calcium homeostasis; spatacsin is also necessary for neurite formation and the GSK3β signalling pathway [10]. Most variants predict protein truncation (loss of function); the few missense variants are associated with milder phenotypes [10]. Sequencing detects approximately 81% of variants and deletion/duplication analysis approximately 19%; 10-20% of disease-associated variants are exon-sized or larger deletions or duplications [10].

2.7 SPG76 — CAPN1 (calpain 1)

CAPN1 (11q13.1; OMIM 114220) encodes the catalytic subunit of calpain 1 (µ-calpain), a calcium-regulated non-lysosomal cysteine protease widely expressed in the central nervous system, where it participates in synaptic plasticity and restructuring and axonal maturation and maintenance [12]. Through exome sequencing, Gan-Or and colleagues identified homozygous or compound heterozygous nonsense, missense, frameshift and splicing variants in nine individuals from three AR-HSP families; calpain 1 loss of function caused neuronal and axonal dysfunction and degeneration in C. elegans, locomotor defects with axonal abnormalities in Drosophila, and altered branchiomotor neuron migration with axonal network disorganisation in zebrafish [12]. Wang and colleagues described cerebellar ataxia with limb spasticity in four families and demonstrated mild ataxia due to altered cerebellar development in calpain 1 knockout mice: increased neuronal apoptosis from absence of calpain 1-mediated PHLPP1 phosphatase cleavage, inhibiting the Akt survival pathway in granule cells, with phenotype reversal using the indirect Akt activator bisperoxovanadate or crossing with PHLPP1-deficient mice [31]. In patients' fibroblasts, calpain 1 loss increases calpain 2 activation, causes autophagosome accumulation, inhibits Akt and ERK1/2 pathways and increases susceptibility to ER stress and apoptosis; the recurrent p.Tyr320Leufs*73 variant causes complete protein absence [12, 46]. In the largest published series, truncating variants were most frequent and missense variants were associated with earlier onset, suggesting an additional deleterious effect [46].

2.8 Genotype–phenotype correlations

An international cohort of 206 patients with SPG4 (seven centres, supplemented by literature cases) identified 136 distinct variants and two longitudinal trajectories: a severe, rapidly progressive subgroup enriched for de novo missense variants, with early developmental delay and rapid loss of walking; and a moderate, biphasic subgroup enriched for inherited truncating variants, with late but accelerated progression [47]. Classification of missense variants by residue essentiality within spastin's hexameric structure, combined with known modifiers (biallelic or modifying variants in trans), separated the two subgroups, and plasma neurofilament light was elevated in both (26 patients and 101 controls), particularly in severe early-onset disease [47]. This stratification remains prospectively unvalidated, useful for trial design but not individual prognosis.

2.9 SPG47, SPG50, SPG51 and SPG52 — AP-4 adaptor complex deficiency

Adaptor protein complexes (AP-1 to AP-5) are conserved heterotetramers that recognise sorting signals in the cytosolic tails of transmembrane proteins and organise the formation of transport vesicles; Saudubray's textbook includes AP4B1, AP4E1, AP4M1 and AP4S1 among the disorders of vesicular trafficking that cause childhood-onset complex spastic paraparesis [48]. AP-4 consists of the β4, ε, μ4 and σ4 subunits, encoded respectively by AP4B1 (1p13.2), AP4E1 (15q21.2), AP4M1 (7q22.1) and AP4S1 (14q12); it is expressed in all tissues, including the central nervous system, and is located in the trans-Golgi network, where it selects transmembrane proteins for export towards the cell periphery [14]. Loss of one subunit prevents assembly of a stable complex and leads to degradation of the other three, so that biallelic loss-of-function variants in any of the four genes produce the same molecular defect [14]. In neurons, polarised cells with long axons, AP-4 takes part in anterograde axonal transport; the autophagy protein ATG9A is the best-studied cargo and, in deficient cells, it accumulates in the trans-Golgi network and is missing from peripheral compartments [14, 49]. This abnormal distribution of ATG9A is the basis of the functional diagnostic assays and small-molecule screens described in sections 5.6 and 7.11.

The variant spectrum is dominated by loss of function. In the 2020 cohort, the more than 70 unique variants were mainly nonsense or frameshift variants, followed by canonical splice variants; about 80% of patients were homozygous and 20% compound heterozygous, and most variants were private [15]. According to the GeneReviews review, frameshift variants predominate in AP4B1 and AP4E1, nonsense variants represent more than one third in AP4M1 and AP4S1, splice variants about one quarter in AP4E1, AP4M1 and AP4S1, and copy-number variants about one quarter in AP4E1 and AP4S1 [14]. The table brings together the recurrent or historical variants, normalised according to HGVS on the MANE Select transcripts; for AP4B1 and AP4S1, the MANE transcripts (NM_001253852.3 and NM_001128126.3) differ from those used in the initial publications (NM_006594 and NM_007077), but the coding numbering of these variants is the same [50, 51].

No genotype–phenotype correlations have been described between the four genes or within each gene; severity and the main phenotypes were similarly distributed across the four subtypes [14, 15]. Spastic paraparesis was universal in the patients with biallelic variants described in the cohorts, and heterozygotes have no known neurological disease; the only allelic association described is that of heterozygous AP4E1 variants in people with persistent stuttering, with no other phenotype associated with AP4B1, AP4M1 or AP4S1 [14].

3. Clinical Spectrum and Manifestations by System

3.1 Common presentation: spastic paraparesis

The predominant manifestations of pure forms are progressive bilateral lower-limb spasticity, maximal in the hamstrings, quadriceps, adductors and triceps surae, and weakness, maximal in the iliopsoas, hamstrings and tibialis anterior [2]. Spasticity and weakness vary: some patients have spasticity without demonstrable weakness, while others have both in similar proportions; increased tone is disproportionately severe relative to weakness [2, 4]. Onset may occur in early childhood, with typically non-progressive manifestations resembling spastic diplegic cerebral palsy, or later in childhood or adulthood, with slowly progressive evolution; in AD families, 25% of cases are symptomatic at 5 years, and presentation may involve delayed motor milestones, clumsiness and difficulty with school physical education [2, 4]. Patients report walking difficulty, falls and trips due to impaired foot dorsiflexion, with wear of shoe toes; after several years, patients with progressive deterioration may reach a “functional plateau”, where deterioration proceeds at a rate similar to that attributable to ageing [2, 4].

Severity ranges from asymptomatic individuals (10-20%) to wheelchair-dependent individuals (10-20%), with confinement to bed very rare [4]. Fatigue, cold and alcohol worsen symptoms [4]. Upper-limb strength and dexterity are typically normal, and speech, chewing and swallowing are unaffected in pure forms [2]. On examination, tendon reflexes are increased, particularly in the lower limbs (99%); 7% of patients have reduced strength and hyperreflexia in the upper limbs; sustained clonus (more than 4 beats) is present in 45% and an extensor plantar response in 80%; foot deformity (pes cavus) occurs in 65% [4]. Signs are predominantly motor, but vibration sense is impaired in 40% of patients (dorsal column involvement), and mild distal amyotrophy may occur in longstanding cases [4]. A hypertonic bladder (urgency, frequency, hesitancy) affects fewer than 50% of patients across HSP, and men may have erectile dysfunction [4, 2]. Progressive gait disturbance combined with frank, symmetrical corticospinal impairment is the clinical hallmark of uncomplicated HSP [2].

The French PNDS adds semiological details useful for early recognition. Hypertonia predominates in the antigravity muscles (triceps surae, quadriceps and sometimes hip adductors), and early in the disease examination may show only brisk reflexes, an absent or unilateral Babinski sign and fatigable clonus, without weakness or spasticity on passive movement; at this stage spasticity is mainly dynamic, evident during walking and barely apparent on the examination couch [21]. The initial complaint of “stiff legs” often appears only with intense exertion or in winter, because cold worsens spasticity, and may become apparent after prolonged immobilisation, with a period of morning stiffness or after sitting for a long time; pelvic girdle weakness manifests as difficulty rising from a low chair or climbing stairs [21]. Triceps surae spasticity causes equinovarus and an arched foot with the appearance of pseudo-steppage and pseudo-pes cavus, which may wrongly suggest a neuromuscular disease; in children, the picture may be limited to equinovarus feet or toe walking, and some patients undergo Achilles tendon tenotomy before HSP is considered [21]. Secondary osteoarticular deformities include pseudo-pes cavus, claw toes, equinovarus and knee flexion or recurvatum, the latter dependent on the age at onset and sometimes a cause of overt postural instability, so the initial assessment should characterise tendon contractures and their reducibility [21]. Reduced vibration sense in the lower limbs, due to posterior column involvement, without proprioceptive ataxia, is not sufficient to classify the form as complex; in long-standing pure forms, alterations of executive functions and social cognition, slowed information processing and fatigue with motor fatigability are frequent, without meeting criteria for dementia but with an impact on working and social life [21]. In a cohort of patients with SPG4, sexual dysfunction affected about 25% of patients, with similar frequency in men and women, and is rarely reported spontaneously [21].

3.2 Complicated forms: accompanying signs

Signs defining complicated forms include amyotrophy (for example, Silver syndrome), cerebellar signs, dystonia, dementia, epilepsy, sensory neuropathy, optic atrophy, central retinal degeneration (Kjellin syndrome), ichthyosis (Sjögren–Larsson syndrome), skin pigmentation changes and adducted thumbs with intellectual impairment (MASA syndrome, mental retardation, aphasia, shuffling gait, adducted thumbs, associated with L1CAM) [4]. Discriminating extraneurological signs include cataract (SPG9A, SPG46), gastro-oesophageal reflux (SPG9A), hypogonadism and male infertility (SPG46) and basal ganglia calcifications (SPG56) [2, 10, 37]. In childhood, a Brazilian series of 35 patients recorded 12 pure and 23 complicated forms, with mean ages at onset of 2.9 and 1.6 years, respectively; intellectual impairment was the most frequent associated finding, and magnetic resonance imaging (MRI) showed non-specific white matter lesions, cerebellar atrophy, corpus callosum thinning and the “ears of the lynx” sign [53]. In a Japanese paediatric series of 37 patients, MRI was abnormal in 69% of complex and 5% of pure forms [54]. In SPG5A (CYP7B1), usually pure, posterior column involvement is more marked than in other pure forms and may reach loss of vibration sense with distal amyotrophy; cerebellar ataxia, when present, makes it complex, and loss of walking occurs after a median of about 20 years of disease [21].

3.3 SPG4 (SPAST)

The cardinal SPG4 feature is insidious progressive bilateral lower-limb spasticity with brisk reflexes, Achilles clonus and extensor plantar responses [8]. Sphincter disturbances occur in 77%, predominantly urinary urgency and incontinence; upper-limb hyperreflexia in 65%, with other upper-limb signs rare; reduced malleolar vibration sense (diminished but not absent) in 60%; and proximal lower-limb weakness in approximately 50% [8]. Age at onset ranges from childhood to the eighth decade and is bimodal, with a main peak before 10 years and a second, smaller peak between the third and fifth decades; age-at-onset differences between relatives with the same variant ranged from 27 to 69 years [8, 41]. Severity increases with disease duration: after 20 years, approximately 50% require walking assistance and approximately 10% a wheelchair; progression is faster in late-onset patients (after 35 years) [8]. Symptomatic women more frequently have upper-limb hyperreflexia, reflecting a more diffuse presentation despite lower penetrance [8, 41].

Lower-limb spasms are frequent, may precede spasticity, worsen after physical activity and tend to disappear when spasticity becomes severe [8]. Bladder dysfunction is among the most frequent problems: Schneider and colleagues recorded urological involvement in 91.2% of SPG4 patients, against 74.5% across HSP, with incontinence, hesitancy, frequency, urgency and incomplete emptying [8]. In a urodynamic study of 29 HSP patients, Fourtassi and colleagues described central neurogenic bladder in 82.7%, detrusor overactivity in 52% and detrusor–sphincter dyssynergia in 65.5%; anal sphincter involvement caused flatus incontinence in 31.4% and faecal incontinence in 8.7% in another series [8]. Pes cavus and mild spastic dysarthria may occur [8].

Non-motor symptoms are more frequent than previously recognised: in a study of 30 patients, fatigue, pain and depression were frequent and sometimes severe [8]. Subtle late-onset cognitive impairment has been documented, not always present in all affected family members, usually limited to executive dysfunction without daily-life impact; neuropsychological assessment of nine adults identified non-amnestic multidomain mild cognitive impairment, with executive and social cognition dysfunction; in the Parodi cohort, intellectual impairment was described in 4.2% [8]. Hand tremor was recorded in 10% of a Dutch cohort, restless legs syndrome was associated but not confirmed, and seizures, intellectual impairment and cerebellar ataxia are rare; some cases of severe dementia have been described [8]. Some people with SPAST variants develop lower motor neuron degeneration with an amyotrophic lateral sclerosis (ALS)-like phenotype [8]. MRI is frequently normal; spinal cord atrophy may occur, less pronounced than in other HSP, and mild vermian atrophy, a thin corpus callosum and subtle white matter changes have been described; advanced techniques (diffusion tensor imaging, resting-state functional MRI, magnetoencephalography) show diffuse grey and white matter involvement correlated with duration and severity [8].

The rare patients with biallelic SPAST variants may have severe intellectual disability [21].

3.4 SPG3A (ATL1)

SPG3A is characterised by bilateral, mostly symmetrical, slowly progressive lower-limb spasticity and weakness, with mean onset at 4 years; more than 80% of affected individuals have a spastic gait before 10 years [11]. Unlike other AD-HSP, where reduced vibration sense and bladder overactivity are widespread, these signs occur in a minority of SPG3A patients, as does upper-limb hyperreflexia [11]. Most early-onset cases have a pure phenotype; complicated phenotypes have been described with axonal motor neuropathy and/or distal amyotrophy with lower motor neuron involvement (Silver syndrome phenotype) and, in forms with spastic tetraparesis, dysphagia and dysarthria from bulbar involvement [11]. Pes cavus and scoliosis are frequent, possibly because of early onset, and wheelchair or walking-aid dependence is relatively rare [11]. Compared with SPG4, SPG3A has earlier onset, more lower-limb amyotrophy and scoliosis, fewer sphincter disturbances and less reduction of vibration sense [8].

The spectrum includes adult-onset forms with slower progression, whose existence contradicts the hypothesis of a purely neurodevelopmental disease [11]. It also includes hereditary sensory neuropathy type ID (HSN1D), an AD axonal form with marked sensory loss causing painless injuries [11, 43]. In the index family, trophic foot changes with painless ulcers led to amputation of toes, feet or more proximal leg segments in most affected adults; younger patients had increased tone and older patients patellar hyperreflexia, and nerve conduction studies showed axonal motor and sensory neuropathy, leading the authors to recommend ATL1 testing in hereditary sensory neuropathy type I with pyramidal signs [43]. Pure autonomic failure followed by spastic paraplegia (a splicing variant in exon 2) and a severe neonatal-onset cerebral palsy-like presentation with tetraparesis, pseudobulbar palsy and axonal polyneuropathy associated with a de novo variant have also been described [11].

3.5 SPG7

The phenotypic spectrum of SPG7-associated neurological disease includes uncomplicated spastic ataxia, complicated spastic ataxia, spinocerebellar ataxia and isolated optic atrophy [9]. Onset is usually in adulthood, with a mean of 35.5 ± 14.3 years, but ranges from childhood to 72 years; the first manifestation is insidious bilateral leg weakness due to spasticity, cerebellar signs (gait and limb ataxia, dysarthria, dysphagia) or both [9, 45]. In the international cohort of 241 patients, initial presentation was spasticity in 89, ataxia in 74 and both in 45 [45]. The following table summarises progression in 98 patients examined twice, with a median interval of 5.0 ± 5.9 years [9, 45].

Spasticity is progressive and may cause loss of independent walking; severe gait impairment develops 8-10 years after onset in approximately one-third of patients, and a very slowly progressive late-onset form (median 39 years; 18-52 years) associated with cerebellar ataxia (39%) or cerebellar atrophy (47%) has also been described [9]. Gait ataxia and dysarthria are the first cerebellar manifestations, with oculomotor abnormalities and dysphagia developing later [9]. Lower motor neuron involvement includes axonal motor and sensory neuropathy, neuropathic pain and amyotrophy, first in the lower limbs and later the hands; in rare variant carriers, fasciculations or mitochondrial encephalomyopathy manifestations (myalgia, fatigue, exercise intolerance, progressive external ophthalmoplegia) occurred in 14.2% [9]. Ophthalmological manifestations include optic neuropathy with reduced visual acuity, diplopia, ptosis, nystagmus and strabismus, which may occur without spasticity or ataxia; optical coherence tomography detects subclinical optic neuropathy [9]. Cognitive impairment is usually mild, with visuoconstructive and executive dysfunction, and frank dementia is rare; rare dystonia predominantly affects the lower limbs, with multifocal dystonia involving craniocervical regions described [9]. Scoliosis, pes cavus and mild conductive, sensorineural or mixed hearing loss may coexist, often detected only through active surveillance [9]. In Coarelli's cohort, patients with more than 20 years' disease had more cerebellar dysarthria, deep sensory loss, amyotrophy, ophthalmoplegia and sphincter dysfunction than those with less than 10 years, and the SARA (Scale for the Assessment and Rating of Ataxia) increased by a mean of 1.0 ± 1.4 points per year in a subgroup of 30 patients; neuropathology in one case showed a reduced pyramidal tract in the medulla and moderate loss of Purkinje cells and substantia nigra neurons [45]. Primary lateral sclerosis with biallelic SPG7 variants has also been described in one family, and heterozygous variants have been proposed as an ALS risk factor [9].

3.6 SPG11

SPG11 is characterised by progressive lower-limb spasticity and weakness, frequently associated with mild intellectual impairment with childhood learning difficulties and/or progressive cognitive decline, peripheral neuropathy, pseudobulbar involvement and upper-limb hyperreflexia [10]. Onset is predominantly in childhood or adolescence (1-31 years), rarely at 50-60 years, presenting with gait disturbance or, less frequently, intellectual impairment; the Oxford Desk Reference emphasises that learning difficulties and cognitive impairment may precede paraplegia, which typically begins in the second decade [4, 10]. Approximately ten years after onset, most patients show the full picture — progressive spasticity, corpus callosum atrophy with intellectual impairment and/or cognitive decline — and most become wheelchair-dependent one to two decades after onset [10].

Cognitive decline, present in most patients, worsens over time and includes severe short-term memory impairment, emotional lability, childish behaviour, reduced verbal fluency and impaired attention indicating executive dysfunction; psychiatric problems occur, and many patients show little concern about motor progression, correlated with frontal atrophy [10]. Axonal motor or sensory-motor peripheral neuropathy occurs in more than 80%, particularly after ten years' disease, and sural nerve biopsies show unmyelinated fibre loss and accumulation of pleomorphic membranous material [10]. Less frequent findings include cerebellar signs (ataxia, nystagmus, saccadic pursuit), retinal degeneration or macular excavation (Kjellin syndrome, initially associated with SPG15 but more frequent in SPG11), pes cavus, scoliosis, parkinsonism, epilepsy, orthostatic hypotension and, with longer disease duration, dysphagia; visual evoked potentials show increased latencies and reduced amplitudes [10]. The same variants may manifest as longstanding juvenile ALS (ALS5) or axonal CMT type 2X, regardless of variant nature, and SPG11 variants have been identified in patients diagnosed with multiple sclerosis [10, 36].

3.7 SPG76 (CAPN1)

SPG76 is a complicated form, best characterised as spastic ataxia, usually with young-adult onset [46]. Progressive lower-limb spasticity and weakness are associated, in most patients, with cerebellar ataxia with cerebellar atrophy, dysarthria and lower-limb weakness [46]. CAPN1 variants explained approximately 1.4% of a young-adult-onset spastic ataxia cohort at two European university hospitals (Paris and Stockholm), with 21 new patients and nine new variants [46]. The authors recommend specific assessment of cognitive impairment and depression during follow-up [46]. Cases with bulbar involvement and emotional disturbance [55] and one case of congenital-onset pure spastic paraplegia have been described [56].

MRI may mislead: in a retrospective analysis of 15 patients, 12 (80%) had bilateral periventricular white matter involvement, associated in 5 with multifocal subcortical abnormalities, and 3 (20%) had only multifocal subcortical involvement; in the proband with longitudinal studies, lesion number and periventricular involvement area increased over time [57]. In a patient with progressive spastic paraparesis, this pattern may incorrectly suggest acquired white matter disease, delaying diagnosis in the series' index case [57].

3.8 AP-4 deficiency (SPG47, SPG50, SPG51, SPG52)

AP-4-associated HSP is a complicated childhood-onset form, with neurodevelopmental and neurodegenerative components; no pure form has been described [14]. The 2020 cross-sectional cohort included 156 patients from 101 families, of diverse ethnic origins and aged 1.0 to 49.3 years; the mean age at symptom onset was 0.8 ± 0.6 years (range 0.2–5.0) and the mean age at diagnosis 10.2 ± 8.5 years (range 0.1–46.3), which corresponds to a mean diagnostic delay of about a decade [15]. The 2025 GeneReviews review combines these data with those of the registry and natural history study NCT04712812 (303 patients from 289 families, largely still unpublished), from which the following frequencies are drawn unless another source is given [14].

The picture begins in the first year of life, with mild hypotonia (92%, 244/264), global developmental delay and postnatal microcephaly. Motor delay is almost constant (99%, 277/281) and is often the first manifestation: median ages of acquisition were 12 months for sitting (interquartile range 9–18), 18 months for crawling (12.3–24), 24 months for standing unsupported (18–36) and 25 months for walking with support (20–36) [14]. Only about 45% (110/243) achieve independent walking, at a median age of 30 months, and this ability is often lost as the disease progresses [14]. Speech is severely impaired or absent (97%, 269/276); most communicate non-verbally and intellectual disability in older children is usually moderate to severe [14]. Microcephaly, generally 2 to 3 standard deviations below the mean for age and sex, was recorded in 67% (73/109) in the natural history study and in 83% in the 2020 cohort [14, 15].

Spastic paraparesis is universal. It typically begins between 4 and 6 years of age (85%, 256/301) and is present in 60% (47/78) of children younger than 4 years and in 96% (216/226) of those aged 4 years or older [14]. Spasticity first sets in at the ankles, often with an in-toeing gait, and progresses proximally; in the 2020 cohort, spastic diplegia appeared at a mean of 8.4 ± 5.1 years and tetraplegia at 16.1 ± 9.8 years, and at the last assessment 36% walked with assistance (mean age 8.9 ± 6.4 years) and 54% were wheelchair-dependent (mean age 13.4 ± 9.8 years) [15]. In adolescence or young adulthood, spasticity may involve the upper limbs, generally without relevant disability [14].

Seizures usually begin in the first 3 years of life, rarely in the neonatal period. About 60% (96/160) have a diagnosis of epilepsy and 70% (94/134), with or without epilepsy, have had seizures with fever; in the 2020 cohort, 37% had at least one episode of status epilepticus, often at the first seizure, and refractory epilepsy (persistent seizures despite two antiseizure medications) occurred in 27% [14, 15]. Seizures become less frequent with age and most respond to conventional antiseizure medications, except in patients with malformations of cortical development, such as polymicrogyria, who have more seizures and more episodes of status epilepticus; epilepsy, generally beginning before 3 years, was associated with a worse motor outcome [15, 16].

The remaining manifestations have variable frequencies: urinary incontinence (81%, 112/139) and faecal incontinence (72%, 105/146); foot deformities, mainly clubfoot (50%, 118/238; 69% in the 2020 cohort); episodes of stereotypic laughter, possibly a pseudobulbar affect (46%, 113/245; 56% in the 2020 cohort); reduced pain sensitivity (38%, 36/96), still without formal studies of sensory neuropathy; short stature (34%, 25/74); dysphagia in advanced stages (24%, 54/224), with gastrostomy feeding in 5% (4/88); dysarthria (21%, 30/140); and behavioural abnormalities, with attention deficit (33%, 100/301), hyperactivity (12%, 35/301) and impulsivity (12%, 37/301) [14, 15]. Movement disorders occur in 28% (83/301), mainly dystonia (14%) and ataxia (7%), and bradykinesia and rigidity (3% each) appear in late stages [14]. In the multicentre study of 428 children and young adults with early-onset HSP, dystonia was the pattern associated with AP-4 deficiency (as with SPG4 and SPG3A) and was associated with greater disability on the SPATAX and SPRS scales and with poorer quality of life [58]. Health-related quality of life, measured by the CPCHILD questionnaire (Caregiver Priorities and Child Health Index of Life with Disabilities) in 64 patients, was reduced in all domains (mean 59.6 ± 12.6), with no differences between the four subtypes, and correlated inversely with the SPRS (r = −0.52) and with age (r = −0.31); tetraplegia and wheelchair dependence were associated with lower scores, but seizures, including refractory epilepsy, were not [59].

Brain MRI has a characteristic pattern. In the systematic analysis of 107 scans from 76 patients, the diagnostic signature comprises four findings — a thin corpus callosum, with posterior predominance; a thin or absent anterior commissure; hyperintensity of the forceps minor around the frontal horns (the “ears of the grizzly” sign); and periventricular white matter changes —; the presence of two or more has a sensitivity of about 99% and the combination of all four occurs in about 45%; compared with other HSPs with a thin corpus callosum, absence of the anterior commissure appears specific to AP-4 deficiency [16]. In the 2020 cohort, a thin corpus callosum was present in 90%, ventriculomegaly in 65%, often with colpocephaly, and periventricular white matter changes in 68%; reduced white matter volumes correlated with motor severity [15, 16]. Less frequent are delayed myelination, bilateral perisylvian polymicrogyria, incomplete hippocampal inversion and the “ears of the lynx” sign, and six patients have been described with signs suggestive of iron deposition in the globus pallidus and substantia nigra (five with SPG50 and one with SPG52) [14]. In four registry patients with fetal or perinatal MRI (SPG47, SPG50 and SPG52), scans between 22 and 38 weeks showed ventriculomegaly, corpus callosum hypoplasia, reduced periventricular white matter and incomplete hippocampal inversion; the patient with SPG52 also had gyral immaturity and hypoplasia of the pons and vermis, and postnatal follow-up revealed progressive white matter loss and delayed myelination [60].

The plasma concentration of neurofilament light chain (NfL) was increased in 46 patients compared with 46 age- and sex-matched controls (area under the curve 0.87; 95% CI: 0.80–0.94), with higher values in the subgroup with severe generalised seizures and developmental stagnation; glial fibrillary acidic protein (GFAP) did not differ from controls (14 patients and 21 controls) [61]. Longitudinal natural history data are still being published; the oldest known patient is 55 years old, according to unpublished registry data [14].

4. Clinical Recognition and Red Flags

4.1 History and examination elements

History should include a three-generation family history, focusing on gait and mobility disturbances, and document age at onset, gait change pattern (shoe toe wear), falls, maximum walking distance on level ground and aid use, severity, aggravating factors (fatigue, cold, alcohol), bladder or bowel dysfunction and signs of complication [4]. Examination should include observation of gait and footwear, assessment of foot deformity, quantification of tone relative to strength, reflexes, clonus and plantar response, and vibration sense testing [4]. The SPRS (Spastic Paraplegia Rating Scale) is the reference instrument for quantifying severity [2] and has been translated and validated in Brazilian Portuguese [62].

When faced with spastic paraparesis, the strongest arguments for distinguishing a hereditary form from an acquired disease are the mode of onset, the course and the familial nature: insidious onset over years and continuous progression favour a neurodevelopmental or neurodegenerative disease, whereas onset over hours, days or weeks suggests an acquired cause, and a vascular or traumatic spinal cord lesion produces a sudden deficit that is fixed or even improves over time [21]. The absence of a family history is frequent and does not exclude HSP, because of incomplete penetrance, variable expressivity and de novo variants; an active search for relatives with gait disturbance, with their age at onset and course, should be completed by a pedigree [21].

4.2 Criteria for a firm clinical diagnosis

The Oxford Desk Reference stipulates that a firm HSP diagnosis should be made only when at least two family members have definite clinical signs — progressive spastic gait disturbance with frank lower-limb corticospinal signs, namely hyperreflexia with bilateral extensor plantar responses or bilateral sustained clonus (5 or more beats) — or when a proven pathogenic variant has been identified [4]. Without a causal variant in the family, HSP is a diagnosis of exclusion [4]. When assessing at-risk relatives, anxiety may cause mildly increased tone, hyperreflexia and unsustained clonus, subtle signs that should not be overinterpreted [4].

The French PNDS does not require two affected relatives: it retains the clinical diagnosis of spastic paraparesis in the presence of a lower-limb pyramidal syndrome with progressive spasticity and motor deficit, after exclusion of acquired causes, and treats molecular confirmation as the next step [21]. This divergence from the Oxford Desk Reference criterion is mainly one of framing: the two-relative criterion serves a firm diagnosis without genetic testing, whereas the PNDS addresses a pathway in which genomic testing is accessible. The same protocol considers complex an HSP associated with ataxia, epilepsy, intellectual disability, inaugural or severe cognitive impairment, optic neuropathy, an extrapyramidal syndrome or peripheral neuropathy, but not one with only reduced vibration sense or bladder and sphincter disturbances [21].

4.3 Red flags for specific forms or alternative diagnoses

5. Diagnosis: Criteria, Tests and Algorithms

5.1 Exclusion and characterisation investigations

Investigations to consider, according to the Oxford Desk Reference, include: DNA for a gene panel, exome sequencing (ES) or genome sequencing (GS), given genetic heterogeneity; very-long-chain fatty acids for adrenoleukodystrophy if inheritance is compatible with XLR; a levodopa trial in cases without sensory symptoms; brain and spinal cord MRI in one affected member of each family and particularly in sporadic or complex cases; and leucocyte enzymes when atypical signs such as significant peripheral neuropathy are present [4]. In pure AD-HSP, MRI may show spinal cord atrophy, particularly cervical and thoracic, while some AR forms are associated with corpus callosum thinning [4]. Electromyography (EMG) with nerve conduction velocities excludes peripheral nervous system involvement; in SPG4, markedly prolonged central motor conduction time argues against diagnosis [8].

Saudubray's textbook proposes an approach targeting metabolic causes of acute myelopathy or spastic paraparesis without leukoencephalopathy, classifying them into four groups [7]:

*Disorders with available specific treatment, according to the textbook. HHH syndrome denotes hyperammonaemia, hyperornithinaemia and homocitrullinuria.

The same textbook identifies three signs suggesting an inborn error of metabolism in a patient with spastic paraparesis: polyneuropathy on EMG, leukoencephalopathy on MRI, and an acute or subacute course with sensory ataxia suggesting subacute combined degeneration of the spinal cord [7]. Some of these disorders may begin with isolated spastic paraparesis lasting years, notably X-linked adrenoleukodystrophy, homocysteine remethylation defects, biotinidase deficiency, HHH syndrome, arginase deficiency, Segawa disease and some lipid metabolism defects [63]. Dopamine synthesis defects (GTP cyclohydrolase and tyrosine hydroxylase deficiencies) may produce lower-limb dystonia mimicking spastic paraparesis and responding to levodopa [7]. The textbook also highlights the ataxia–spasticity spectrum: Purkinje cells, spinocerebellar and corticospinal tracts are frequently affected by the same genes, ataxia may precede spasticity or develop simultaneously, and the concept of “spastic-ataxic” disease has replaced the classical separation between SPG and AR ataxias [63].

The French PNDS proposes a more extensive and systematic set of investigations, intended to exclude acquired forms, define the pure or complex character and identify forms with urgent specific treatment [21]:

Compared with the GeneReviews and Oxford Desk Reference algorithm (section 5.4), the PNDS makes systematic MRI in all patients (and not only in one affected member per family), the metabolic work-up with oxysterols, cholestanol and hexosaminidases, audiometry, optical coherence tomography and EMG, and does not include vitamin E in the list; vitamin E remains relevant when ataxia is associated [2, 4, 21]. For dopa-responsive dystonias (GTP cyclohydrolase 1 and tyrosine hydroxylase deficiencies), the protocol recommends a systematic dopaminergic therapeutic trial in any patient with a compatible phenotype, especially with worsening of hypertonia during the day, without waiting for the genetic result, and allows analysis of neurotransmitters in cerebrospinal fluid; the response then guides interpretation of the genetic test [21]. The Oxford Desk Reference criterion, which reserves the trial for patients without sensory signs, is more restrictive [4].

5.2 Genetic tests

The molecular approach combines targeted testing (multigene panel) and comprehensive genomic testing (ES, GS) [2]. A panel including genes for pure AD and AR forms is more likely to identify the cause, limiting identification of variants of uncertain significance (VUS) in genes not explaining the phenotype; for SPG4 and SPG11, inclusion of deletion/duplication analysis in the panel is recommended [2, 8, 10]. Genomic tests are preferred when the phenotype is indistinguishable from many other disorders with spastic paraparesis; if ES is non-diagnostic and inheritance appears AD, an exon array may detect single- or multiexon deletions or duplications [8]. In SPG7, single-gene testing is rarely useful, and GS should be considered whenever only one pathogenic variant is identified, given the role of deep intronic variants [9]. In population-based studies, the proportion of families without a genetic diagnosis after systematic testing ranged from 45% to 67% in AD-HSP and 71% to 82% in AR-HSP [18], and in the Portuguese survey the genetic cause was not identified in 39.7% of patients with ataxia or HSP [19]. Recent reviews emphasise increased diagnostic yield with long-read sequencing [22].

The French PNDS considers next-generation sequencing (panel, exome or genome) the first-line test and reserves targeted single-gene testing for families with an already identified variant, given that pure forms cannot be distinguished clinically from one another [21]. In the series cited, the yield of panels and clinical exomes reaches 50% or more and that of the genome about 60%, so close to half of patients with a pure form remain undiagnosed [21]. Genomic rearrangements, particularly of SPAST, REEP1 and SPG11, account for 6% of all patients and 10–25% of SPAST variants, and the protocol recommends searching for them with techniques based on high-throughput sequencing; the difference from the classical MLPA recommendation (section 5.5) is one of method, not of objective, and either route requires validated detection of copy-number variants [4, 8, 21]. Without a diagnosis after NGS, reanalysis of the data is relevant from 1 year onwards, and DNA from patients with no identified cause should be stored [21]. HSPs are a pre-indication of the France Genomic Medicine Plan 2025 (PFMG 2025), with access to genome sequencing according to criteria: no genome in a patient without a family history and with onset after 50 years; with onset before 50 years, with or without a family history, genome sequencing at least as a duo (as a trio, if possible) after measurement of the biomarkers of AR forms and a panel including at least the 10 most frequent genes [21]. These access criteria reflect French organisation and should not be read as evidence of the absence of a genetic cause in patients with late onset.

5.3 Subtype-specific criteria

SPG4. Should be suspected with insidious progressive bilateral leg stiffness, with or without resting spasticity, mild proximal weakness and frequently urinary urgency; examination showing bilateral corticospinal impairment; mildly reduced malleolar vibration sense in most; and an AD-compatible family history, or exclusion of other causes in simplex cases [8]. Signs of a complicated form reduce likelihood but do not exclude SPG4 [8]. Diagnosis is established with a heterozygous pathogenic SPAST variant; failure to detect a variant or deletion does not absolutely exclude diagnosis [8].

SPG3A. Should be suspected with onset between childhood and 10 years (mean 4 years), progressive symmetrical lower-limb spasticity and weakness, reduced vibration sense, bladder overactivity and an AD-compatible family history; absence of family history does not exclude diagnosis [11]. A heterozygous ATL1 VUS neither establishes nor excludes diagnosis [11].

SPG7. There are no consensus criteria; diagnosis should be considered with mean onset at 35.5 years, pure or complicated spastic paraparesis (deep sensory loss, pes cavus, amyotrophy, cerebellar dysarthria, dysphagia, palatal tremor, subtle cognitive impairment), cerebellar ataxia, ophthalmological abnormalities (optic atrophy with or without spasticity, progressive external ophthalmoplegia, ptosis, nystagmus, strabismus) and MRI showing cerebellar atrophy or diffusion tensor white matter changes in frontal lobes, corticospinal tracts and brainstem; it is established with biallelic pathogenic or likely pathogenic variants [9].

SPG11. Suggested by the clinical combination described in section 3.6 with MRI showing corpus callosum thinning in more than 90% of patients, particularly with long T1 and T2 times in the forceps minor (“ears of the lynx” sign, hyperintense on FLAIR and hypointense on T1), frequent cortical atrophy, white matter hyperintensities (initially frontal and occipital periventricular, subsequently confluent and progressive), and sometimes brainstem and cerebellar atrophy and basal ganglia changes [10]. Sixty per cent of individuals with a thin corpus callosum, cognitive impairment and spastic paraparesis had biallelic SPG11 variants [10]. Diagnosis is established with biallelic pathogenic variants [10].

SPG76. Diagnosis is molecular (biallelic CAPN1 variants); neurological examination, brain and spinal MRI (cerebellar atrophy, periventricular and subcortical white matter pattern), neurophysiological studies and neuropsychological and mood assessment characterise the phenotype without specificity [46, 57].

5.4 Practical algorithm

The diagnostic pathway can be organised into four steps. The first confirms the lower-limb upper motor neuron syndrome and classifies it as pure or complicated, collecting a three-generation family history and examining parents, because the frequency of asymptomatic carriers requires examining parents of apparently sporadic or recessive cases [4]. The second excludes acquired and treatable causes: brain and whole spinal cord MRI, vitamin B12, vitamin E, copper, serology for syphilis, human T-cell lymphotropic virus type 1 (HTLV-1) and human immunodeficiency virus (HIV), very-long-chain fatty acids and, according to red flags, a levodopa trial, plasma amino acids, total homocysteine, ammonia and sterol studies [2, 4, 7]. The third chooses genetic testing: an HSP panel with deletion/duplication analysis or ES/GS according to phenotype and inheritance pattern, with SPAST MLPA if the panel does not include deletions and inheritance is AD [2, 8]. The fourth interprets results using the five-class ACMG/AMP system (pathogenic, likely pathogenic, VUS, likely benign, benign), recalling that “pathogenic” and “likely pathogenic” are equivalent for clinical decisions and that familial segregation and parental studies are indispensable in AR forms [9, 10]. DNA storage for future reanalysis should be considered in probands without a molecular diagnosis [2].

5.5 Genetic test limitations and report content

Before accepting a negative result, test coverage should be checked. Exon deletions and duplications account for 20–25% of SPAST and approximately 19% of SPG11 variants, and a panel or exome without validated copy-number variant analysis may miss them [4, 8, 10]. Exome sequencing does not detect deep intronic or most regulatory variants, particularly relevant in SPG7 with only one identified variant, where genome sequencing is indicated [9]. None of these tests reliably detects repeat expansions entering the differential diagnosis — the FXN GAA expansion in Friedreich ataxia with retained reflexes and polyglutamine expansions in spinocerebellar ataxias — requiring targeted tests [4, 9]. Low-level parental mosaicism may escape leucocyte DNA testing (documented in SPG4 and SPG11), incomplete age-dependent penetrance prevents considering a clinically normal parent a non-carrier without testing, and SPG7 pseudodominance may mimic dominant inheritance [8–10].

The report should state the gene, versioned reference transcript, HGVS coding DNA and protein nomenclature, zygosity (and, in recessive forms, variant phase confirmed through parental studies), ACMG/AMP classification with applied criteria and technical limitations [64]. MANE Select transcripts (v1.5) for the genes detailed in this article are NM_014946.4 for SPAST, NM_015915.5 for ATL1, NM_003119.4 for SPG7, NM_025137.4 for SPG11 and NM_005186.4 for CAPN1 [50]. The SPAST MANE transcript is the long isoform including exon 4, whereas some literature refers to the isoform without exon 4 (NM_199436.1), requiring exon numbering confirmation before comparing variants [8, 50]. On the MANE transcript, cited variants are written NM_014946.4:c.131C>T p.(Ser44Leu) and c.134C>A p.(Pro45Gln) in SPAST — severity modifiers when in trans, classified in ClinVar as benign/likely benign/risk factor and with conflicting classifications, respectively, and which should not be reported as the disease cause [8, 65]; NM_015915.5:c.715C>T p.(Arg239Cys), c.470T>G p.(Leu157Trp) and c.1243C>T p.(Arg415Trp) in ATL1, pathogenic or likely pathogenic in ClinVar, the last with approximately 30% penetrance [11, 66]; NM_003119.4:c.1529C>T p.(Ala510Val) in SPG7, pathogenic/likely pathogenic in ClinVar, present on at least one allele in 58% of patients in the European cohort and with a phenotype dependent on the allele in trans [45, 67]; and NM_005186.4:c.853C>T p.(Arg285*) in CAPN1, pathogenic/likely pathogenic [55, 68]. For the remaining CAPN1 variants in section 10.2, the authors' transcript was not confirmed in this review, and “c.213dupG” should be written c.213dup under HGVS. A variant of uncertain significance does not become pathogenic through therapeutic response or partial segregation — incomplete penetrance and variable expressivity make cosegregation across few meioses poorly informative — and should not underpin predictive testing, prenatal diagnosis or preimplantation testing [2, 11, 64].

5.6 AP-4 deficiency: diagnosis

No published consensus clinical criteria exist. The diagnosis is established in a patient with suggestive findings — global developmental delay, infantile hypotonia, postnatal microcephaly, seizures (including febrile seizures), progressive spastic paraparesis and incontinence, with the MRI described in section 3.8 and a family history compatible with autosomal recessive inheritance — and biallelic pathogenic or likely pathogenic variants in one of the four genes [14]. Two biallelic VUS, or one pathogenic variant and one VUS, neither establish nor exclude the diagnosis [14]. The absence of consanguinity does not exclude the disease: in the registry, 46% (142/311) of patients had consanguineous parents [14].

Molecular testing may be a multigene panel including the four genes and the differentials in section 6.5, or exome or genome sequencing, recommended by the ACMG as a first- or second-tier test in children with intellectual disability or multiple congenital anomalies; most variants lie in coding regions and are detectable by exome sequencing [14]. The table summarises the contribution of each gene and method.

Source: GeneReviews, table 1, with registry data updated in January 2025 [14]. The proportions in the table refer to known pathogenic variants and are not equivalent to the clinical sensitivity of the test. The text of the review itself attributes about one quarter of variants in AP4E1 and AP4S1 to copy-number variants, which does not match the table for AP4S1; when a single allele is identified or there is a typical phenotype with a negative exome, copy-number analysis should be requested (MLPA, microarray or validated read-depth analysis), since the first AP4E1 deletion was detected by microarray [14, 33]. Splice variants represent about one quarter of variants in AP4E1, AP4M1 and AP4S1, which justifies RNA studies when classification depends on the effect on processing [14]. The report should state the MANE transcript with its version; for AP4B1 and AP4S1, the MANE transcript is not the one used in the initial publications (section 2.9).

For variants of uncertain significance, quantification of the subcellular distribution of ATG9A in patient fibroblasts (ATG9A ratio between the trans-Golgi network and the cytoplasm) has been used as a functional assay, initially in a research setting [14]. In a series of eight patients with suspected AP-4 deficiency, the assay demonstrated loss of AP-4 function in six and allowed novel variants to be reclassified; the series is small and does not establish sensitivity or specificity [69]. Plasma NfL distinguishes patients from controls in group studies, but is not an individual diagnostic test [61]. The combination of two or more findings of the MRI signature has a sensitivity of about 99% for AP-4 deficiency; specificity was not quantified in the abstract consulted, so MRI guides, but does not replace, molecular testing [16].

When both familial variants have been identified, carrier testing, prenatal diagnosis and preimplantation genetic testing are possible [14]. Fetal MRI may show ventriculomegaly, corpus callosum hypoplasia and reduced white matter from 22 weeks, but these findings are not specific and, without a molecular diagnosis, do not allow prenatal diagnosis [60].

6. Differential Diagnosis

6.1 Acquired and non-genetic causes

Non-genetic causes to consider include space-occupying lesions and other brain or spinal structural abnormalities (tethered cord, cord compression), arteriovenous vascular malformations and fibrocartilaginous embolism, primary progressive multiple sclerosis and neuromyelitis optica, paraneoplastic myelopathies (including those associated with anti-GAD65 antibodies), nutritional deficiencies (vitamin B12, vitamin E, copper) and infections (syphilis, HTLV-1-associated myelopathy, also termed tropical spastic paraparesis, and AIDS myelopathy) [4, 2]. Diplegic cerebral palsy with corresponding MRI abnormalities is another differential diagnosis, although several patients presumed to have cerebral palsy have had pathogenic variants in HSP genes [2]. In simplex cases, all possible causes of lower-limb spasticity should be considered because several non-genetic causes are more frequent than SPG4 [8]. The French PNDS adds functional neurological disorders to this list and, for diplegic cerebral palsy, highlights the features that distinguish it from early-onset HSP: a suggestive obstetric history (prematurity, anoxia, difficult delivery), compatible lesions on MRI and a fixed, very early-onset and non-progressive deficit [21].

6.2 Treatable genetic causes

Among treatable genetic disorders, GeneReviews highlights: dopa-responsive dystonia (GTP cyclohydrolase 1 and tyrosine hydroxylase deficiencies), to be considered in everyone with progressive gait disturbance, particularly children; MTHFR deficiency homocystinuria; cobalamin C disease; urea cycle defects (arginase deficiency, HHH syndrome); biotinidase deficiency; phenylalanine hydroxylase deficiency; glycine encephalopathy; FOLR1-associated cerebral folate transport deficiency; cerebrotendinous xanthomatosis; and purine nucleoside phosphorylase deficiency [2]. In dopa-responsive dystonia, levodopa response is marked and sustained [4]. Adrenomyeloneuropathy may mimic HSP and should be considered in families without male-to-male transmission; normal brain MRI does not exclude this disorder [4]. In SPG5 (CYP7B1), accumulation of neurotoxic oxysterols underpinned trials of cholesterol-lowering drugs (atorvastatin, simvastatin, ezetimibe), which consistently reduced 27-hydroxycholesterol, a validated biomarker, without demonstrated motor or functional benefit to date [22].

In arginase 1 deficiency, the differential diagnosis with HSP is difficult; progression of spasticity, spontaneous avoidance of protein-rich foods and the rarity of a hypertonic bladder point towards argininaemia, and seizures with electroencephalographic abnormalities, microcephaly and cerebral atrophy are frequent [70].

6.3 Other genetic disorders

6.4 Differential diagnosis within HSP

For SPG4, the main pure AD alternatives are SPG3A (earlier onset, more amyotrophy and scoliosis, fewer sphincter abnormalities), SPG10 (KIF5A; more peripheral neuropathy, amyotrophy or parkinsonism) and SPG31 (REEP1; more peripheral neuropathy or amyotrophy, the third most frequent AD form) [8]. For SPG3A, the differential includes SPG10 (frequent axonal motor neuropathy, probably the second cause of early-onset AD-HSP), SPG6 (NIPA1; occasionally childhood onset, probably the fastest-progressing AD form), SPG72 (REEP2; onset before 4 years and mild postural tremor), SPG42 (SLC33A1, a gene whose biallelic variants cause Huppke-Brendel syndrome, with congenital cataracts, deafness and low serum copper and caeruloplasmin), SPG4 and SPG12 (RTN2; usual onset before 10 years), as well as diplegic or tetraplegic cerebral palsy [11]. For SPG7, there are no significant differences from other pure AD and AR forms; polyglutamine expansion SCAs (SCA1, SCA2, SCA3, SCA7, SCA17, DRPLA) usually begin before 30 years, progress faster and involve the brainstem on MRI, while SCA6 presents with slowly progressive adult ataxia and gaze-evoked nystagmus [9].

For SPG11, clinical differentiation from other HSP with a thin corpus callosum and cognitive impairment is frequently difficult [10]:

When amyotrophy is marked, lower motor neuron degeneration in SPG11 may mimic ALS, and typical ALS lesions have been found in the brains of SPG11 patients [10]. For SPG76, the differential includes other young-adult spastic ataxias and, on imaging, acquired white matter disorders [46, 57].

6.5 Differential diagnosis of AP-4 deficiency

The initial manifestations of AP-4 deficiency are not very specific, and many children receive a diagnosis of cerebral palsy before genetic testing [14, 16]. The table brings together the genetic differentials proposed by the GeneReviews review, with the features that distinguish them and the test that discriminates between them.

Source of the genetic rows: GeneReviews, table 3 [14]. The review also includes leukodystrophies and inborn errors of metabolism, highlighting treatable forms such as dopa-responsive dystonia [14]. In practice, exome or genome sequencing resolves most of this differential in a single test; MRI, with attention to the anterior commissure, the corpus callosum and the frontal horns, guides variant interpretation [16].

7. Treatments and Risk Management Strategies

7.1 General principles

No treatment prevents or reverses axonal degeneration in HSP; supportive care aims to improve quality of life, maximise function and reduce complications, ideally within a multidisciplinary team including general medicine, neurology, clinical genetics, physical and rehabilitation medicine, physiotherapy, occupational therapy, speech therapy, urology, orthopaedics, social services and psychology [2, 8]. No professional society clinical practice guidelines have been published for pure HSP, SPG4, SPG7, SPG11 or SPG3A; GeneReviews recommendations rest on the authors' experience and Spastic Paraplegia Foundation information [2, 9]. The only national guidance document identified is the Haute Autorité de Santé PNDS of September 2024 for pure HSP, prepared by the French neurogenetics CRMR according to HAS methodology, based on a literature review and consensus of a multidisciplinary group; its recommendations and divergences are set out in sections 7.12 to 7.16 [21]. None of the antispastic drugs has an approved indication specifically naming HSP; their use rests on generic spasticity indications in the respective labels [71–74].

7.2 Assessments after diagnosis

7.3 Physiotherapy and rehabilitation

Individualised physiotherapy includes stretching to improve flexibility, reduce spasticity, maintain joint range and prevent contractures; aerobic exercise for cardiovascular capacity, strength, coordination and balance; and strengthening exercises for posture, walking and walking-aid use [2]. In SPG4, stretching should be manual at all levels (hip, knee, ankle) and preceded by heat application; the authors note that early intensive physiotherapy delays spasticity-related symptoms and prolongs walking, although physiotherapy efficacy in HSP is documented only in case reports and uncontrolled studies [8]. Physical measures include massage, ultrasound, electrical stimulation and hydromassage; anodal transcutaneous spinal direct current stimulation (2 mA, 20 minutes, 5 days, over T10–T12), in a randomised, crossover, sham-controlled trial of 11 patients (8 with SPG4), reduced Ashworth scale spasticity for up to two months, without differences in the 5-minute walk test or SPRS [2, 75]. In a randomised Brazilian pilot trial of 16 patients, virtual reality vestibular rehabilitation (balance games, with or without strengthening) mainly improved the visual contrast domain of the physiological fall-risk profile, without a non-intervention control group, so the authors consider it only a potential adjunct [76]. A randomised crossover trial of strength and functional exercise supervised by videoconference is ongoing in Porto Alegre (HSPMOVE; SPRS, Timed Up and Go and goal attainment scale outcomes) [77]. Ankle–foot orthoses attenuate dorsiflexion weakness, and gait, transfer and falling technique training reduces injury risk [2, 11]. In SPG7, balance training, strengthening and occupational therapy maintain function but do not slow progression; weight control is recommended because obesity worsens mobility [9]. The role of surgical hamstring and Achilles tendon lengthening and adductor longus release remains undefined but should be considered if contractures develop [2].

7.4 Pharmacological treatment of spasticity

GeneReviews recommends baclofen, botulinum toxin, dantrolene and tizanidine, used one at a time, particularly early in disease, to reduce cramps, decrease leg muscle stiffness and facilitate walking; baclofen may be tried first and, in some patients, administered through an intrathecal pump; drugs may be taken at bedtime when nocturnal cramps are the main problem or otherwise three to four times daily, with effects apparent after a few days [2]. In SPG3A, GeneReviews proposes a sequence: oral baclofen or tizanidine; chemodenervation with botulinum toxin A or B if oral drugs are not tolerated; an intrathecal baclofen pump for patients improving on oral baclofen but experiencing significant systemic effects [11]. Dantrolene and diazepam are also mentioned in SPG7 [9]. Dantrolene should be avoided in ambulatory patients because it may induce irreversible weakness impairing mobility [2, 11]. A double-blind crossover gabapentin trial showed no improvement in SPG4 spasticity [8].

Table sources: FDA labels for oral baclofen [71], oral baclofen solution and granule formulations [78, 79], tizanidine [72], intrathecal baclofen [73, 80], onabotulinumtoxinA [74] and abobotulinumtoxinA [81]. In the European Union and Portugal, oral and intrathecal baclofen, tizanidine and botulinum toxins are authorised through national or decentralised procedures, with indications and dosages defined in each country's summary of product characteristics (SmPC), which should be consulted before prescribing.

7.5 HSP-specific evidence

Controlled evidence in HSP is scarce. In SPASTOX, a randomised, double-blind, placebo-controlled crossover trial, 55 patients (41 with a pure phenotype; mean age 43 years; mean disease duration 17 years) received one botulinum toxin A session (100 IU/2 mL Prosigne® into each adductor magnus and each triceps surae) or saline; maximum walking speed (primary endpoint) and secondary endpoints did not differ, except for reduced adductor tone, and adverse effects were transient and similar in both arms [82]. This regimen is outside the indication of the product used and cannot be transferred in units to other toxins. In an open-label study of 33 adults, botulinum toxin A reduced adductor spasticity and fatigue without functional motor improvement [83]. In a prospective open-label study of intrathecal baclofen in 16 HSP patients refractory to oral drugs, screening was positive in 14, who underwent implantation; after mean follow-up of 25.8 months, the modified Ashworth scale decreased from 2.6 ± 0.8 to 0.7 ± 0.9 and the functional walking scale increased from 5.9 ± 1.7 to 7.4 ± 2.0; two patients underwent further surgery for catheter fracture [84]. The same Brazilian SPASTOX group conducted ZISPAST, a randomised, placebo-controlled, double-blind crossover trial of topical Alpinia zerumbet oil (Ziclague®, a topical antispastic marketed in Brazil): in 57 patients (88% with a pure phenotype; mean disease duration 21 years), 0.8 mL applications to adductor magnus and triceps surae did not change self-selected walking speed (primary endpoint) or secondary endpoints, with mild adverse effects similar to placebo [85].

Dalfampridine (4-aminopyridine), a potassium channel blocker approved to improve walking in multiple sclerosis, is off-label in HSP. In an open-label study of 12 patients receiving 10 mg twice daily for 2 weeks, SPRS and MSWS-12 improved without serious adverse effects [86]; in a randomised, placebo-controlled, triple-blind pilot trial of 8 patients treated for 8 weeks with 10 mg twice daily plus physiotherapy, the 25-foot walk test (T25FWT) improved more in the active group, without recorded adverse effects [87]. These small-sample data do not establish efficacy.

7.6 Bladder and bowel dysfunction and other manifestations

Spastic bladder dysfunction (urgency, frequency, voiding difficulty, incontinence) may be treated with antimuscarinics such as oxybutynin (Ditropan XL®) and solifenacin (Vesicare®), or the β3 agonist mirabegron (Myrbetriq®), and intravesical botulinum toxin injections; urodynamic studies should be performed early in all patients with urgency, voiding difficulty, retention or recurrent urinary infections [8, 9, 11]. The doses of these drugs according to their labels are given in the table in section 7.14. Constipation and faecal incontinence are treated with stool softeners; dysphagia with food consistency modification, head positioning and swallowing exercises, and in SPG7 with aspiration risk, consideration of gastrostomy; dysarthria with speech therapy and, if necessary, augmentative and alternative communication [2, 9]. Scoliosis requires orthopaedic surveillance (orthosis, possible surgery), and pes cavus may benefit from physiotherapy, orthoses, botulinum toxin or corrective surgery [2]. In SPG7, optic atrophy requires corrective lenses, prisms and low-vision services [9]. SPG11 and SPG7 psychiatric manifestations are treated according to usual practice [9, 10]. In the ATL1 spectrum with sensory neuropathy, protecting insensate feet (inspection, footwear, early ulcer treatment) is the priority, given the natural history of painless ulcers, osteomyelitis and amputation [43].

7.7 Functional adaptation and risk management

With insidious progression, disability may have advanced substantially since the last review, requiring periodic reassessment of driving (HSP may compromise pedal use; patients should inform the traffic authority of the diagnosis and may need adaptation to hand controls), home adaptations (step-free shower, rails, stairlift, ground-floor bedroom) and psychological and social support [4, 2]. In bedridden patients with advanced disease, aspiration pneumonia, urinary infection and pulmonary embolism risks should be monitored [9]. In pregnancy, regional anaesthesia (epidural or spinal) was traditionally avoided because of a theoretical risk of worsening weakness, but several successful uses in women with HSP have been described [11]. The French PNDS sets this reservation aside: it considers that epidural anaesthesia is not contraindicated and that spinal cord anatomy does not increase the risk of locoregional anaesthesia (section 7.15) [21].

7.8 Investigational therapies

In SPG11, miglustat (a glucosylceramide synthase inhibitor) was tested on the basis of lysosomal glycosphingolipid accumulation and biochemical and locomotor improvement in zebrafish and fly models: in TreatSPG11 (phase 2, open-label, non-randomised, single-centre; 100 mg three times/day for 4 weeks and 200 mg three times/day for 8 weeks), the drug was tolerated without serious adverse effects but did not change neuromotor function or plasma sphingolipid and ganglioside profiles [88]. Miglustat (Zavesca®) is not authorised for this indication. Naringenin and SMER28, acting on lysosomal reformation, corrected SPG11 and SPG15 phenotypes in laboratory models [89]. In SPG50 (AP4M1), AAV9 vector gene therapy administered intrathecally to one patient (phase 1 trial) showed no serious adverse effects or neurotoxicity over 12 months, with effective immunosuppression and suggested clinical stabilisation [22, 52]. In SPG76, therapeutic evidence is exclusively preclinical: in calpain 1-deficient mice, the indirect Akt activator bisperoxovanadate prevented cerebellar granule cell apoptosis and restored motor coordination, without human trials [31]. In SPG50, the programme progressed to a US phase 1–2 trial (NCT05518188) and an open-label phase 3 trial with a concurrent prospective control group (NCT06692712; recruitment started 01/04/2026; primary endpoint: GMFM-88-defined motor milestones at 156 weeks) [17]. In SPG4, AAV9 “silencing and replacement” gene therapy (microRNA against endogenous SPAST and complementary DNA of normal M1 and M87 isoforms), administered intracerebroventricularly to newborn mice carrying human spastin p.Cys448Tyr, prevented corticospinal degeneration and gait abnormalities; this is exclusively preclinical evidence [90]. In a complex form caused by a de novo SPTSSA variant, D-cycloserine was tested in a single-patient protocol completed in 2026, without published results at consultation [91]. None of these approaches should be presented to families as available treatment.

7.9 Clinical trials and limits of indirect comparison

There are no pivotal registration trials for any HSP therapy; the table brings together controlled studies and the main open-label studies cited in this article.

These studies' results cannot be compared with each other or with label trials. Trials supporting antispastic regulatory indications were conducted in other disorders (multiple sclerosis, spinal injury, stroke, cerebral palsy), mainly using tone endpoints, whereas HSP trials used walking endpoints, where reduced tone did not translate into functional gain (SPASTOX, tsDCS) [71, 72, 75, 82]. Studies with favourable results are open-label or include fewer than 15 patients, and mean scale improvement (Ashworth, SPRS) does not indicate the proportion achieving clinically relevant benefit; in crossover trials, a single session per period and carryover effects limit extrapolation to prolonged treatment [82, 86, 87]. Negative trials (botulinum toxin for walking, gabapentin, Ziclague®, miglustat) are as informative as positive trials and should not be omitted when discussing expectations with patients [82, 85, 88, 92]. Slow natural SPRS progression — approximately 0.9 points per year in the Austrian cohort — makes disease-modification trials dependent on long follow-up or more sensitive endpoints [93].

7.10 Comparative regulatory status and evidence hierarchy

The evidence hierarchy should be explicit: a peer-reviewed article, a trial registry, a sponsor statement, an EMA Committee for Medicinal Products for Human Use (CHMP) opinion and an authorisation decision represent different levels, and authorisation does not guarantee marketing, reimbursement or access in each country — in Portugal these depend on INFARMED and in Brazil on ANVISA registration and incorporation into the SUS. No medicine has a specifically named HSP indication in either jurisdiction; the table summarises the position on 6 October 2026.

Distinguishing extrapolation from off-label use is essential. Oral baclofen or tizanidine use in HSP extrapolates evidence obtained in other causes of spasticity but lies within generic label wording; toxin injection into thigh adductors, dalfampridine, miglustat in SPG11 and statins in SPG5 are off-label uses defined by label wording, age and jurisdiction, requiring documented decisions and informed consent [22, 74, 86, 88]. The SPASTOX toxin regimen (Prosigne®) cannot be transferred in units to other toxins [82].

7.11 AP-4 deficiency: management and investigational therapies

There are no clinical practice guidelines for AP-4 deficiency; the recommendations of the GeneReviews review are based on the authors' experience, and no treatment prevents, halts or reverses neurological progression [14]. After diagnosis, the initial evaluation includes neurological examination (tone, pyramidal signs, seizures), assessment of motor, adaptive and cognitive development with referral for early intervention or special education, speech therapy with consideration of augmentative and alternative communication, ophthalmological examination, nutritional and swallowing assessment with aspiration risk, respiratory assessment, neuro-urological assessment with urodynamic study, orthopaedic and physiatric assessment (mobility, contractures, scoliosis and foot deformities), pain assessment in patients with contractures, behavioural assessment (self-injury, attention-deficit/hyperactivity disorder and autism spectrum disorder), genetic counselling and assessment of family and social needs [14].

Treatment is supportive and multidisciplinary. Spasticity is treated with physiotherapy, oral antispastic drugs, botulinum toxin, intrathecal baclofen and orthopaedic surgery, according to the principles and doses in section 7.4, which include the paediatric doses of botulinum toxin (FDA indication for spasticity from 2 years of age) and intrathecal screening in young children; intrathecal baclofen has no established safety below 4 years [73, 74]. Regular physiotherapy together with antispastic drugs may delay the progression of contractures, scoliosis and foot deformities [14]. Epilepsy responds in most patients to conventional antiseizure medications, with no drug of choice defined for this disease; parental education about the presentation of seizures and a plan for prolonged seizures are relevant, given that status epilepticus is frequent at the first seizure [14, 15]. Augmentative communication should be introduced early; strabismus may be corrected surgically; dysphagia and aspiration risk may justify gastrostomy, which also secures a route for medication; sialorrhoea is treated with anticholinergics, botulinum toxin or surgery; constipation and gastro-oesophageal reflux with laxatives, prokinetics and proton pump inhibitors, with fundoplication considered in refractory cases; urinary urgency with anticholinergics and urological follow-up; and osteopenia with vitamin D and calcium supplements [14]. Behavioural abnormalities benefit from guidance by developmental paediatrics, including medicines for attention-deficit/hyperactivity disorder when indicated, and the family should have access to social services, care coordination, home care and, when appropriate, palliative care [14]. The doses of these supportive medicines follow the paediatric label of each one; there are no doses specific to AP-4 deficiency.

Gene therapy is the most advanced line of research. In the preclinical studies of the AAV9/AP4M1 vector, intrathecal injection in Ap4m1-KO mice with 1.25 × 10¹¹ to 5 × 10¹¹ vector genomes (vg) at 7–10 or at 90 days of life had age- and dose-dependent effects, with greater benefit from early intervention and from higher doses; toxicology studies in mice, rats and non-human primates showed an acceptable safety profile up to the target human dose of 1 × 10¹⁵ vg, with minimal to mild dorsal root ganglion toxicity in rats and primates [99]. The same dose was given intrathecally to the single patient described in section 10.3, and other patients in Canada and Spain received the vector in expanded access programmes [14, 52]. For SPG47, administration of AAV9/hAP4B1 into the cisterna magna of mice corrected ATG9A distribution, the calbindin-positive spheroids of the deep cerebellar nuclei, the anatomical abnormalities and the motor dysfunction, and normalised plasma NfL, and the toxicology study in non-human primates showed no significant adverse effects [100]. In a high-throughput screen of 28,864 small molecules, the compound BCH-HSP-C01 restored ATG9A distribution in patient fibroblasts and in neurons derived from induced pluripotent stem cells, and a CRISPR/Cas9 loss-of-function screen in a neuronal model identified regulators of ATG9A trafficking; both studies are preclinical [49, 101].

None of these therapies has a marketing authorisation in the USA or in the European Union, and access outside a trial is limited to expanded access programmes [14, 17]. The trials are open-label and small, and the phase 3 trial compares treated patients with non-randomised controls, some with other subtypes of AP-4 deficiency; the evidence that the four subtypes share the same natural history supports this choice, but does not remove selection bias [15, 17]. The preclinical data of greater efficacy with early treatment explain the age restriction of the trials and reinforce the value of early diagnosis; conversely, dorsal root ganglion toxicity, the immunosuppression required and the slow and variable progression of the disease call for prolonged follow-up before any conclusion about efficacy [52, 99].

7.12 Spasticity, pain and fatigue: PNDS recommendations and divergences

The French PNDS does not propose systematic treatment of spasticity: treatment is justified only when spasticity limits transfers, walking, grasping, hygiene, dressing or elimination, or when it causes pain [21]. The neurologist may start the antispastic drug, but severe or complicated spasticity should be referred to a physical and rehabilitation medicine (PRM) clinic, where impairments, activity limitations and participation restrictions are described according to the International Classification of Functioning, in order to agree realistic goals with the patient [21]. Oral baclofen is first line, especially in diffuse spasticity; in case of contraindication, dantrolene or tizanidine are proposed, the latter available in France only through compassionate access [21]. Any antispastic drug should be introduced very gradually, aiming at the minimum effective dose, for two reasons: lower-limb spasticity sometimes compensates for pelvic girdle weakness and its abrupt reduction may worsen walking, and baclofen is sedative and often poorly tolerated [21].

Botulinum toxin is described as an effective and well-tolerated treatment, but usually as second line, because access is more difficult; it is particularly indicated in focal spasticity (spastic equinovarus that hinders walking, finger flexors that hinder grasping) or when oral drugs cause adverse effects, and it may be applied to several muscle groups (hip adductors, quadriceps, among others) according to the goals [21]. The effect appears on average 2 weeks after the injection and lasts 3 to 6 months, in adults and children, so effective injections are repeated 2 to 4 times a year; a prior motor block allows the functional gain to be estimated and avoids post-injection weakness detrimental to walking [21]. Because the maximum dose per session limits the number of muscles treated, the toxin is often combined with systemic treatment, alcohol neurolysis of nerve trunks, tenotomies or neurotomies; injections should be guided by ultrasound or electrical stimulation, and anaesthetic cream (lidocaine/prilocaine) and an equimolar mixture of oxygen and nitrous oxide should be offered to children and adults who wish it [21]. In patients with painful contractures or spasms refractory to conventional treatments, the protocol recommends referral for a trial of intrathecal baclofen by pump [21]. Medical cannabis is under evaluation in France in other neurological diseases with spastic paraparesis (mainly multiple sclerosis), for refractory neuropathic pain and painful spasticity; it cannot be prescribed for HSP, the data on neuropathic pain are heterogeneous and psychiatric adverse effects have been described [21].

Lower-limb pain results mainly from spasticity and its musculoskeletal consequences, but many patients describe it in neuropathic terms (tingling, burning), attributed to involvement of the central sensory pathways; in these cases the PNDS favours pregabalin and gabapentin and recommends using amitriptyline sparingly, because of its anticholinergic effect on micturition [21]. The physiotherapy sheet appended to the protocol adds non-pharmacological measures for neuropathic pain (low-frequency transcutaneous electrical nerve stimulation, sensory hyperstimulation, neurodynamic stretching, mirror therapy and motor imagery, cognitive behavioural therapy) and for mechanical low back pain secondary to the altered gait (strengthening of the lumbar girdle and joint mobility) [21]. Fatigue has no specific treatment; workplace adaptations, therapeutic part-time work and technical aids improve quality of life [21].

The PNDS does not give doses. For the drugs it introduces or whose place in the therapeutic sequence it changes, the table presents the doses of the FDA labels; for pregabalin and gabapentin, central neuropathic pain in HSP is not an approved indication, and the doses are extrapolated, respectively, from neuropathic pain associated with spinal cord injury and from postherpetic neuralgia. The doses of baclofen, tizanidine, intrathecal baclofen and the toxins for spasticity are given in section 7.4.

Divergences between the PNDS and the other sources. First, dantrolene: GeneReviews recommends avoiding it in patients who walk, because of the risk of irreversible weakness, whereas the PNDS places it as an alternative to baclofen in case of contraindication; the GeneReviews reservation and the hepatotoxicity warning in the label justify limiting it to non-ambulant patients or to those with a defined functional goal, and to 45-day trials with liver monitoring [2, 11, 21, 102]. Second, tizanidine: the FDA label and GeneReviews place it alongside baclofen, and the secondary place it has in the PNDS reflects mainly its availability in France [2, 21, 72]. Third, botulinum toxin: the PNDS describes it as “frequently effective”, but the only controlled trial of botulinum toxin in HSP (SPASTOX) showed no gain in walking speed (section 7.5), so the goals should be focal (tone, pain, hygiene, orthosis) and not overall improvement of walking [21, 82]; the interval of 2 to 4 injections per year is compatible with the minimum of 12 weeks between sessions in the label [74]. Fourth, gabapentin: the crossover trial in SPG4 showed no effect on spasticity, and its place in the PNDS is neuropathic-type pain, not spasticity [21, 92].

7.13 Rehabilitation, physical activity and surgery according to the PNDS

The PNDS states that physical rehabilitation is the only factor with a demonstrated favourable impact on the course of the disease [21]. This wording is stronger than the available evidence: the SPG4 GeneReviews indicates that the efficacy of physiotherapy is documented only in case reports and uncontrolled studies, and the rehabilitation trials cited in section 7.3 are small [8, 75, 76]. The practical recommendation coincides — continuous physiotherapy and regular physical activity — but the expectation to convey is one of functional preservation, not of demonstrated modification of progression.

Physical activity should be prescribed, and not merely advised, to avoid deconditioning, worsening of spasticity and excess weight, with the aim of reaching the World Health Organization recommendations; muscle strengthening is not contraindicated [21]. The prescription follows an initial assessment of impairments and limitations and the levels of the HAS framework (adapted physical activity of level 2, autonomous or supervised physical or sporting activity of levels 3 and 4) and may include aerobic endurance, strengthening, stretching and balance work; in France, adapted physical activity has been prescribable since March 2017 for patients with a long-term condition, by health professionals or graduates in sport sciences, but is not reimbursed, and the BRAIN-TEAM filière provides free home exercise tools (BrainMouv) [21]. The good practice sheet recommends physical exercise for 20 minutes to 1 hour a day, 4 to 5 times a week, with an example session alternating three times 1 minute of weight-bearing triceps surae stretching and 30 seconds of strengthening of the dorsiflexors through the full range, followed by 20 to 30 minutes of functional gait training [21].

The physiotherapy prescription should be detailed (goals, targets, frequency and limits, such as pain), because physiotherapists rarely know the disease; the usual frequency is 1 to 2 sessions a week, adjusted to the capacity for self-exercise and the degree of dependence — the more dependent the patient, the more frequent and continuous the sessions — and the duration should respect fatigability [21]. The appended prescription template proposes ideally 2 weekly sessions, a prescription for 6 months renewable once and an increase in frequency from 15 days after each botulinum toxin injection and up to 2 months afterwards; the follow-up section allows 1 to 3 weekly sessions, so the protocol itself presents a range that varies with the context [21]. Each session includes stretching and teaching of self-stretching; the work extends to targeted strengthening, balance, transfers, standing (which prevents equinus from becoming fixed), reconditioning to effort, falls prevention and learning to rise independently from the floor; balneotherapy helps with painful contractures, but does not replace the dry-land session and may be precluded by incontinence; on the day of a toxin injection, the session may be held as usual [21].

The good practice sheet distinguishes predominantly overactive muscles (adductors, hamstrings, recti femoris, triceps surae) from predominantly paretic muscles (psoas, gluteus maximus and medius, foot dorsiflexors), identifies cold, nociceptive stimuli (“irritative thorns”) and stress as extrinsic factors that worsen spasticity, and proposes stretching of the spastic muscles always combined with active work, analytical strengthening of the paretic muscles, task-specific training, functional electrical stimulation of the paretic muscles and balneotherapy [21]. For nocturnal cramps and spasms, it advises avoiding prolonged immobility and performing at bedtime several isometric contractions of 6 seconds with an equal period of relaxation [21]. Therapeutic goals should be SMART (specific, measurable, adapted, achievable and time-bound), for example a further 0.10 m in step length, a further 50 m in the 6-minute walk test, a further 2 seconds of single-leg stance, a further 10° of knee extension through hamstring stretching or a further 2 kg of dorsiflexor strength on the dynamometer [21].

Source: good practice sheet for physiotherapists appended to the PNDS [21].

The PRM physician also coordinates podiatry (nail cutting and removal of corns and calluses, ideally every 1 to 2 months, generally not reimbursed), walking aids (adapted footwear, ankle stabilisation orthoses, a stick or crutches depending on strength and the upper limbs, a 2-, 3- or 4-wheeled walker or one with forearm support when there is imbalance or fatigability, a manual wheelchair by simple prescription and a powered wheelchair after multidisciplinary assessment), occupational therapy (adaptation of the home, school and workplace), custom-made orthopaedic footwear, orthoses and braces for scoliosis, rehabilitation or respite stays and, in children, special education services and home care [21].

Surgery is proposed as a last resort, when medical treatment and rehabilitation neither correct nor stabilise the deformities, with the indication discussed in a multidisciplinary meeting (orthopaedic surgeon, neurosurgeon, PRM physician, neurologist, physiotherapist and, if possible, carers) and with an explicit and achievable goal — barefoot walking, giving up an orthosis, resolution of pain, better positioning in the wheelchair or easier transfers —, distinguishing the symptoms due to the deformity from those that surgery will not improve [21]. The most commonly operated deformities are equinovarus (rarely valgus) and claw toes; the principles are selective neurotomy (section of 50 to 75% of the motor branches of the spastic muscle), tendon lengthening for irreducible contractures, simple tenotomy of a contracted muscle with no functional usefulness, tendon transfers to rebalance forces and, rarely, joint stabilisation; tendon procedures are preferably performed percutaneously and under local anaesthesia [21]. After isolated neurotomy, weight-bearing is immediate; after tendon surgery, a walking cast is generally used for 5 weeks; after bone or joint surgery, immobilisation without weight-bearing for 2 months; immobilisation should be as short as possible, because it worsens spasticity, and should be followed by mobilisation guided by a physiotherapist [21]. GeneReviews considers the role of surgical lengthening still to be defined (section 7.3); the PNDS does not contradict this reservation, but defines the conditions in which surgery should be considered [2, 21].

7.14 Bladder, bowel and sexuality: the PNDS approach

The treatments for neurogenic bladder in acquired diseases (multiple sclerosis, spinal cord injury) apply to HSP with the same modalities; in the cohorts, 50 to 80% of patients have detrusor overactivity with a post-void residual, but other patterns exist, so treatment should be guided by a physician experienced in bladder and sphincter disorders and, ideally, by a urodynamic study [21]. In detrusor overactivity, antimuscarinics are first line, with the drawback of worsening constipation; transcutaneous posterior tibial neuromodulation (adhesive electrodes on the inner aspect of the ankle, in sessions whose frequency and duration are set by the specialist) is the best alternative and may be used first or second line, for urinary and anorectal symptoms; alpha-blockers may be tried and are first line in detrusor-sphincter dyssynergia [21]. In nocturia, desmopressin may be added at the lowest effective dose, with monitoring of serum sodium, initially close and later spaced out; in France it is not authorised for this use after the age of 65 years [21]. Intradetrusor botulinum toxin injections have efficacy and safety in HSP similar to those in acquired diseases; before them, the ability to self-catheterise should be checked (Pencil Paper Test), followed by therapeutic education in neuro-urology [21]. Intermittent self-catheterisation is the technique of choice in dyssynergia after failure of alpha-blockers and in the underactive bladder, exceptional in HSP; sphincterotomy and urinary diversion require joint assessment by PRM, urology and neurology, and permanent drainage (indwelling or suprapubic catheter) is palliative, with a silicone catheter changed monthly [21].

Bowel disturbances are rare in HSP and arise mainly from reduced mobility; they are treated as in the general population, with fibre, hydration, abdominal massage, standing and osmotic or bulk laxatives, suppositories of potassium bitartrate with sodium bicarbonate in terminal dyschezia and, if they persist, transanal irrigation with therapeutic education; Malone surgery and stoma formation are exceptional [21]. Erectile dysfunction is treated with phosphodiesterase 5 inhibitors after a satisfactory cardiovascular assessment or with intracavernosal papaverine injections, started under the supervision of a trained physician (PRM or urology); female sexual dysfunction may benefit from local treatments, such as lubricating gel [21].

The PNDS does not give doses; the table presents those of the FDA labels. Tamsulosin is approved only for benign prostatic hyperplasia and oral desmopressin only for central diabetes insipidus and primary nocturnal enuresis, so their use in bladder dysfunction in HSP is an extrapolation (primary nocturnal enuresis being the reference for the night-time dose).

The PNDS does not mention mirabegron, which GeneReviews includes among the options (section 7.6); in practice, β3 agonists are an alternative without anticholinergic effects on constipation and cognition, while tibial neuromodulation, absent from GeneReviews, is the non-pharmacological alternative favoured by the French protocol [2, 11, 21].

7.15 Pregnancy, anaesthesia and vaccination

Fertility is not affected in pure forms, and contraception follows the criteria of the general population [21]. A preconception consultation is recommended, ideally with the partner, to discuss the interaction between HSP and pregnancy, physical access to care and to the delivery room, the capacity to care for the newborn and the management of medicines; support for disability in parenting may begin with the plan for pregnancy, through parenting support services for people with disabilities (in France, the SAPPH) [21]. Pregnancy does not appear to alter the long-term course, but weight gain and the discontinuation of symptomatic treatments may temporarily reduce independence, and bladder and sphincter disturbances may appear or worsen, generally transiently [21].

Vaginal delivery is possible in the great majority of cases, even with limited or absent expulsive efforts, with the use of instrumentation; epidural analgesia is not contraindicated and has no adverse effects of its own, and caesarean section should be anticipated if sphincter disturbances or severe spasticity justify it [21]. Since prolonged immobilisation worsens spasticity, the pregnant woman should avoid it and, postpartum, mobilisation and physiotherapy should resume as early as possible; breastfeeding should be planned according to the medicines to be restarted, and the return home prepared with exercises simulating newborn care (bathing, nappy changing, transfers, dressing) and environmental adaptations [21].

Regarding medicines, the PNDS refers to the Centre de Référence sur les Agents Tératogènes (CRAT) and details three situations [21]. Baclofen is not teratogenic in the rat and first-trimester data show no concerning signals, but it crosses the placenta, with neonatal concentrations close to maternal ones, and may cause neonatal withdrawal syndrome (irritability, hypertonia, tremor); it may be continued until pregnancy is confirmed, is in principle withdrawn until the 10th week (the end of organogenesis) and, near term, is either tapered gradually, if the disease allows, or continued with close monitoring of the newborn [21]. Among urinary antispasmodics, for which data in pregnancy are scarce, oxybutynin is preferred if an anticholinergic is needed and tamsulosin if an alpha-blocker is needed, ideally after 10 weeks of gestation (tamsulosin, only during pregnancy) [21]. Botulinum toxin should be avoided whenever possible and considered case by case, because the available information is limited to isolated cases [21]. The FDA label for baclofen also mentions neonatal withdrawal (section 7.4), and that for pregabalin describes a possible small increase in major malformations in observational studies and advises against breastfeeding [71, 103].

The PNDS identifies no contraindication or specific precautions for general or locoregional anaesthesia in pure forms [21]. Suxamethonium (suxamethonium), a depolarising neuromuscular blocking agent with a risk of hyperkalaemia, contraindicated in neuromuscular diseases, is by extension considered contraindicated in HSP, and non-depolarising blocking agents warrant monitoring because of the risk of an exaggerated response [21]. This position differs from the traditional reservation about regional anaesthesia reported by the SPG3A GeneReviews (section 7.7) [11].

No vaccine is contraindicated; the general population schedule is followed, with annual influenza vaccination if there is respiratory involvement [21]. As pure forms are not usually life-threatening and do not alter the capacity to express one's wishes, advance directives and the designation of a trusted person follow the general rules [21].

7.16 Transition, psychological support and carers

The transition from paediatric to adult care should be prepared from the age of 16 years, with follow-up shared between paediatric and adult teams between 16 and 18 years, in separate or joint consultations; when there is no direct link between the paediatric neurologist and the adult neurologist, or after an interruption in follow-up, child-to-adult transition units offer a multidisciplinary assessment and the link to the new team [21]. The aim is the young person's medical, social and occupational autonomy: knowledge of the disease, of warning signs, of treatments and of available support, and the choice of a career path; the team should be attentive to relationships with peers and family, to the conflict between the wish for autonomy and functional capacities or parental apprehension, to isolation and to fear of rejection, and should see the young person alone in consultation as soon as appropriate [21].

Psychological support should be offered systematically to patients and carers, from the disclosure of the diagnosis, in individual, couple or family interviews, with attention to times of fragility: disclosure, the young person's move to independence, educational and occupational choices, plans for children, deterioration, the move to a wheelchair and medical decisions; depressive and anxiety disorders are more frequent than in the general population [21]. Follow-up should identify the main carer, look for signs of physical or psychological deterioration, assess whether the available resources (technical, medico-social and financial support) match the needs and refer to psychologists, social workers, patient associations or peer-support groups; when there is a risk of burnout, respite options should be offered (at home, temporary admission to a medico-social facility, rehabilitation-respite stays, respite platforms) and information given about carer's leave and carer training [21].

8. Reproductive Options and Family Counselling

8.1 Determining inheritance mode

Counselling depends on identification of the genetic cause; several pure HSP genes are associated with both AD and AR inheritance (for example, ATL1, ERLIN2, REEP2, ALDH18A1), and in a patient with pure HSP without affected relatives or molecular diagnosis, another cause (primary lateral sclerosis or acquired disease) cannot be excluded and inheritance mode cannot be established [2]. In pure HSP, inheritance is AD in 70-80% of families, AR in 10-20% and XLR in less than 5% [4]. The frequency of asymptomatic carriers requires examination of parents of apparently sporadic or recessive cases; if both are normal, recessive inheritance is most likely but not certain [4]. When both parents of affected siblings cannot be examined, empirical values indicate a 1/6 probability of one parent being affected, conferring a 1/12 risk on the children of affected siblings [4]. No empirical values exist for offspring of apparently sporadic cases, where AR inheritance, a non-penetrant AD parent and a de novo variant must be considered [4].

8.2 Autosomal dominant forms (SPG4, SPG3A)

Most AD-HSP patients have an affected parent; the proportion of de novo variant cases is unknown across pure AD forms and low in SPG4 [2, 8]. Each child of a carrier has a 50% risk of inheriting the variant; in AD families, because 10-20% of carriers are asymptomatic, the risk of symptomatic disease in a carrier's offspring is 40-45% [4, 2]. If the proband's variant is undetected in either parent's leucocyte DNA, sibling recurrence risk is estimated at approximately 1% because of possible gonadal mosaicism, already described in SPG4 [2, 8]. If parents are clinically normal but untested, sibling risk is low but above population risk because of possible reduced penetrance or mosaicism [2].

In SPG4, penetrance is age-dependent and almost complete: 85% at 45 years and complete at 70 years, with approximately 6% of carriers asymptomatic on examination and approximately 20% having signs without awareness of disease; overall penetrance was 0.9 in the 842-patient cohort, lower in women (0.88 versus 0.94 in men), with greater sex discordance for onset before the third decade (91% versus 70%) [8, 41]. In SPG3A, more than 95% of patients have an affected parent; penetrance is high (approximately 80-90%) and does not appear significantly age-dependent, but p.Arg415Trp has only 30% penetrance and, because unaffected women predominated, erroneously suggested X-linked inheritance [11]. Marked intrafamilial variability prevents predicting onset age, course or severity in offspring [8].

For clinically normal children of affected individuals in AD families with onset predominantly before 35 years, the Oxford Desk Reference provides age-dependent residual risk of having inherited the gene [4]:

Based on the type I cumulative age-at-onset frequency curve, Harding estimated that an asymptomatic child of an affected parent has approximately a 20% probability of developing disease at 25 years, probably less if examination is normal [26].

8.3 Autosomal recessive forms (SPG7, SPG11, SPG76)

A proband's parents are presumed heterozygotes; molecular testing of both is recommended to confirm variant phase and allow reliable risk assessment [2]. If only one parent is a carrier, one variant may have arisen de novo in the proband or a mosaic parent, as described in one SPG11 case [10]; if the proband appears homozygous, an undetected deletion on sequencing and uniparental isodisomy must be excluded [2]. At conception, each proband's sibling has a 25% probability of being affected, 50% of being a carrier and 25% of inheriting neither variant; an affected individual's offspring are obligate carriers, and each sibling of the proband's parents has a 50% carrier risk [2, 10]. Heterozygotes are typically asymptomatic; in SPG11, they may occasionally have fundus abnormalities [10].

SPG7 raises specific problems. Several studies described patients with only one pathogenic variant, suggesting AD inheritance, but this hypothesis is controversial; alternative explanations include pseudodominance — described in a family segregating three SPG7 variants and in two families with p.Ala510Val homozygous patients in different generations — and “missing heritability” from non-coding variants [9]. A family has been described in which the compound heterozygous proband inherited one variant from an affected heterozygous father and another from an unaffected mother [2]. Heterozygotes may show subtly reduced corpus callosum white matter integrity on diffusion tensor imaging but are not typically at risk of disease [9]. SPG7 intrafamilial variability includes onset ages from 7 to 35 years and phenotypes from pure paraplegia to spastic ataxia [9].

8.4 Predictive testing, prenatal and preimplantation diagnosis

Predictive testing is possible when the familial variant is known (or, historically, when a family was sufficiently large for linkage studies) but should be restricted to adults able to participate fully in decision-making, following approaches used in other adult-onset neurodegenerative diseases [4, 8]. Predictive testing in asymptomatic minors is considered inappropriate because it removes the child's autonomy without benefit, while testing symptomatic individuals in a family with an established diagnosis is appropriate at any age [8]. Prenatal diagnosis (technically possible through chorionic villus sampling) and preimplantation genetic diagnosis are possible once familial variants are identified; because of variable expressivity, a prenatal SPG4 result does not predict onset age, course or disability degree [4, 8]. The ideal time for these discussions is before pregnancy, and counselling should be offered to affected, carrier or at-risk young adults [2]. Differences in perspectives between professionals and within families on use of these tests should be discussed [2].

8.5 AP-4 deficiency

AP-4 deficiency has autosomal recessive inheritance. If both parents are heterozygous, each sibling of an affected individual has, at conception, a 25% risk of being affected, a 50% risk of being an asymptomatic carrier and a 25% chance of being neither affected nor a carrier [14]. Consanguinity was recorded in about half of the registry patients (46%, 142/311) and in about two thirds of the 2020 cohort; the difference reflects mainly the ascertainment bias of the first series, centred on countries with a high rate of consanguinity [14, 15]. In consanguineous couples and in families with recurrent variants, such as the AP4B1 variant of the Jain community of Rajasthan, carrier testing extended to relatives and to the spouses of carriers has a higher yield [15]. Once both variants have been identified, carrier testing, prenatal diagnosis and preimplantation genetic testing are possible [14]. Younger siblings of an affected child may be tested soon after birth, because early diagnosis determines access to gene therapy trials with age limits (section 7.11); this is a diagnostic test in a childhood-onset disease, and not a predictive test for an adult-onset disease [14, 17].

8.6 Disclosure of the diagnosis, counselling and legal framework according to the French PNDS

The PNDS devotes a dedicated consultation to the disclosure of the genetic diagnosis, given by the physician who requested the analysis, with a clinical psychologist present when desirable and a further consultation soon afterwards if the information cannot be given all at once; in paediatric neurology, it may be useful to see the young person without the parents [21]. That consultation explains the mode of transmission, recalls the legal obligation to inform relatives at risk, plans multidisciplinary follow-up and gives information about patient associations and the national network coordinated by the reference centre [21]. The protocol recommends taking a sample from the parent likely to be a carrier at the initial assessment, for analysis in parallel with the index case, because studying several relatives, especially several affected individuals, increases the probability of a molecular diagnosis and immediately clarifies the parents' status; in apparently de novo cases, testing the parents for the variant should be offered systematically, preferably within the framework of a presymptomatic testing consultation [21].

Counselling is reliable only with a molecular diagnosis, that is, with one (AD or X-linked) or two (AR) pathogenic or likely pathogenic variants (classes 5 and 4); it should never be based on a variant of uncertain significance (class 3) [21]. In a sporadic case without a molecular diagnosis, the clinician should not reassure the counsellee by assuming an AR form, because an isolated case may result from an AD form with a paucisymptomatic or asymptomatic relative, from a de novo variant, from non-paternity or from an X-linked form transmitted by an asymptomatic mother; and, given intrafamilial variability, future severity cannot be predicted for the patient or for the offspring, an uncertainty that must be explained [21]. In AR forms, the risk that the spouse of an affected person carries a variant in the same gene lies, in the great majority of cases, between about 1/10 000 and 1/1 000,000 (the frequency of heterozygotes in the general population for each gene), but becomes relevant in the presence of consanguinity or of a variant with a founder effect, situations in which targeted sequencing of the gene may be offered to the spouse [21]. For the siblings of a patient with an AD form whose variant was not detected in the leucocyte DNA of either parent, the PNDS indicates a risk slightly higher than that of the general population, but below 1%, because of the possibility of germline mosaicism (described in SPG4 in four French families, without conclusive epidemiological data); GeneReviews rounds the same risk to about 1% (section 8.2), a difference of wording without practical impact [2, 8, 21].

In France, presymptomatic testing in adults may be performed only in clinical genetics services with teams declared to the Agence de la biomédecine and with a multidisciplinary protocol including at least one clinician (geneticist or neurologist) and a psychologist or psychiatrist; the counsellee first meets the clinician and then the psychologist, sometimes over several consultations before the sample is taken [21]. In asymptomatic minors, presymptomatic testing for a disease with usual adult onset and without beneficial early treatment is not permitted by law, because it deprives the child of autonomy and of the right not to know without any counterbalancing benefit; as the request is frequent, this rule must be explained to parents [21]. Prenatal diagnosis and preimplantation genetic testing require identified familial variant(s), known status of the parent at risk and a disease considered of particular severity and incurable: each request is considered by a multidisciplinary prenatal diagnosis centre (CPDPN), the only body competent to issue, at the woman's request, the certificate of severity that makes the procedure lawful, and which may refer the couple to another CPDPN in case of refusal [21]. Prenatal diagnosis without the intention of terminating the pregnancy is not accepted, as it would be equivalent to presymptomatic testing in a minor, and the result does not predict severity, age at onset or the degree of disability [21]. These rules coincide in essence with the general recommendations cited in section 8.4, but add the French legal framework, which is not transposable to other countries [4, 8, 21].

9. Special Considerations: Population Diversity and Epidemiology

9.1 Global and regional prevalence

In Ruano and colleagues' systematic review (22 studies, 14,539 patients from 16 countries), AD-HSP prevalence ranged from 0.5 to 5.5/100 000 and AR-HSP from 0.0 to 5.3/100 000, with pooled means of 1.8/100 000 (95% CI: 1.0-2.7) and 1.8/100 000 (95% CI: 1.0-2.6), respectively; SPG4 was the most frequent AD form in all populations, followed by SPG3A, and SPG11 the most frequent AR form, followed by SPG15 [18]. Population-based studies using multiple sources produced higher values than hospital- or genetics-centre-based studies, and large world regions remain without prevalence studies [18]. Vander Stichele and colleagues' integrated epidemiological model estimated global prevalence of 3.6/100 000 for all HSP and 0.90 (SPG4), 0.22 (SPG7), 0.34 (SPG11) and 0.13 (SPG15) per 100,000, corresponding to approximately 3365 symptomatic SPG4 patients and 872 SPG11 patients in the United States [20].

Orphanet data record the following validated point prevalences of HSP per 100,000 population: Europe 4.8; Spain 9.6; Norway 7.4; Tunisia 5.75; Estonia 4.4; Portugal 4.1; Italy 2.7; Libya 2.1; worldwide value 4.2; and, for Sweden, 1.3 (not validated) [5]. The Portuguese national survey (1994-2004), population-based and with the participation of family doctors, identified 1336 patients with hereditary ataxia or HSP in a population of 10.3 million (12.9/100 000); the prevalence of HSP was 4.1/100 000 (2.4 for dominant and 1.6 for recessive forms), and the most prevalent forms were SPG4 (0.91/100 000), SPG11 (0.26/100 000) and SPG3 (0.14/100 000) [19]. In Portugal, SPG11 accounts for 13% of all spastic paraplegias, irrespective of the mode of inheritance [10]. Orphanet classifies SPG76 in the prevalence class below 1/1 000,000 [5]. For France, the 2024 PNDS estimates a prevalence of about 1/20 000 (close to 3500 patients), a figure presented without a description of the method or the reference period and which should be read as an order of magnitude [21]; according to the same protocol, SPG4 accounts for more than 40% of pure AD forms (an estimate supported by the French series of 1550 probands), SPG3A for 5–10%, SPG30 and SPG31 for about 5% each and SPG10 for about 3% (section 2.1) [21].

9.2 Subtype distribution and consanguinity effects

SPG4 accounts for 40% of inherited AD-HSP and 20% of simplex HSP in series from several European countries, the United States, Canada, Japan and China; one study found fewer SPG4 families in North America than expected from European data [8]. SPG11 is more frequent in families originating from countries with high consanguinity rates, particularly the Mediterranean basin and Middle East, although described worldwide, and GeneReviews estimates prevalence of 1.25/100 000 from its 19-31% proportion of AR-HSP, above the Portuguese population study estimate [10, 19]. AR forms overall are more frequent in more consanguineous populations [2]. In SPG7, p.Ala510Val is present on at least one allele in 58% of patients in the European cohort, making its detection relevant in populations of European origin [45].

9.3 Special populations

In children, HSP may be confused with cerebral palsy; in Ortega and Rosemberg's Brazilian series, complicated forms predominated (23 of 35) and had earlier onset [53], and in Ikeda and colleagues' Japanese series genetic diagnosis was obtained in 45% of pure and 81% of complex forms, with SPAST predominant in pure and KIF1A in complex forms [54]. SPASTOX, conducted in Brazil, is the largest controlled botulinum toxin trial in HSP [82], and SPRS has a validated Brazilian Portuguese version [62]. Women with SPG4 have lower penetrance but a more diffuse phenotype when symptomatic [41].

9.4 Data gaps

No population prevalence studies of HSP in Latin America (including Brazil), sub-Saharan Africa, most of Asia or the Middle East were found in the consulted sources — which does not demonstrate their non-existence; available regional data come from hospital series and studies of relative gene frequencies. For SPG7 and SPG3A, no direct population prevalences exist outside Portugal and modelled estimates, and for SPG76 Orphanet records only published cases and families, without measurable population prevalence [5, 20].

9.5 AP-4 deficiency

There are no population prevalence estimates for AP-4 deficiency in the sources consulted. More than 300 patients are known, 311 of whom were included in the registry and natural history study NCT04712812 up to January 2025, from families in North America, South America, Europe, the Middle East, China and the Indian subcontinent [14]. In the registry, AP4B1 (~36%) and AP4M1 (~35%) account for most cases, followed by AP4S1 (~19%) and AP4E1 (~10%) [14]. The 2020 cohort included European, North American, Middle Eastern, Indian and Brazilian (Salvador) centres, and the study of 428 cases of early-onset HSP included a Brazilian centre (Campinas); neither of these works presents frequencies by country [15, 58]. In a Lebanese series of 21 children with HSP and a molecular diagnosis (2010–2024), 71.4% were the children of consanguineous couples and variants were found in AP4B1, AP4M1 and AP4S1, which illustrates the weight of recessive forms in populations with high consanguinity [112]. The possible founder variant of AP4B1 in the Jain community of Rajasthan is the only founder effect suggested in the sources consulted; no data from Portugal were found in these sources [15].

10. Illustrative Clinical Case

10.1 Illustrative case (composite for teaching purposes): SPG4 with an exon deletion

A 42-year-old man is referred to genetics for progressive gait disturbance of six years' duration, occasional falls and shoe toe wear. He reports urinary urgency and nocturnal calf cramps. His father, who died at 70 years, had “dragged his feet” from 50 years and an unconfirmed “multiple sclerosis” diagnosis; a paternal aunt uses a walking stick. Examination shows a spastic gait, increased adductor and triceps surae tone disproportionate to weakness (mild iliopsoas and tibialis anterior weakness), brisk reflexes in all four limbs, bilateral sustained Achilles clonus, extensor plantar responses, reduced malleolar vibration sense and pes cavus. These findings, present in at least two relatives, fulfil clinical HSP criteria with an AD-compatible pattern [4]. Brain and whole spinal cord MRI shows only mild thoracic cord atrophy; vitamin B12, vitamin E, copper, syphilis, HIV and HTLV-1 serology and very-long-chain fatty acids are normal [4, 2].

A sequencing-only HSP panel identifies no pathogenic variants. Because 20-25% of SPAST variants are exon deletions or duplications, SPAST MLPA is requested, revealing a heterozygous multiexon deletion classified as pathogenic based on established gene haploinsufficiency [8, 40]. Urodynamic studies demonstrate detrusor overactivity, treated with an antimuscarinic; oral baclofen titration begins at 5 mg three times daily, with a usual effective dose of 40–80 mg/day and maximum 80 mg/day, alongside a physiotherapy programme with stretching [8, 71]. Counselling explains to his two asymptomatic adult sons (24 and 27 years) the 50% risk of inheriting the deletion, age-dependent penetrance (85% at 45 years, complete at 70) and inability to predict onset age or severity; the 24-year-old opts for predictive testing after structured counselling consultations, while the 27-year-old decides to defer [4, 8]. The proband's 15-year-old daughter is not tested, following recommendations on predictive testing in asymptomatic minors [8].

This case illustrates four practical points: a relative's previous misdiagnosis, the need to test for deletions when sequencing is negative, the importance of early neurogenic bladder treatment and the ethical framework for predictive testing.

10.2 Published cases: SPG76 caused by CAPN1 variants

Zhu and colleagues described three Chinese SPG76 patients aged 48, 39 and 48 years, two from consanguineous families and one sporadic [55]. Examination showed lower-limb hypertonia, hyperreflexia and Babinski signs; patients 2 and 3 also had dysarthria and depression [55]. Exome sequencing, confirmed by Sanger and cosegregation studies, identified homozygous c.213dupG (p.Asp72Glyfs95) in patient 1 and c.1729+1G>A in patient 3, and compound heterozygosity for c.853C>T (p.Arg285) and c.1324G>A (p.Gly442Ser) in patient 2; c.213dupG and c.1324G>A were described for the first time [55]. Functional studies showed that wild-type calpain 1 localises in the cytoplasm and colocalises with tubulin, while p.Asp72Glyfs*95 and p.Gly442Ser forms produce intracellular aggregates with little tubulin colocalisation [55]. On the MANE transcript (NM_005186.4), the nonsense variant is written c.853C>T p.(Arg285*), classified as pathogenic/likely pathogenic in ClinVar; the others should be confirmed on the authors' transcript before comparison, and the duplication is written c.213dup under HGVS [55, 68]. The series confirms the adult spastic ataxia phenotype, with or without bulbar involvement and emotional disturbance, and reinforces follow-up mood assessment [46, 55].

10.3 Published case: individualised gene therapy in SPG50

Dowling and colleagues described a Canadian boy with SPG50 diagnosed at 18 months by exome sequencing, performed because of developmental delay (he neither stood nor walked without support and produced no words) and microcephaly; he had two AP4M1 variants, described by the authors as c.916C>T p.(Arg306*) and c.696dupG p.(Glu232Glyfs*21), without indication of the transcript in the text consulted [52]. On the MANE Select transcript NM_004722.4, the first corresponds to c.916C>T p.(Arg306*); in the second, position c.696 is an adenine and the published protein consequence corresponds to c.694dup p.(Glu232Glyfs*21), a variant classified as pathogenic in ClinVar [50, 51]. Before treatment, at 4 years of age, he crawled about 1.5 m, stood momentarily with support on a table and took a few steps with assistance, and examination showed diffuse spasticity, predominant in the lower limbs, and hyperreflexia [52]. On the initiative of a foundation created by the family, a self-complementary AAV9 vector with codon-optimised AP4M1 complementary DNA was developed and administered intrathecally at a dose of 1 × 10¹⁵ vector genomes in March 2022, 2 years and 11 months after the diagnosis, in a single-patient phase 1 trial authorised by Health Canada and with immunosuppression with prednisolone, sirolimus and tacrolimus [52]. Over 12 months there were no serious adverse events; transient neutropenia occurred on day 6 and a prolonged episode of abdominal symptoms attributed to tacrolimus and to Clostridioides difficile infection, without dorsal root ganglion toxicity and without inflammatory changes on MRI [52]. The spasticity scales were poorly tolerated, with missing data, but without worsening in the joints assessed; the Bayley scores increased in most domains, and the patient became able to stand with his heels on the ground (global clinical impression of minimal improvement) [52]. The authors stress that a single patient does not allow conclusions about efficacy in a disease with slow and variable progression and that the estimated preclinical cost was 3.5 million Canadian dollars [52]. The figures of the original open-access article (Figure 1: developmental timeline and trial scheme; Figure 2: adverse events and change in the developmental scales) are available at https://pmc.ncbi.nlm.nih.gov/articles/PMC11271397/ and are not reproduced here. The case illustrates the pathway, the ethical safeguards and the limits of the individualised gene therapy that led to the ongoing phase 1–2 and phase 3 trials [17].

11. Prognosis and Long-Term Follow-Up

11.1 Subtype-specific prognosis

In pure forms, manifestations may be disabling but life expectancy is not reduced [2]. In SPG4, severity increases with duration: after 20 years, approximately 50% need walking assistance and 10% a wheelchair, with faster progression in onset after 35 years [8]. In SPG3A, progression is slow and walking-aid dependence relatively rare, particularly in adult-onset forms [11]. In SPG7, progression is continuous: severe gait disability may develop from 8 years after onset, both spasticity and ataxia may lead to wheelchair dependence, and SARA increases by a mean of 1.0 point per year [9, 45]. In SPG11, most patients become wheelchair-dependent one to two decades after onset, and corpus callosum thinning correlates with severity [10]. In SPG76, long-term natural history is uncharacterised beyond published series [46]. In a prospective Austrian cohort from a rare movement disorder centre (126 patients, 103 with baseline data, 54.8% complicated forms, genetic confirmation in 54.0% and SPAST in 36.8% of confirmed cases), SPRS increased by a mean of 0.9 points per year over up to 5 years' follow-up (mean 2.3 years), 1.3 points annually in complicated and 0.6 in pure forms; this is a hospital series, not representative of the population [93]. In early-onset SPG4, Alecu and colleagues' international cohort distinguished a severe subgroup with rapid walking loss from a moderate subgroup with biphasic progression (section 2.8) [47].

11.2 Surveillance

Because of the insidious progression, it is advisable that reassessment include driving, housing and urinary symptoms, since disability may have progressed appreciably since the last consultation [4]. Follow-up of sphincter disturbances prevents bladder dysfunction and infections, and early, regular physiotherapy prevents contractures [10]. Storing DNA from probands without a molecular diagnosis allows reanalysis as methods and knowledge evolve [2]. The PNDS intervals (neurology, PRM and bladder and sphincter follow-up 1 to 2 times a year) are compatible with those of GeneReviews for pure forms and for SPG3A; the shorter GeneReviews intervals in SPG4 (6 to 12 months) and in SPG11 (6 months) refer to consultations for treatment adjustment and, in SPG11, to a complex form outside the scope of the French protocol [2, 8, 10, 11, 21].

11.3 Resources

The European Reference Network for Rare Neurological Diseases (ERN-RND) includes cerebellar ataxias and HSP among its disease groups and provides clinical information and intercentre case discussion [23]. Patient associations and registries mentioned by GeneReviews include EURO HSP, the Spastic Paraplegia Foundation (United States), Tom Wahlig Stiftung (Germany), HSP Research Foundation (Australia), A.I.Vi.P.S. (Italy), NORD and NINDS, and the SPATAX network for ataxia and spastic paraparesis registries [2, 10]. The Oxford Desk Reference mentions the UK Familial Spastic Paraplegia Group [4].

In France, care is organised around the rare disease reference centre (CRMR) “Neurogénétique et maladies génétiques rares du système nerveux”, of the BRAIN-TEAM filière, with 10 reference centres (coordination at the Pitié-Salpêtrière, in Paris, and constituent centres in Angers, Bordeaux, Dijon, Fort-de-France, Lille, Montpellier, Paris-Trousseau for the paediatric age, Rouen and Strasbourg) and 18 competence centres, including Saint-Pierre, on Réunion; the patient association is ASL-HSP-France (Association Strümpell-Lorrain), which offers social guidance, psychological support and information about research [21]. Follow-up is coordinated by the general practitioner, in liaison with the reference neurologist or paediatric neurologist and the PRM physician, ideally in a day hospital or multidisciplinary clinic and with an up-to-date list of direct contacts for all those involved [21].

HSPs are recognised as a long-term condition (ALD 30; ALD 17 in metabolic forms), declared by the general practitioner through a care protocol validated by the health insurance fund; transport to a specialised centre more than 150 km away requires prior approval at least 15 days in advance, stating on the form that the destination is a reference centre [21]. The application to the MDPH (or MDA) is the main gateway to entitlements and should be opened early; the first medical certificate should be completed by the reference or competence centre, with the wording “progressive rare neurological disease” and the genetic form when known, and the application may be accompanied by the supplementary form for rare diseases, the Orphanet “focus handicap” sheet and the PNDS itself [21]. Support includes schooling (a personalised schooling plan with a monitoring team meeting at least once a year, support staff for pupils with disabilities, SESSAD, ULIS or motor education institutes, an individualised care plan), benefits (AEEH in children, AAH and PCH in adults, an invalidity pension below 62 years of age if the capacity to work or to earn is reduced by at least two thirds), recognition of the status of disabled worker (RQTH), the APA application above 60 years of age in certain situations, care and support services at home or in an institution and, for carers, respite periods, leave (AJPP, parental presence leave, carer's leave) and training; the assessment of needs at home or in an institution may draw on the SERAFIN-PH nomenclature [21]. These measures describe the French system; in Portugal and Brazil, referral is to local neurogenetics and medical genetics centres and, in Europe, to the ERN-RND centres [23].

11.4 AP-4 deficiency: prognosis and surveillance

The disease is progressive: spasticity advances from diplegia to tetraplegia (mean ages of 8.4 and 16.1 years in the 2020 cohort), independent walking, when acquired, is frequently lost, and more than half of patients are wheelchair-dependent in adolescence [15]. Early epilepsy is associated with a worse motor outcome, whereas seizures tend to decrease with age [14, 15]. Dysphagia, aspiration, scoliosis and restrictive lung disease are the late complications with the greatest respiratory impact [14]. There are no published data on survival; the oldest known patient is 55 years old [14]. The SPRS, the SPATAX scale, the motor milestones of the GMFM-88 and the CPCHILD are the instruments used in natural history studies and in trials, and plasma NfL is under study as a disease marker [17, 59, 61].

Specific resources include the SP-CERN research network (Spastic Paraplegia Centers of Excellence Research Network), based at Boston Children's Hospital, the registry and natural history study NCT04712812 and the patient foundation CureAP4; in Europe, the ERN-RND covers the HSPs [14, 23, 58].

12. Summary and Final Considerations

Hereditary spastic paraplegias encompass more than 80 genetic entities with a common denominator: length-dependent corticospinal axon degeneration producing progressive lower-limb spastic paraparesis. Harding's pure/complicated classification continues to guide the approach, but the boundary is permeable within each gene and family, and the ataxia–spasticity spectrum requires considering HSP and hereditary ataxias together. AD inheritance predominates in pure and AR in complicated forms, with minor XLR and mitochondrial contributions.

Diagnosis rests on three pillars. The first is clinical and familial: three-generation history, parental examination and firm diagnostic criteria. The second is systematic exclusion of acquired causes (spinal cord compression, vitamin B12, vitamin E and copper deficiencies, HTLV-1, HIV, syphilis, multiple sclerosis) and treatable inborn errors (dopa-responsive dystonia, adrenomyeloneuropathy, remethylation defects, arginase deficiency, HHH syndrome, biotinidase deficiency, cerebrotendinous xanthomatosis). The third is genetic testing covering deletions and duplications, particularly for SPAST and SPG11, and genome sequencing when only one SPG7 variant is identified.

The five subtypes detailed here illustrate distinct mechanisms: spastin haploinsufficiency, a microtubule-severing enzyme, in SPG4, the most frequent form, with bimodal onset, age- and sex-dependent penetrance, and bladder dysfunction in more than 90%; altered ER membrane fusion by atlastin-1 in SPG3A, the main cause of AD-HSP beginning before 10 years and allelic to HSN1D sensory neuropathy; loss of mitochondrial m-AAA protease in SPG7, adult spastic ataxia with optic atrophy, ophthalmoplegia, pseudodominance and missing heritability issues; failure of lysosomal reformation in SPG11, a complicated form with a thin corpus callosum, “ears of the lynx” sign, cognitive decline, neuropathy and continuity with juvenile ALS and CMT2X; and calpain 1 deficiency in SPG76, young-adult spastic ataxia with white matter lesions potentially mimicking acquired disease.

Treatment remains symptomatic and multidisciplinary. Oral baclofen and tizanidine, with label-guided titrated dosage, botulinum toxin and intrathecal baclofen for severe refractory spasticity, ongoing physiotherapy and early neurogenic bladder treatment form the basis; HSP-specific controlled evidence is limited and, for botulinum toxin, did not demonstrate functional walking benefit. Mechanism-targeted therapies such as SPG50 gene therapy (in a phase 3 trial since April 2026) or SPG5 oxysterol reduction open prospects but are not available for the most frequent subtypes; in SPG4, silencing and replacement gene therapy remains preclinical, and recent controlled trials (ZISPAST) again proved negative for walking. For the geneticist, counselling requires integrating incomplete age-dependent penetrance, intrafamilial variability, parental mosaicism and recessive-form particularities, and applying ethical principles for predictive testing in adult-onset neurodegenerative diseases.

The 2024 French PNDS for pure HSP is the only published national guideline for this group: it makes systematic brain and spinal cord MRI, a metabolic, audiological, ophthalmological and electrophysiological work-up and a dopaminergic trial in the presence of a compatible phenotype; it places next-generation sequencing first line, with genome sequencing before 50 years of age; and it structures follow-up into neurology, PRM and neuro-urology consultations 1 to 2 times a year, weekly physiotherapy, prescribed physical activity and attention to pain, fatigue, sexuality, pregnancy, anaesthesia (suxamethonium contraindicated) and transition [21].

AP-4 complex deficiency (SPG47, SPG50, SPG51 and SPG52) brings together four genes with a common phenotype beginning in the first year of life — global developmental delay, microcephaly, febrile seizures and status epilepticus, stereotypic laughter and spastic paraparesis from 4–6 years of age, progressing to tetraplegia — and an MRI signature with a thin corpus callosum and a thin or absent anterior commissure; many children are first diagnosed as having cerebral palsy. Treatment is supportive, and intrathecal gene therapy with AAV9/AP4M1 is in phase 1/2, phase IIb and phase 3 trials, which reinforces the importance of early diagnosis.

13. Bibliographic References

1. Walusinski O. A historical approach to hereditary spastic paraplegia. Rev Neurol (Paris). 2020;176(4):225-34. doi:10.1016/j.neurol.2019.11.003.
2. Hedera P. Uncomplicated (Pure) Hereditary Spastic Paraplegia Overview. 2000 Aug 15 [updated 2025 Jun 5]. In: Adam MP, Bick S, Mirzaa GM, Pagon RA, Wallace SE, Amemiya A, editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993-2026. Available at: https://www.ncbi.nlm.nih.gov/books/NBK1509/.
3. Harding AE. Classification of the hereditary ataxias and paraplegias. Lancet. 1983;1(8334):1151-5. doi:10.1016/s0140-6736(83)92879-9.
4. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 448-50 (Hereditary spastic paraplegias).
5. Orphanet. Hereditary spastic paraplegia (ORPHA:685); Pure hereditary spastic paraplegia (ORPHA:102012); Complex hereditary spastic paraplegia (ORPHA:102013); SPG4 (ORPHA:100985); SPG7 (ORPHA:99013); SPG11 (ORPHA:2822); SPG3 (ORPHA:100984); SPG76 (ORPHA:488594) [Internet]. Paris: INSERM; and Orphadata, epidemiology of rare diseases (en_product9_prev) [accessed 2026 Oct 3]. Available at: https://www.orpha.net/en/disease/detail/685.
6. Orphanet. Severe intellectual disability and progressive spastic paraplegia (ORPHA:280763); genes associated with AP4B1, AP4E1, AP4M1 and AP4S1 [Internet]. Paris: INSERM [accessed 2026 Nov 1]. Available at: https://www.orpha.net/en/disease/detail/280763.
7. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 2, Inborn Errors of Metabolism in Adults – A Diagnostic Approach to Neurological and Psychiatric Presentations, section 2.3.8 Spastic Paraparesis (Table 2.10).
8. Parodi L, Rydning SL, Tallaksen C, Durr A. Spastic Paraplegia 4. 2003 Apr 17 [updated 2019 Jun 13]. In: Adam MP, Bick S, Mirzaa GM, Pagon RA, Wallace SE, Amemiya A, editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993-2026. Available at: https://www.ncbi.nlm.nih.gov/books/NBK1160/.
9. Hedera P. SPG7-Related Neurologic Disorder. 2006 Aug 24 [updated 2026 Jan 8]. In: Adam MP, Bick S, Mirzaa GM, Pagon RA, Wallace SE, Amemiya A, editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993-2026. Available at: https://www.ncbi.nlm.nih.gov/books/NBK1107/.
10. Stevanin G. Spastic Paraplegia 11. 2008 Mar 27 [updated 2019 Dec 19]. In: Adam MP, Bick S, Mirzaa GM, Pagon RA, Wallace SE, Amemiya A, editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993-2026. Available at: https://www.ncbi.nlm.nih.gov/books/NBK1210/.
11. Hedera P. Spastic Paraplegia 3A. 2010 Sep 21 [updated 2020 Jun 18]. In: Adam MP, Bick S, Mirzaa GM, Pagon RA, Wallace SE, Amemiya A, editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993-2026. Available at: https://www.ncbi.nlm.nih.gov/books/NBK45978/.
12. Gan-Or Z, Bouslam N, Birouk N, Lissouba A, Chambers DB, Vérièpe J, et al. Mutations in CAPN1 cause autosomal-recessive hereditary spastic paraplegia. Am J Hum Genet. 2016;98(5):1038-46. doi:10.1016/j.ajhg.2016.04.002.
13. Online Mendelian Inheritance in Man (OMIM). Spastic paraplegia 47, autosomal recessive (#614066); spastic paraplegia 50, autosomal recessive (#612936); spastic paraplegia 51, autosomal recessive (#613744); spastic paraplegia 52, autosomal recessive (#614067); AP4B1 (*607245), AP4M1 (*602296), AP4E1 (*607244), AP4S1 (*607243) [Internet]. Baltimore (MD): Johns Hopkins University [accessed 2026 Nov 1]. Available at: https://www.omim.org/.
14. Alecu J, Schierbaum L, Ebrahimi-Fakhari D. AP-4-Associated Hereditary Spastic Paraplegia. 2018 Dec 13 [updated 2025 Feb 6]. In: Adam MP, Bick S, Mirzaa GM, Pagon RA, Wallace SE, Amemiya A, editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993-2026. PMID: 30543385. Available at: https://www.ncbi.nlm.nih.gov/books/NBK535153/.
15. Ebrahimi-Fakhari D, Teinert J, Behne R, Wimmer M, D'Amore A, Eberhardt K, et al. Defining the clinical, molecular and imaging spectrum of adaptor protein complex 4-associated hereditary spastic paraplegia. Brain. 2020;143(10):2929-44. doi:10.1093/brain/awz307. PMID: 32979048. Erratum in: Brain. 2021;144(3):e33.
16. Ebrahimi-Fakhari D, Alecu JE, Ziegler M, Geisel G, Jordan C, D'Amore A, et al. Systematic analysis of brain MRI findings in adaptor protein complex 4-associated hereditary spastic paraplegia. Neurology. 2021;97(19):e1942-54. doi:10.1212/WNL.0000000000012836. PMID: 34544818.
17. ClinicalTrials.gov. NCT05518188: phase 1/2 open-label intrathecal administration of MELPIDA in spastic paraplegia type 50 (SPG50); NCT06692712: phase 3, open-label trial of intrathecal MELPIDA (AAV9/AP4M1) with matched prospective concurrent control arm. Elpida Therapeutics [Internet]. Bethesda (MD): National Library of Medicine [accessed 2026 Oct 6]. Available at: https://clinicaltrials.gov/study/NCT05518188 ; https://clinicaltrials.gov/study/NCT06692712
18. Ruano L, Melo C, Silva MC, Coutinho P. The global epidemiology of hereditary ataxia and spastic paraplegia: a systematic review of prevalence studies. Neuroepidemiology. 2014;42(3):174-83. doi:10.1159/000358801.
19. Coutinho P, Ruano L, Loureiro JL, Cruz VT, Barros J, Tuna A, et al. Hereditary ataxia and spastic paraplegia in Portugal: a population-based prevalence study. JAMA Neurol. 2013;70(6):746-55. doi:10.1001/jamaneurol.2013.1707.
20. Vander Stichele G, Durr A, Yoon G, Schüle R, Blackstone C, Esposito G, et al. An integrated modelling methodology for estimating global incidence and prevalence of hereditary spastic paraplegia subtypes SPG4, SPG7, SPG11, and SPG15. BMC Neurol. 2022;22(1):115. doi:10.1186/s12883-022-02595-4.
21. Haute Autorité de Santé. Protocole National de Diagnostic et de Soins (PNDS): Paraparésies spastiques héréditaires pures. Saint-Denis La Plaine: HAS; 2024 Sep 25 [text of September 2024, Centre de Référence Neurogénétique et maladies génétiques rares du système nerveux, filière BRAIN-TEAM]. Available at: https://www.has-sante.fr/jcms/p_3542901.
22. Cipriano L, Angelini C, Santorelli FM. Hereditary spastic paraplegia: from decades of therapy to future innovations. Ther Adv Neurol Disord. 2026;19:17562864251406589. doi:10.1177/17562864251406589.
23. European Reference Network on Rare Neurological Diseases (ERN-RND). Disease knowledge hub: cerebellar ataxia and hereditary spastic paraplegias [Internet] [accessed 2026 Oct 6]. Available at: https://www.ern-rnd.eu/disease-knowledge-hub/hereditary-spastic-paraplegia/
24. Shribman S, Reid E, Crosby AH, Houlden H, Warner TT. Hereditary spastic paraplegia: from diagnosis to emerging therapeutic approaches. Lancet Neurol. 2019;18(12):1136-46. doi:10.1016/S1474-4422(19)30235-2.
25. Faber I, Pereira ER, Martinez ARM, França M, Teive HAG. Hereditary spastic paraplegia from 1880 to 2017: an historical review. Arq Neuropsiquiatr. 2017;75(11):813-8. doi:10.1590/0004-282X20170160.
26. Harding AE. Hereditary "pure" spastic paraplegia: a clinical and genetic study of 22 families. J Neurol Neurosurg Psychiatry. 1981;44(10):871-83. doi:10.1136/jnnp.44.10.871.
27. Casari G, De Fusco M, Ciarmatori S, Zeviani M, Mora M, Fernandez P, et al. Spastic paraplegia and OXPHOS impairment caused by mutations in paraplegin, a nuclear-encoded mitochondrial metalloprotease. Cell. 1998;93(6):973-83. doi:10.1016/s0092-8674(00)81203-9.
28. Hazan J, Fonknechten N, Mavel D, Paternotte C, Samson D, Artiguenave F, et al. Spastin, a new AAA protein, is altered in the most frequent form of autosomal dominant spastic paraplegia. Nat Genet. 1999;23(3):296-303. doi:10.1038/15472.
29. Zhao X, Alvarado D, Rainier S, Lemons R, Hedera P, Weber CH, et al. Mutations in a newly identified GTPase gene cause autosomal dominant hereditary spastic paraplegia. Nat Genet. 2001;29(3):326-31. doi:10.1038/ng758.
30. Stevanin G, Santorelli FM, Azzedine H, Coutinho P, Chomilier J, Denora PS, et al. Mutations in SPG11, encoding spatacsin, are a major cause of spastic paraplegia with thin corpus callosum. Nat Genet. 2007;39(3):366-72. doi:10.1038/ng1980.
31. Wang Y, Hersheson J, Lopez D, Hammer M, Liu Y, Lee KH, et al. Defects in the CAPN1 gene result in alterations in cerebellar development and cerebellar ataxia in mice and humans. Cell Rep. 2016;16(1):79-91. doi:10.1016/j.celrep.2016.05.044.
32. Verkerk AJ, Schot R, Dumee B, Schellekens K, Swagemakers S, Bertoli-Avella AM, et al. Mutation in the AP4M1 gene provides a model for neuroaxonal injury in cerebral palsy. Am J Hum Genet. 2009;85(1):40-52. doi:10.1016/j.ajhg.2009.06.004. PMID: 19559397.
33. Moreno-De-Luca A, Helmers SL, Mao H, Burns TG, Melton AM, Schmidt KR, et al. Adaptor protein complex-4 (AP-4) deficiency causes a novel autosomal recessive cerebral palsy syndrome with microcephaly and intellectual disability. J Med Genet. 2011;48(2):141-4. doi:10.1136/jmg.2010.082263. PMID: 20972249.
34. Abou Jamra R, Philippe O, Raas-Rothschild A, Eck SH, Graf E, Buchert R, et al. Adaptor protein complex 4 deficiency causes severe autosomal-recessive intellectual disability, progressive spastic paraplegia, shy character, and short stature. Am J Hum Genet. 2011;88(6):788-95. doi:10.1016/j.ajhg.2011.04.019. PMID: 21620353.
35. Erfanian Omidvar M, Torkamandi S, Rezaei S, Alipoor B, Omrani MD, Darvish H, et al. Genotype-phenotype associations in hereditary spastic paraplegia: a systematic review and meta-analysis on 13,570 patients. J Neurol. 2021;268(6):2065-82. doi:10.1007/s00415-019-09633-1.
36. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 44, Disorders of Cellular Trafficking, section 44.1.3 Other Types of Cellular Trafficking.
37. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 35, section 35.4 Inborn Errors of Phospholipid Remodelling (35.4.2 PLA2G6; 35.4.4 PNPLA6; 35.4.5 DDHD1/DDHD2; 35.4.6 CYP2U1).
38. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 40, section 40.1 Disorders of Sphingolipid Synthesis (40.1.5 FA2H/SPG35; 40.1.7 GM2/GD2 synthase/SPG26).
39. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 21, section 21.3 Δ1-Pyrroline-5-Carboxylate Synthetase Deficiency.
40. ClinGen. Dosage Sensitivity Curation: SPAST (HGNC:11233) [Internet]. Clinical Genome Resource [accessed 2026 Oct 3]. Available at: https://search.clinicalgenome.org/kb/gene-dosage/HGNC:11233.
41. Parodi L, Fenu S, Barbier M, Banneau G, Duyckaerts C, Tezenas du Montcel S, et al. Spastic paraplegia due to SPAST mutations is modified by the underlying mutation and sex. Brain. 2018;141(12):3331-42. doi:10.1093/brain/awy285.
42. Beetz C, Nygren AO, Schickel J, Auer-Grumbach M, Bürk K, Heide G, et al. High frequency of partial SPAST deletions in autosomal dominant hereditary spastic paraplegia. Neurology. 2006;67(11):1926-30. doi:10.1212/01.wnl.0000244413.49258.f5.
43. Guelly C, Zhu PP, Leonardis L, Papić L, Zidar J, Schabhüttl M, et al. Targeted high-throughput sequencing identifies mutations in atlastin-1 as a cause of hereditary sensory neuropathy type I. Am J Hum Genet. 2011;88(1):99-105. doi:10.1016/j.ajhg.2010.12.003.
44. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 10, Disorders of Oxidative Phosphorylation, section 10.3.2 Nuclear Gene Defects.
45. Coarelli G, Schule R, van de Warrenburg BPC, De Jonghe P, Ewenczyk C, Martinuzzi A, et al. Loss of paraplegin drives spasticity rather than ataxia in a cohort of 241 patients with SPG7. Neurology. 2019;92(23):e2679-90. doi:10.1212/WNL.0000000000007606.
46. Méreaux JL, Firanescu C, Coarelli G, Kvarnung M, Rodrigues R, Pegoraro E, et al. Increasing involvement of CAPN1 variants in spastic ataxias and phenotype-genotype correlations. Neurogenetics. 2021;22(1):71-9. doi:10.1007/s10048-020-00633-2.
47. Alecu JE, Schierbaum LM, Tam A, Quiroz V, Bernardi K, Yang K, et al. Genotype-structure-phenotype correlations define divergent natural history in early-onset spastic paraplegia type 4. Brain. 2026 May 5 [Epub ahead of print]. doi:10.1093/brain/awag150. PMID: 42083457.
48. García-Cazorla Á, Dionisi-Vici C, Saudubray JM. Disorders of Cellular Trafficking. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 44, secções 44.1.1 Membrane Trafficking e 44.3.1 Neurological Manifestations. doi:10.1007/978-3-662-63123-2_44.
49. Saffari A, Brechmann B, Böger C, Saber WA, Jumo H, Whye D, et al. High-content screening identifies a small molecule that restores AP-4-dependent protein trafficking in neuronal models of AP-4-associated hereditary spastic paraplegia. Nat Commun. 2024;15(1):584. doi:10.1038/s41467-023-44264-1. PMID: 38233389.
50. Morales J, Pujar S, Loveland JE, Astashyn A, Bennett R, Berry A, et al. A joint NCBI and EMBL-EBI transcript set for clinical genomics and research. Nature. 2022;604(7905):310-5. doi:10.1038/s41586-022-04558-8. PMID: 35388217. Ficheiro MANE GRCh38 v1.5: https://ftp.ncbi.nlm.nih.gov/refseq/MANE/
51. ClinVar. NM_004722.4(AP4M1):c.694dup (p.Glu232fs), VCV000818030 [Internet]. Bethesda (MD): National Center for Biotechnology Information [accessed 2026 Nov 1]. Available at: https://www.ncbi.nlm.nih.gov/clinvar/variation/818030/.
52. Dowling JJ, Pirovolakis T, Devakandan K, Stosic A, Pidsadny M, Nigro E, et al. AAV gene therapy for hereditary spastic paraplegia type 50: a phase 1 trial in a single patient. Nat Med. 2024;30(7):1882-7. doi:10.1038/s41591-024-03078-4.
53. Ortega RPM, Rosemberg S. Hereditary spastic paraplegia: a clinical and epidemiological study of a Brazilian pediatric population. Arq Neuropsiquiatr. 2019;77(1):10-8. doi:10.1590/0004-282X20180153.
54. Ikeda A, Kumaki T, Tsuyusaki Y, Tsuji M, Enomoto Y, Fujita A, et al. Genetic and clinical features of pediatric-onset hereditary spastic paraplegia: a single-center study in Japan. Front Neurol. 2023;14:1085228. doi:10.3389/fneur.2023.1085228.
55. Zhu Z, Hou W, Cao Y, Zheng H, Tian W, Cao L. Spastic paraplegia type 76 due to novel CAPN1 mutations: three case reports with literature review. Neurogenetics. 2023;24(4):243-50. doi:10.1007/s10048-023-00726-8.
56. Travaglini L, Bellacchio E, Aiello C, Pro S, Bertini E, Nicita F. Expanding the clinical phenotype of CAPN1-associated mutations: a new case with congenital-onset pure spastic paraplegia. J Neurol Sci. 2017;378:210-2. doi:10.1016/j.jns.2017.05.014.
57. Alkhalifa A, Chen S, Hasiloglu ZI, Filosto M, Cali E, Houlden H, et al. White matter abnormalities in 15 subjects with SPG76. J Neurol. 2023;270(12):5784-92. doi:10.1007/s00415-023-11918-5.
58. Resch D, Alecu JE, Yang K, Quiroz V, Schierbaum L, Bernardi K, et al. Spectrum of movement disorders in early-onset hereditary spastic paraplegia: a study of 428 cases. Mov Disord. 2026;41(3):615-24. doi:10.1002/mds.70141. PMID: 41328529.
59. Jordan C, Geisel G, Alecu JE, Zhang B, Sahin M, Ebrahimi-Fakhari D. Disease severity and motor impairment correlate with health-related quality of life in AP-4-associated hereditary spastic paraplegia. Neurol Genet. 2021;7(4):e605. doi:10.1212/NXG.0000000000000605. PMID: 34295967.
60. Brooks AK, Quiroz V, Schierbaum L, Alecu JE, Bernardi K, Battaglia N, et al. Fetal and perinatal brain MRI findings in adaptor protein complex 4-associated hereditary spastic paraplegia. Ann Child Neurol Soc. 2025;3(3):226-31. doi:10.1002/cns3.70035. PMID: 41815511.
61. Alecu JE, Saffari A, Ziegler M, Jordan C, Tam A, Kim S, et al. Plasma neurofilament light chain is elevated in adaptor protein complex 4-related hereditary spastic paraplegia. Mov Disord. 2023;38(9):1742-50. doi:10.1002/mds.29524. PMID: 37482941.
62. Servelhere KR, Faber I, Coan AC, França M. Translation and validation into Brazilian Portuguese of the Spastic Paraplegia Rating Scale (SPRS). Arq Neuropsiquiatr. 2016;74(6):489-94. doi:10.1590/0004-282X20160047.
63. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 1, Clinical Approach to Inborn Errors of Metabolism in Paediatrics, section 1.5.2.4 Category 4: With Complex Motor Disorders.
64. Richards S, Aziz N, Bale S, Bick D, Das S, Gastier-Foster J, et al. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genet Med. 2015;17(5):405-24. doi:10.1038/gim.2015.30. PMID: 25741868.
65. National Center for Biotechnology Information. ClinVar: VCV000005671, NM_014946.4(SPAST):c.131C>T (p.Ser44Leu); VCV000005672, NM_014946.4(SPAST):c.134C>A (p.Pro45Gln) [Internet]. Bethesda (MD): NLM [accessed 2026 Oct 6]. Available at: https://www.ncbi.nlm.nih.gov/clinvar/variation/5671/ ; https://www.ncbi.nlm.nih.gov/clinvar/variation/5672/
66. National Center for Biotechnology Information. ClinVar: VCV000004346, NM_015915.5(ATL1):c.715C>T (p.Arg239Cys); VCV000004353, c.470T>G (p.Leu157Trp); VCV000004352, c.1243C>T (p.Arg415Trp) [Internet]. Bethesda (MD): NLM [accessed 2026 Oct 6]. Available at: https://www.ncbi.nlm.nih.gov/clinvar/variation/4346/
67. National Center for Biotechnology Information. ClinVar: VCV000042016, NM_003119.4(SPG7):c.1529C>T (p.Ala510Val) [Internet]. Bethesda (MD): NLM [accessed 2026 Oct 6]. Available at: https://www.ncbi.nlm.nih.gov/clinvar/variation/42016/
68. National Center for Biotechnology Information. ClinVar: VCV000489121, NM_005186.4(CAPN1):c.853C>T (p.Arg285Ter) [Internet]. Bethesda (MD): NLM [accessed 2026 Oct 6]. Available at: https://www.ncbi.nlm.nih.gov/clinvar/variation/489121/
69. Agianda HAP, Kim HM, Battaglia N, Rong J, Tam A, Gonzalez Saez-Diez E, et al. Diagnostic utility of the ATG9A ratio in AP-4-associated hereditary spastic paraplegia. Ann Clin Transl Neurol. 2026;13(4):834-9. doi:10.1002/acn3.70308. PMID: 41491634.
70. Nyhan WL, Hoffmann GF, Al-Aqeel AI, Barshop BA. Atlas of Inherited Metabolic Diseases. 4th ed. Boca Raton: CRC Press; 2020. p. 241–246 (Argininemia; p. 242).
71. Advagen Pharma Ltd. Baclofen tablets, USP [label]. DailyMed. Bethesda (MD): National Library of Medicine; revision dated 2026 Jan 23. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=04dac0f6-cc6c-4959-9bca-a6e4a99690f3.
72. Covis Pharma US. Zanaflex (tizanidine hydrochloride) tablets and capsules [label]. DailyMed. Bethesda (MD): National Library of Medicine; revision dated 2025 May 5. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=043d9e51-bfa2-4add-9058-5ece332f7e99.
73. Amneal Pharmaceuticals. Lioresal Intrathecal (baclofen injection) [label]. DailyMed. Bethesda (MD): National Library of Medicine; revision dated 2022 Dec 9. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=6f3bdcc0-ffc8-4911-8c35-958c8103a2c5.
74. Allergan. BOTOX (onabotulinumtoxinA) for injection [label]. DailyMed. Bethesda (MD): National Library of Medicine; revision dated 2023 Nov 18. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=33d066a9-34ff-4a1a-b38b-d10983df3300.
75. Ardolino G, Bocci T, Nigro M, Vergari M, Di Fonzo A, Bonato S, et al. Spinal direct current stimulation (tsDCS) in hereditary spastic paraplegias (HSP): a sham-controlled crossover study. J Spinal Cord Med. 2021;44(1):46-53. doi:10.1080/10790268.2018.1543926.
76. Santos GJBD, Severiano MIR, Gonçalves FM, Araujo CM, Zeigelboim BS, Teive HAG, et al. Vestibular rehabilitation in hereditary spastic paraplegia: a randomized pilot study. Arq Neuropsiquiatr. 2026;84(7):1-10. doi:10.1055/s-0046-1827050. PMID: 42674425.
77. ClinicalTrials.gov. NCT07561359 (HSPMOVE): 12-week strength and functional exercise protocol for treating motor symptoms in individuals with hereditary spastic paraplegia: adaptive randomized crossover clinical trial. Hospital de Clínicas de Porto Alegre [Internet]. Bethesda (MD): National Library of Medicine [accessed 2026 Oct 6]. Available at: https://clinicaltrials.gov/study/NCT07561359
78. Rosemont Pharmaceuticals. OZOBAX DS (baclofen oral solution) [label]. DailyMed. Bethesda (MD): National Library of Medicine; revision dated 2026 Mar 13. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=d8ddcec5-debe-42b4-a803-e65917cc6a54.
79. Amneal Pharmaceuticals. LYVISPAH (baclofen oral granules) [label]. DailyMed. Bethesda (MD): National Library of Medicine; revision dated 2023 Apr 23. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=f8992a56-9010-4c4b-84d5-ddcc3e2c3317.
80. Piramal Critical Care. Gablofen (baclofen injection), intrathecal [label]. DailyMed. Bethesda (MD): National Library of Medicine; revision dated 2026 May 7. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=00d3e846-dd92-448d-9ab8-6a07be823cc1.
81. Galderma Laboratories. Dysport (abobotulinumtoxinA) for injection [label]. DailyMed. Bethesda (MD): National Library of Medicine; revision dated 2024 Oct 24. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=71313a04-1349-4c26-b840-a39e4a3ddaed.
82. Diniz de Lima F, Faber I, Servelhere KR, Bittar MFR, Martinez ARM, Piovesana LG, et al. Randomized trial of botulinum toxin type A in hereditary spastic paraplegia – the SPASTOX trial. Mov Disord. 2021;36(7):1654-63. doi:10.1002/mds.28523.
83. Servelhere KR, Faber I, Martinez A, Nickel R, Moro A, Germiniani FMB, et al. Botulinum toxin for hereditary spastic paraplegia: effects on motor and non-motor manifestations. Arq Neuropsiquiatr. 2018;76(3):183-8. doi:10.1590/0004-282X20180013.
84. Margetis K, Korfias S, Boutos N, Gatzonis S, Themistocleous M, Siatouni A, et al. Intrathecal baclofen therapy for the symptomatic treatment of hereditary spastic paraplegia. Clin Neurol Neurosurg. 2014;123:142-5. doi:10.1016/j.clineuro.2014.05.024.
85. de Lima FD, Servelhere KR, Bittar MFR, González-Salazar C, Martinez ARM, Benaglia T, et al. Ziclague® (Alpinia zerumbet oil) in patients with hereditary spastic paraplegia – the randomized controlled ZISPAST trial. Orphanet J Rare Dis. 2025;20(1):481. doi:10.1186/s13023-025-04007-5. PMID: 40993748.
86. Béreau M, Anheim M, Chanson JB, Tio G, Echaniz-Laguna A, Depienne C, et al. Dalfampridine in hereditary spastic paraplegia: a prospective, open study. J Neurol. 2015;262(5):1285-8. doi:10.1007/s00415-015-7707-6.
87. Selcuk Muhtaroglu F, Belgen Kaygisiz B, Usar Incirli S, Kahraman T. Dalfampridine as a promising agent in the management of hereditary spastic paraplegia: a triple-blinded, randomized, placebo-controlled pilot trial. J Clin Neurosci. 2023;117:136-42. doi:10.1016/j.jocn.2023.09.026.
88. Mero S, Ricca I, Rossi S, Mellone S, Musumeci O, Zanna GD, et al. Miglustat does not impact clinical progression in patients with spastic paraplegia type 11. Neurogenetics. 2025;27(1):3. doi:10.1007/s10048-025-00874-z.
89. Vantaggiato C, Guarato G, Brivio F, Panzeri E, Speltoni B, Gumeni S, et al. Naringenin and SMER28 target lysosomal reformation and rescue SPG11 and SPG15 hereditary spastic paraplegia phenotypes. Pharmacol Res. 2025;218:107836. doi:10.1016/j.phrs.2025.107836.
90. Piermarini E, Guha S, Qiang L, Gray-Edwards H, Sena-Esteves M, Baas PW. Intracerebroventricular SPAST-AAV9 gene therapy prevents manifestation of symptoms in a mouse model of SPG4 hereditary spastic paraplegia. Mol Ther. 2026;34(3):1729-42. doi:10.1016/j.ymthe.2025.11.029. PMID: 41311060.
91. ClinicalTrials.gov. NCT07542548: D-cycloserine for serine palmitoyltransferase inhibition: treatment of a single patient with a complex hereditary spastic paraplegia due to a de novo variant in the SPTSSA gene. Massachusetts General Hospital [Internet]. Bethesda (MD): National Library of Medicine [accessed 2026 Oct 6]. Available at: https://clinicaltrials.gov/study/NCT07542548
92. Scheuer KH, Svenstrup K, Jennum P, Rogvi-Hansen B, Werdelin L, Fenger K, et al. Double-blind crossover trial of gabapentin in SPG4-linked hereditary spastic paraplegia. Eur J Neurol. 2007;14(6):663-6. doi:10.1111/j.1468-1331.2007.01812.x. PMID: 17539946.
93. Amprosi M, Indelicato E, Eigentler A, Boesch D, Fritz J, Nachbauer W, et al. Natural history in hereditary spastic paraplegias: real-world data from an Austrian cohort. J Neurol. 2026;273(2):97. doi:10.1007/s00415-025-13606-y. PMID: 41586880.
94. U.S. Food and Drug Administration. Drugs@FDA (via openFDA): Lioresal NDA 017851 (1977 Nov 22); Ozobax NDA 208193 (2019 Sep 18); Lyvispah NDA 215422 (2021 Nov 22); Fleqsuvy NDA 215602 (2022 Feb 4); Zanaflex NDA 020397 (1996 Nov 27) e NDA 021447 (2002 Aug 29); Lioresal Intrathecal NDA 020075 (1992 Jun 17); Gablofen NDA 022462 (2010 Nov 19); Botox BLA 103000 (1991 Dec 9); Dysport BLA 125274 (2009 Apr 29); Ampyra NDA 022250 (2010 Jan 22); Zavesca NDA 021348 (2003 Jul 31) [Internet; accessed 2026 Oct 6]. Available at: https://open.fda.gov/data/drugsfda/
95. European Medicines Agency. Fampyra (fampridine): EPAR. EMEA/H/C/002097. Authorisation 2011 Jul 20 [accessed 2026 Oct 6]. Available at: https://www.ema.europa.eu/en/medicines/human/EPAR/fampyra
96. European Medicines Agency. Zavesca (miglustat): EPAR. EMEA/H/C/000435. Authorisation 2002 Nov 20 [accessed 2026 Oct 6]. Available at: https://www.ema.europa.eu/en/medicines/human/EPAR/zavesca
97. Clinical Trials Information System (CTIS). EU CT 2025-522603-15-00: a phase IIb, monocentric, non-profit, open-label trial for the intrathecal administration of AAV9/AP4M1 for hereditary spastic paraplegia type 50 (SPG50) (protocol AAV9/AP4M1-ITA-01). Sponsor: Azienda Socio Sanitaria Territoriale Papa Giovanni XXIII [Internet]. Amsterdam: European Medicines Agency [accessed 2026 Nov 1]. Available at: https://euclinicaltrials.eu/search-for-clinical-trials/?lang=en&EUCT=2025-522603-15-00.
98. ClinicalTrials.gov. NCT06948019: safety and efficacy of AAV9/AP4B1 (BFB-101) for patients with AP4B1-related hereditary spastic paraplegia type 47 (SPG47). BlackfinBio Ltd [Internet]. Bethesda (MD): National Library of Medicine; last updated 2025 Apr 28 [accessed 2026 Nov 1]. Available at: https://clinicaltrials.gov/study/NCT06948019.
99. Chen X, Dong T, Hu Y, De Pace R, Mattera R, Eberhardt K, et al. Intrathecal AAV9/AP4M1 gene therapy for hereditary spastic paraplegia 50 shows safety and efficacy in preclinical studies. J Clin Invest. 2023;133(10):e164575. doi:10.1172/JCI164575. PMID: 36951961.
100. Wiseman JP, Scarrott JM, Alves-Cruzeiro J, Saffari A, Böger C, Karyka E, et al. Pre-clinical development of AP4B1 gene replacement therapy for hereditary spastic paraplegia type 47. EMBO Mol Med. 2024;16(11):2882-917. doi:10.1038/s44321-024-00148-5. PMID: 39358605.
101. Ziegler M, Günter C, Alecu JE, Xue X, Kim HM, Saffari A, et al. CRISPR/Cas9 loss-of-function screen in a neuronal model of AP-4 deficiency identifies ATG9A trafficking modulators. JCI Insight. 2026;11(14):e202204. doi:10.1172/jci.insight.202204. PMID: 42154535.
102. Par Health USA. Dantrium (dantrolene sodium) capsules [label]. DailyMed. Bethesda (MD): National Library of Medicine; revision dated 2026 Aug 25. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=57b4eec1-49e1-432c-95ae-0fe36e32d64b.
103. Viatris Specialty. Lyrica (pregabalin) capsules and oral solution [label]. DailyMed. Bethesda (MD): National Library of Medicine; revision dated 2025 Apr 15. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=d4734e7d-5079-455e-8ff5-8f4539c998a9.
104. Viatris Specialty. Neurontin (gabapentin) capsules, tablets and oral solution [label]. DailyMed. Bethesda (MD): National Library of Medicine; revision dated 2026 May 12. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=97935fd9-1d4a-43b6-a5d9-de994591187b.
105. A-S Medication Solutions. Oxybutynin chloride tablets, USP [label]. DailyMed. Bethesda (MD): National Library of Medicine; revision dated 2025 Nov 18. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=0697b3cf-b146-47c0-b5a5-5905ae8d1c97.
106. American Health Packaging. Oxybutynin chloride extended-release tablets [label]. DailyMed. Bethesda (MD): National Library of Medicine; revision dated 2026 Jun 24. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=0c71bdfb-01db-49ac-a8e9-0037df0d24af.
107. Glenmark Pharmaceuticals USA. Solifenacin succinate tablets [label]. DailyMed. Bethesda (MD): National Library of Medicine; revision dated 2026 Jan 28. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=010c0657-c907-477f-8bd7-5d455a5f50eb.
108. Astellas Pharma US. Myrbetriq (mirabegron) extended-release tablets and extended-release oral suspension [label]. DailyMed. Bethesda (MD): National Library of Medicine; revision dated 2024 Aug 1. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=ba9e9e15-e666-4c56-9271-2e24739cfa2d.
109. Cardinal Health 107. Tamsulosin hydrochloride capsules [label]. DailyMed. Bethesda (MD): National Library of Medicine; revision dated 2026 Aug 25. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=0e3206ba-db3e-40b3-b46b-c7ae4703fdea.
110. American Health Packaging. Desmopressin acetate tablets [label]. DailyMed. Bethesda (MD): National Library of Medicine; revision dated 2025 Nov 13. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=0d447b48-3f4b-430e-9e68-d8b9401c002e.
111. Pfizer Laboratories. Viagra (sildenafil citrate) tablets [label]. DailyMed. Bethesda (MD): National Library of Medicine; revision dated 2017 Dec 15. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=0b0be196-0c62-461c-94f4-9a35339b4501.
112. Habanjar D, Kreidly S, Trad S, Boustany RM. Clinical and genetic characteristics of pediatric hereditary spastic paraplegia in the Eastern Levant. Eur J Paediatr Neurol. 2026;63:69-75. doi:10.1016/j.ejpn.2026.07.003. PMID: 42462372.