Leigh Syndrome and Leigh Spectrum

Subacute necrotising encephalomyelopathy: a spectrum of disorders of mitochondrial energy production of nuclear and mitochondrial origin, including treatable forms due to thiamine transport deficiency, pyruvate dehydrogenase complex deficiency, biotinidase deficiency and primary coenzyme Q10 deficiencies

Literature and regulatory information updated: 6 October 2026.

0. Initial Reference Table

1. Introduction

Leigh syndrome was described in 1951 by Denis Leigh, a neuropathologist at the Maudsley Hospital in London, in an infant whose autopsy revealed bilateral symmetrical necrotic foci in the basal ganglia, thalamus, brainstem, dentate nuclei and optic nerves, which he termed “subacute necrotising encephalomyelopathy” [1, 17, 18]. Lesions have a spongiform appearance, with demyelination, gliosis and vascular proliferation, and relative neuronal preservation, a pattern resembling Wernicke encephalopathy but, unlike the latter, usually sparing the mammillary bodies [1, 19]. For decades, diagnosis was mainly neuropathological and post-mortem; magnetic resonance imaging (MRI) enabled recognition of the pattern during life, and next-generation sequencing transformed the syndrome into an entity defined by a neuroradiological and biochemical phenotype shared by more than a hundred monogenic diseases [2, 6].

The first molecular cause identified was a mitochondrial DNA (mtDNA) variant. In 1990, Holt and colleagues described a family with maternally transmitted neurodegenerative disease across three generations, with neurogenic muscle weakness, ataxia and pigmentary retinopathy, which they named NARP (neurogenic muscle weakness, ataxia, and retinitis pigmentosa), associated with the NC_012920.1:m.8993T>G variant in the ATP synthase subunit 6 gene [20, 21]. In 1992, Tatuch and colleagues found the same variant, at a high proportion, in a female infant who died at 7 months with typical Leigh neuropathology, and in 1993 Santorelli and colleagues showed that the variant at nt 8993 was a frequent cause of LS; that same year, the second substitution at the same position, m.8993T>C, was described [21–23]. In 1998, two groups identified the first nuclear gene, SURF1, an assembly factor for complex IV (cytochrome c oxidase, COX) [24, 25]. In 2003, integrative genomics identified LRPPRC as the cause of the Saguenay-Lac-Saint-Jean French-Canadian form, a classic example of a founder effect [26].

Genetic expansion was rapid. In 2016, Lake and colleagues counted more than 75 monogenic causes; current GeneReviews chapters list more than 115 nuclear genes and 15 mitochondrial genes, and the 2023 ClinGen curation assessed 113 primary mitochondrial disease genes for this spectrum [1, 2, 8, 27]. This heterogeneity prompted a change in nomenclature: the term “Leigh-like syndrome”, used when the presentation suggested LS but did not meet strict criteria (atypical neuroimaging, normal lactate, atypical neuropathology or incomplete assessment), was incorporated into the concept of Leigh syndrome spectrum, and the GeneReviews authors recommend that it no longer be used [1, 2]. Similarly, the designation MILS for mtDNA forms may mislead because approximately one quarter of probands have de novo variants rather than maternal inheritance [2].

The clinical interest of the Leigh spectrum lies in three aspects. First, it is the most frequent phenotype of paediatric-onset mitochondrial disease [2, 6]. Second, a minority of causes have targeted therapy whose efficacy depends on speed: biotin-thiamine-responsive basal ganglia disease may reverse within days, and pyruvate dehydrogenase complex deficiency benefits from an early ketogenic diet [9, 10]. Third, genetic counselling varies radically according to the affected genome, with recurrence risks ranging from near zero (de novo variants) to 25% (autosomal recessive inheritance) and to estimates dependent on maternal heteroplasmy in mtDNA forms [2, 3]. This article treats the ATP synthase form (MT-ATP6) as a subtype; mitochondrial disease in general and MELAS (mitochondrial encephalomyopathy, lactic acidosis and stroke-like episodes) are mentioned only as differential diagnoses.

2. Molecular Genetics and Pathophysiology

2.1. Biochemical basis: why the developing brain is the target organ

The respiratory chain (complexes I to IV) and oxidative phosphorylation system (OXPHOS, complexes I to V) are located in the inner mitochondrial membrane. Complex I (NADH-ubiquinone oxidoreductase) has 44 subunits, seven encoded by mtDNA; complex II (succinate-ubiquinone oxidoreductase) has four subunits, all nuclear; complex III has 11 subunits, one of them (cytochrome b) mitochondrial; complex IV has 14 polypeptides, three mitochondrial; and complex V (ATP synthase) has 16 subunits, two mitochondrial (ATP6 and ATP8) [6]. Complexes I, III and IV pump protons into the intermembrane space, generating a potential of approximately 150 mV that complex V uses to synthesise ATP; coenzyme Q10 and cytochrome c carry electrons between complexes [6]. Human mtDNA, circular, with 16,569 base pairs, encodes 37 genes (13 OXPHOS polypeptides, 22 transfer RNAs and two ribosomal RNAs), without introns, translated by its own machinery with a partly different genetic code [2, 28]. Assembly and functioning of the system depend on more than 200 proteins for mtDNA replication, transcription and translation, assembly factors, cofactors (CoQ10, iron-sulphur clusters, lipoic acid, haem, copper), specific lipids such as cardiolipin and organelle quality control, and up to 10% of the human proteome is estimated to participate in this metabolism [6].

When OXPHOS fails, the cell resorts to anaerobic glycolysis, approximately 20 times less efficient, which generates lactate and, through transamination, alanine; elevated lactate, pyruvate and alanine is typical but not specific to these defects, and there is never complete loss of function because residual activity always exists, varying between tissues [6]. The pyruvate dehydrogenase complex (PDHc) is the rate-limiting step in cerebral glucose oxidation: the fully activated complex is estimated to oxidise approximately 180 g of glucose daily, whereas the brain normally oxidises approximately 125 g, leaving little margin for any reduction in activity [29]. In PDHc deficiency, NADH production decreases but its oxidation is maintained, so the lactate/pyruvate (L/P) ratio remains normal in blood and CSF, unlike respiratory chain defects, in which the ratio tends to rise [30]. The infant brain, with its high proportion of brain mass and dependence on glucose oxidation, is therefore the most vulnerable organ, and the basal ganglia and brainstem, with high energy demands, concentrate the lesions [5, 29].

2.2. Mitochondrial DNA forms: heteroplasmy, threshold and bottleneck

MtDNA variants associated with Leigh spectrum comprise transfer RNA gene variants, which reduce mitochondrial protein synthesis through abnormalities of base modification and aminoacylation, and coding gene variants, which reduce activity of the corresponding complex; the mechanism is loss of function [2]. Each cell contains hundreds to thousands of mtDNA copies; coexistence of mutated and normal molecules (heteroplasmy) and the need for a minimum quantity of normal mtDNA to maintain function (threshold effect) explain clinical variability [6, 28]. The threshold varies by variant: for many frequent variants it lies between 80% and 95% mutant load, for large deletions between 50% and 70%, and some transfer RNA variants with a dominant-negative effect are pathogenic below 20% [6]. The statement, found in some Portuguese-language sources, that at least 95% of mtDNA must be compromised for disease to occur applies only to some variants and should not be generalised [5, 6].

The mitochondrial genes involved and their relative contribution, according to cohorts with acknowledged selection bias, are shown in Table 2.1 [2].

All variants in Table 2.1 are classified as pathogenic in ClinVar by an expert panel (accessed on 6 October 2026); numbering uses reference sequence NC_012920.1 and protein changes are those stated in GeneReviews [2, 31]. Older literature refers to “T8993G” or “8993 T-to-G”; this is the same variant, which should be recorded as m.8993T>G [2, 21]. Large mtDNA deletions are an exceptional cause of LSS [2].

In MT-ATP6, subunit a forms, together with subunit c, the proton channel of the F(0) domain; the m.8993T>G variant replaces a conserved hydrophobic leucine with a hydrophilic arginine at position 156 and reduces ATP synthesis but not hydrolysis, whereas m.8993T>C replaces the leucine with proline and tends to cause less severe disease [4, 21]. The m.8993T>G and m.8993T>C variants show little heteroplasmy variation between tissues and with age and a strong correlation between load and severity [2]. For m.8993T>G, loads below approximately 60% are associated with no symptoms or only mild pigmentary retinopathy or migraine, although asymptomatic adults with up to 75% have been described; loads of 70–90% are associated with the NARP phenotype and loads above 90% with LSS; for m.8993T>C, almost all symptomatic individuals have more than 90%, but there are asymptomatic individuals above this value [2]. Nyhan's Atlas reports, in the initial series, loads above 95% in infantile encephalopathy, of 80–90% in late-onset cases and of only 3–6% in asymptomatic individuals, and documents an asymptomatic mother with 39% in blood and 71% in fibroblasts [21]. There is overlap between asymptomatic individuals, NARP and LSS, particularly for variants other than m.8993T>G, and published regression models should be read as indicative rather than precise [2, 32]. For m.3243A>G, m.8344A>G, m.9176T>C, m.14459G>A, m.14487T>C, m.10158T>C, m.10191T>C and m.13513G>A, the correlation is weak and heteroplasmy has predictive value only when close to 0% or 100% [2]. De novo insertions in MT-ATP6 (for example, NC_012920.1:m.8618dup, published as m.8618_8619insT) also cause NARP, so analysis should not be limited to position 8993 [21].

2.3. Nuclear forms: mechanisms and inheritance

Nuclear-encoded forms result from variants in genes required for OXPHOS activity and assembly, mtDNA maintenance and expression, cofactor biosynthesis (lipoic acid and CoQ10), membrane lipid remodelling, mitochondrial quality control and dynamics, pyruvate metabolism and B-vitamin transport and metabolism, and also from organic acidurias and toxic metabolites that secondarily inhibit OXPHOS [1, 6]. Of the more than 115 nuclear genes, all but six follow autosomal recessive inheritance; AIFM1, NDUFA1, HSD17B10 and PDHA1 are X-linked — AIFM1 accounts for less than 1% of cases and the Leigh spectrum is only one of the gene's presentations —, SLC25A4 and SSBP1 have caused autosomal dominant disease through de novo variants in isolated cases, and DNM1L and OPA1 are associated with both modes of transmission [1]. Table 2.2 summarises the mechanistic groups.

Source: tables 2a, 2b, 3, 4a, 4b and 5 of Rahman, Thorburn and Ball [1].

2.4. Gene-by-gene distinctions

Table 2.3 brings together the genes with the greatest contribution, targeted therapy or clinical, biochemical or imaging markers guiding diagnosis; proportions refer to nuclear-encoded LSS rather than total LSS [1].

SURF1 deserves emphasis as the most frequent single nuclear cause. It encodes an inner mitochondrial membrane protein integrated into the MITRAC complex, which regulates complex IV assembly; biallelic loss leaves the complex unassembled [4]. The multicentre study by Wedatilake and colleagues, involving 44 patients from ten UK and two Australian centres, showed onset in the first year in 73% (median 9.5 months), poor weight gain in 95%, hypotonia in 93%, regression in 71%, oculomotor involvement in 52% and central respiratory failure in 78% (median 31 months), with hypertrichosis in 41% (GeneReviews, the source of Table 2.3, gives 48%) and cardiomyopathy in only 2% [1, 4, 34]. At the mildest extreme, the same gene causes demyelinating neuropathy (Charcot-Marie-Tooth type 4K), and COX activity may be only partly reduced even in the absence of protein, so normal or mildly decreased activity does not exclude the gene [4]. The Oxford Desk Reference (2017) attributed 75% of LS with COX deficiency and 10–20% of all LS to SURF1; the more recent 2025 GeneReviews review, based on larger cohorts, gives approximately 50% of LS with COX deficiency [1, 3]. Two recurrent variants, NM_003172.4:c.845_846del p.(Ser282fs) and c.312_321delinsAT p.(Pro104_Leu105insTer), are classified as pathogenic in ClinVar, with multiple concordant submitters [31].

In pyruvate dehydrogenase complex deficiency, PDHA1 (Xp22.12) accounts for 76–85% of primary defect cases, followed by PDHX (7–11%), PDHB (4–9%), DLD (1–6%), DLAT (1–4%) and PDP1 (approximately 1%) [33]. In a review of approximately 466 cases, PDHc deficiency accounted for 74.2% of pyruvate oxidation disorders, and 77% had E1α deficiency [30]. More than 166 PDHA1 variants have been described, half small deletions, insertions or frameshifts and half missense; premature termination codons, particularly in exons 10 and 11, predominate in females and missense variants in males, and null variants are not known in males except in mosaic form, suggesting lethality [30]. Thiamine pyrophosphate is a cofactor of the E1 subunit, and thiamine-responsive variants are missense variants near the cofactor-binding site, particularly in exon 3 [33, 35]. Activating PDK3 variants cause X-linked dominant Charcot-Marie-Tooth type 6 rather than LSS [30, 33].

In SLC19A3 disease, loss of thiamine transporter 2 (THTR2) prevents thiamine passage across the blood-brain barrier, creating cerebral thiamine deficiency with normal serum thiamine and failure of thiamine pyrophosphate-dependent enzymes (PDHc, α-ketoglutarate dehydrogenase, branched-chain α-ketoacid dehydrogenase, transketolase) [9, 35]. In the Saudi population, founder variant NM_025243.4:c.1264A>G p.(Thr422Ala), pathogenic in ClinVar, accounts for 52% (76/146) of cases [9, 31]. Biallelic loss-of-function variants are more associated with the early infantile form, and a combination of a missense variant and a loss-of-function variant with the classical form [9].

2.5. From energy failure to injury: neuropathology

LS lesions are symmetrical foci of spongiform necrosis with neuropil vacuolation, demyelination, gliosis, necrosis and capillary proliferation, with relative neuronal preservation, in the basal ganglia, thalamus, brainstem, cerebellum and spinal cord [1, 6]. In m.8993T>G cases, autopsy showed reduction of the caudate, pallidum, putamen and cerebellum, with gliosis, demyelination and spongiform changes in the thalamus, hypothalamus and medulla [21]. In PDHc deficiency, corpus callosum dysgenesis is frequently associated with migration defects (absent medullary pyramids, ectopic olives, dentate nucleus dysplasia, heterotopias, pachygyria), reflecting prenatal onset, and girls with PDHA1 variants present with non-progressive encephalopathy modulated by X-inactivation [30].

3. Clinical Spectrum and Manifestations by System

3.1. Age at onset and course

Approximately 75% of patients develop manifestations within the first 24 months, and in mtDNA forms typical onset lies between 3 and 12 months, frequently after intercurrent infection, generally viral, or another metabolic challenge (vaccination, surgery, anaesthesia, prolonged fasting) [1, 2]. Intrauterine growth restriction, cardiomegaly, microcephaly and oligohydramnios have been described prenatally [1]. The early form (before 2 years) presents non-specifically with feeding difficulties, persistent vomiting, failure to thrive, hypotonia and developmental delay; the late form (after 2 years) generally progresses more slowly and is dominated by muscle weakness, ataxia, dystonia, peripheral neuropathy and neurobehavioural manifestations, while adult onset may present with ataxia, neuropathy and psychiatric disorder [1, 2]. The Brazilian reference textbook emphasises that onset may be insidious and that a history of febrile episodes or surgery before manifestations should be actively sought because it frequently precedes subacute or abrupt presentation [5].

The course is stepwise: a decompensation, often with rising lactate in blood and/or CSF, causes sudden loss of acquired skills; partial recovery follows, rarely to the previous level, and a plateau that may last months or years until the next episode, sometimes without an apparent trigger [2, 3, 6]. Most patients have at least one acute decompensation requiring hospitalisation; in a European cohort of 130 patients, approximately 44% had experienced exacerbations requiring hospitalisation in the preceding year, mainly due to infections [2, 12].

3.2. Frequency of manifestations

Table 3.1 reproduces GeneReviews estimates, based on approximately 1000 patients from several cohorts for LSS overall and five cohorts with approximately 150 patients for the mtDNA form; some percentages in the second column are based on fewer than 20 patients and should be interpreted cautiously, and not all manifestations were recorded in all patients [1, 2].

3.3. Central and peripheral nervous system

Neurological manifestations include developmental delay and regression, hypotonia (often with loss of head control in infants), spasticity with a Babinski sign, dystonia and other involuntary movements (including chorea), tremor, cerebellar ataxia, hypo- or hyperreflexia, seizures and peripheral neuropathy; psychiatric disorders occur in adolescents and adults [2, 5, 6]. Brainstem lesions explain the signs carrying the greatest threat to life: apnoea, central hypoventilation or hyperventilation, irregular breathing, dysphagia and dysarthria, persistent vomiting and thermoregulatory abnormalities [2]. During acute deterioration, blood gas analysis may reveal mixed respiratory and metabolic acidosis due to central hypopnoea, which may mask the metabolic component [3]. In Nyhan's Atlas, approximately one third of patients with LS have seizures, and blindness, retinal pigment epithelial abnormalities and, late in the course, areflexia complete the presentation; tracheostomy and ventilation may become necessary [19]. The SURF1 and PDHA1 spectra extend to isolated phenotypes such as Charcot-Marie-Tooth-type peripheral neuropathy and episodic dystonia [1].

3.4. Eye, ear, heart, digestive system, liver, kidney and endocrine glands

Ophthalmological manifestations include optic atrophy, retinitis pigmentosa, strabismus, ptosis, nystagmus and ophthalmoparesis [2]. In NARP, night blindness is frequently the first symptom, followed by peripheral field loss and, in some patients, central vision loss; the fundus may show bone-spicule pigment or a “salt-and-pepper” appearance, and nystagmus, esotropia and optic atrophy have been described [21]. Sensorineural deafness, cochlear or involving the auditory nerve, affects approximately 15% [2].

Cardiac manifestations, in approximately 15% of LSS, include hypertrophic or dilated cardiomyopathy and conduction defects such as Wolff-Parkinson-White syndrome and paroxysmal supraventricular tachycardia; cardiomyopathy has been described in up to 70% of patients with MT-ND5 variants [2]. The mitochondrial cardiomyopathy chapter of Clinical Cardiogenetics adds histiocytoid cardiomyopathy and ventricular pre-excitation among forms described in LS, and hypertrophic cardiomyopathy in NARP [36]. The nuclear genes most associated with hypertrophic cardiomyopathy are SCO2, COX10, COX15, NDUFS4, NDUFS1, NDUFS2, NDUFA2, NDUFA10, ECHS1, ATP5PO, ATP5MK and several mtDNA expression genes [1].

Gastrointestinal manifestations include failure to thrive, vomiting, dysphagia, gastroparesis and intestinal dysmotility; in a British natural history study, they were the most frequent extraneurological manifestation, with dysmotility in 25% and a need for gastrostomy or nasogastric feeding in almost half of patients during follow-up [2, 37]. The liver (elevated transaminases, hepatomegaly, liver failure), kidney (tubulopathy, glomerulocystic lesions) and endocrine glands (diabetes, adrenal insufficiency) are involved in a minority [2]. The Brazilian textbook includes sideroblastic anaemia, short stature, hypoparathyroidism, exocrine pancreatic dysfunction and intestinal pseudo-obstruction among systemic manifestations, features belonging to the general mitochondrial disease spectrum that are rare in LSS [5, 6].

3.5. MtDNA-associated phenotypes in relatives and allelic diseases

Maternal relatives of a patient with mtDNA LSS may present with any combination of LSS, NARP or other mitochondrial disease manifestations: mild learning difficulties, ataxia, muscle weakness, night blindness, deafness, diabetes, migraine or sudden unexpected death [2]. Holt's index family included a 47-year-old woman with night blindness from 12 years, nearly blind by 30, with a generalised seizure at 24 years and progressive ataxia in the fourth decade; her asymptomatic sister had retinal pigment and proximal weakness; one niece had retinitis pigmentosa and ataxia; and the other niece, developing normally until 28 months, stopped walking after a febrile illness and was left with retinopathy and pyramidal signs [21]. One patient followed for 20 years had only retinopathy and mild sensory neuropathy at 39 years and was blind, diabetic and had sensorineural deafness at 59 years [21]. The MT-ATP6 spectrum also includes Charcot-Marie-Tooth-type neuropathy, spinocerebellar ataxia with upper motor neuron signs, bilateral striatal necrosis and myopathy with lactic acidosis and sideroblastic anaemia; MT-ND variants may cause Leber hereditary optic neuropathy (LHON) and MELAS, with some patients presenting an overlapping LSS/MELAS phenotype [2, 6].

An international cohort published in 2025, involving 111 patients with MT-ATP6 or MT-ATP8 variants from centres in Italy, Germany, the USA and Spain, found onset before 1 year in 44%, between 1 and 12 years in 36% and after 12 years in 20%; the central nervous system was involved in 93%, muscle in 75%, eye in 46% and heart in 18%; MRI showed isolated Leigh-type lesions in 58%; 11% were wheelchair users and 8% had died at last observation, and metabolic acidosis or acute deterioration complicated approximately 55% of early-onset cases [38].

3.6. Particular features of forms with targeted therapy

In biotin-thiamine-responsive basal ganglia disease (BTBGD), more than 169 patients and three age-related phenotypes have been described [9]. The early infantile form (before 3 months) involves vomiting, feeding difficulties, irritability, encephalopathy, hypotonia, seizures or infantile spasms and respiratory failure, with elevated lactate and poor prognosis even with treatment [9]. The classical form (3 to 10 years) presents with recurrent episodes of acute or subacute encephalopathy triggered by fever (including SARS-CoV-2 infection), trauma, intense exercise, alcohol or vaccination, with confusion, easily controlled seizures, ataxia, supranuclear facial palsy, external ophthalmoplegia, dysphagia, almost constant dystonia and cogwheel rigidity, pyramidal signs and sometimes intracranial hypertension; without treatment it may progress to coma and death [9]. The adult form, resembling Wernicke encephalopathy, has been described in three patients with status epilepticus, ataxia, nystagmus, diplopia and ophthalmoplegia, or rapidly progressive dementia [9].

In PDHc deficiency, hemizygous males present with neonatal lactic acidosis (frequently with corpus callosum agenesis), Leigh encephalopathy, which is the largest group and generally manifests within the first 5 years with respiratory abnormalities, episodic weakness and ataxia and areflexia, or intermittent ataxia after carbohydrate-rich meals [30]. Heterozygous females have a more uniform presentation, with severity varying with X-inactivation: dysmorphism resembling fetal alcohol syndrome (narrow head with frontal bossing, broad nasal root, upturned nose, long philtrum), microcephaly, intellectual disability, spastic diplegia or tetraplegia and seizures in almost all, with infantile spasms within the first 6 months [29, 30]. Presentations have been described of acute flaccid neuropathy resembling Guillain-Barré syndrome, thiamine-responsive exercise-induced paroxysmal dystonia, alternating hemiplegia and psychiatric manifestations in mildly affected adolescents and adults [30, 33]. The 2026 international consensus reports prenatal onset in approximately 50% of patients, detected on routine ultrasound [10].

The French-Canadian LRPPRC variant is characterised by hypotonia, developmental delay, mild facial dysmorphism and compensated chronic metabolic acidosis, interrupted by severe crises of acidosis and coma with high mortality; in the largest cohort, 90% of 56 patients had acute metabolic or neurological crises, including stroke-like episodes, with survival from 5 days to more than 30 years [4, 6, 39].

4. Clinical Recognition and Warning Signs

There are no pathognomonic signs of LS; suspicion arises from a combination of a temporal pattern (stepwise regression triggered by catabolism), signs of brainstem and basal ganglia dysfunction, and compatible laboratory and imaging findings [1, 6]. Signs requiring immediate investigation are summarised in Table 4.1.

The clinical approach proposed by the Oxford Desk Reference begins by verifying that the proband does indeed have a presentation compatible with LS and constructing a three-generation pedigree, extended as far as possible through the maternal line, investigating consanguinity and stored biological material from deceased relatives [3]. Examination should record head circumference (and slowing of its growth), hypotonia, dystonia, spasticity, ataxia, muscle wasting, ptosis, strabismus, external ophthalmoplegia, optic atrophy and signs of cardiomyopathy [3]. GeneReviews recommends detailed neurological and ophthalmological examination in all patients [1].

5. Diagnosis: Criteria, Tests and Algorithms

5.1. Clinical criteria

The ClinGen mitochondrial disease expert panel (McCormick and colleagues, 2023) defined Leigh spectrum criteria adopted by GeneReviews for nuclear and mitochondrial forms; diagnosis is established in a proband meeting these criteria with one or two pathogenic or likely pathogenic variants in an associated gene, compatible with the mode of inheritance (or a heteroplasmic or homoplasmic variant in one of the 15 mitochondrial genes) [1, 2, 8].

Source: adapted from Table 2 of McCormick 2023, as reproduced in GeneReviews [1, 8]. Characteristic neuropathology remains a historical criterion, but autopsy is now rarely performed outside research [1].

5.2. Initial biochemical investigation

The Mitochondrial Medicine Society recommends that the initial blood assessment include a full blood count, creatine kinase, transaminases, albumin, lactate and pyruvate, amino acids and acylcarnitines, with urinary organic acids; when CSF is collected, lactate, pyruvate, amino acids and 5-methyltetrahydrofolate should be measured [13]. Venous lactate may be falsely elevated if a tourniquet is used or the child is agitated, and postprandial lactate is more sensitive than fasting lactate; an elevated plasma lactate (>3 mmol/L) in a correctly collected sample suggests primary or secondary mitochondrial dysfunction [13]. Studies cited by the society attribute to genuinely elevated lactate a sensitivity of 34–62% and a specificity of 83–100% for primary mitochondrial disease, and to the blood L/P ratio a sensitivity of 31% and a specificity of 100%, interpretable only when lactate is elevated [13]. Pyruvate requires immediate deproteinisation with perchloric acid and cooling, without which falsely low values are obtained; alanine (with an alanine/lysine ratio normally below 3:1) reflects hyperpyruvicaemia [6]. A normal lactate in blood or CSF does not exclude LSS [1, 2].

Some profiles point towards specific genes (Table 5.1) [1, 2, 6].

In the international MT-ATP6/MT-ATP8 cohort, citrulline was reduced in 56%, alanine elevated in 49% and lactate in 71% of patients, supporting their use as biomarkers [38]. Based on the 2023 criteria, a Portuguese group proposed performing amino acids, acylcarnitines and urinary organic acids in parallel while awaiting genetic investigation; in their series, this approach characterised 80% of cases and enabled specific intervention in 10% [41]. Biomarkers FGF-21 and GDF-15 correlate with mitochondrial dysfunction particularly in patients with myopathy [6].

5.3. Neuroimaging

The typical pattern consists of bilateral, generally symmetrical T2 hyperintensities in the brainstem and/or basal ganglia, which may fluctuate over time, sometimes with restricted diffusion, and with or without involvement of the thalamus, cerebellum, white matter and spinal cord; cerebral or cerebellar atrophy may develop additionally [1, 2]. Some patterns point to specific genes: lesions of the mammillothalamic tracts, substantia nigra, medial lemniscus, medial longitudinal fasciculus and spinothalamic tracts in NDUFAF2; pallidal involvement followed by putaminal and caudate swelling with an “eye” of dorsal putaminal preservation in SERAC1; brain malformations in hemizygous males and some females with PDHA1 variants; cystic leukoencephalopathy in NDUFV1, NDUFS1 and NDUFA2; a succinate peak on spectroscopy in SDHA/SDHAF1; and, in SLC19A3, caudate and putaminal, thalamic, cortical and brainstem swelling in the acute phase, with subsequent necrosis and atrophy, and thalamic and perirolandic white matter lesions in infants [1, 6, 9]. Leukoencephalopathy may occur in SURF1 and complex I deficiencies [3]. A lactate peak on spectroscopy, in the absence of acute seizures, suggests mitochondrial disease [1].

5.4. Molecular strategy

Genome sequencing is the preferred approach when available because it analyses nuclear and mitochondrial genomes simultaneously and has the highest diagnostic rate; approximately 30% of children with suspected LSS have an mtDNA form and approximately 70% a nuclear form [1, 2]. If unavailable, dedicated mtDNA sequencing combined with exome sequencing is recommended, preferably as a trio in children; mtDNA data extracted from some exome kits have variable coverage and may not be of clinical quality [2]. Targeted testing for frequent variants (for example, m.8993T>G/C) is justified only with strong suspicion or unavailability of comprehensive tests, and multigene panels should include, at minimum, genes with targeted therapy [1, 2, 6]. The 2023 British laboratory guidelines describe the transition from a “biopsy first” approach to genomic analysis in blood and/or urine and specify minimum test and report content [42].

Test limitations determine interpretation of a negative result (Table 5.2).

The report should state the gene, reference sequence with version (for nuclear genes, the MANE Select transcript, for example NM_003172.4 for SURF1, NM_000284.4 for PDHA1 and NM_025243.4 for SLC19A3; for mtDNA, NC_012920.1), HGVS c. and p. (or m.) nomenclature, zygosity or, for mtDNA, heteroplasmy and tissue analysed, ACMG/AMP classification and technical limitations [31, 42, 43]. Pathogenic and likely pathogenic variants are equivalent for clinical decision-making, but identification of biallelic variants of uncertain significance in SLC19A3 does not establish diagnosis [9].

5.5. Biopsy, enzymology and other functional studies

Muscle biopsy was long the first investigation, but is invasive and frequently requires general anaesthesia, which may precipitate decompensation; it is now reserved for critically ill patients or functional validation of variants [1, 2]. Vastus lateralis is the preferred site; enzymology identifies activities below 30% of the control mean; isolated complex I or complex IV deficiencies are most frequent in LSS; fibroblasts detect only approximately half the deficiencies found in muscle; and 10–20% of patients with normal muscle enzymology have deficiency in liver or heart [2]. Respiratory chain enzymology is frequently normal in MT-ATP6 variants, and muscle histology may be normal or show non-specific changes (lipid droplets, type II fibre predominance), rarely ragged-red or COX-negative fibres [2, 21]. In patients with SURF1, COX deficiency is more marked in fibroblasts than muscle [1].

In PDHc deficiency, the 2026 consensus recommends measuring blood lactate and pyruvate and considering CSF, measuring plasma amino acids and urinary organic acids, not performing carbohydrate loading tests, and prioritising molecular investigation (first or in parallel, by clinical exome or genome sequencing) [10]. PDHc activity in cultured fibroblasts has sensitivity of 97% in females and 91% in males with PDHA1 disease, exceeding that of lymphocytes (36%) and muscle (58%) in affected females; normal activity, particularly in muscle and lymphocytes and in a female, does not exclude diagnosis, and low activity may be secondary [10]. Activity should be measured with low and high thiamine pyrophosphate concentrations to detect thiamine-responsive forms [30]. Carbohydrate loading, still mentioned in some sources, is no longer recommended [5, 10].

A pilot newborn screening programme for PDHc deficiency in Ohio, using alanine/leucine ratios >4 and/or proline/leucine >3 in dried blood spots, identified three newborns with the deficiency and two with other mitochondrial diseases, with a positive predictive value of 2.8% and sensitivity of 80% [33, 44]. Prospective diagnosis of MT-ATP6 disease from newborn screening abnormalities (C3 or C5-OH) has also been described, but no population programme exists for LSS [45].

5.6. Practical algorithm

In a child with stepwise regression, brainstem signs or symmetrical basal ganglia lesions, the first step is to exclude and treat causes with targeted therapy: while samples are collected, initiate thiamine and biotin (BTBGD, thiamine deficiencies, biotinidase deficiency) and consider other indicated cofactors, continuing them until biochemical and/or genetic exclusion and lifelong if diagnosis is confirmed [1, 9]. In parallel: blood and CSF lactate and pyruvate, blood gas analysis, glucose, ammonia, amino acids, acylcarnitines, urinary organic acids, biotinidase activity; MRI with diffusion and spectroscopy; echocardiography and electrocardiography; ophthalmological and hearing assessment; and genome sequencing (or mtDNA + trio exome) with storage of DNA and, if possible, fibroblasts [1–3, 13]. Muscle or skin biopsy is required only to validate VUS or in critically ill undiagnosed patients; and a negative investigation should prompt periodic reanalysis and DNA storage [1].

6. Differential Diagnosis

The differential includes, first, treatable causes within the spectrum itself, followed by non-mitochondrial monogenic diseases and acquired conditions with similar symptoms or imaging (Table 6.1) [1, 2].

In practice, clinical and biochemical overlap between genes prevents clinical features or imaging from directing investigation to a subset of genes in most cases, which justifies the genomic approach [1]. In neonatal-onset forms, congenital lactic acidosis broadens the differential to pyruvate carboxylase deficiency and biotinidase deficiency, and to defects of gluconeogenesis, the Krebs cycle, carbonic anhydrase VA and long-chain fatty acid oxidation [6].

7. Treatments and Risk Management Strategies

7.1. General principles

There is no curative treatment for most forms of Leigh spectrum, and on 6 October 2026 no drug is authorised for Leigh syndrome in the United States or European Union [2, 11, 46]. The strategy rests on four pillars: rapidly identify forms with targeted therapy and initiate it before genetic confirmation whenever suspicion justifies this; prevent catabolism and decompensations; treat organ-specific manifestations with a multidisciplinary team; and avoid drugs and circumstances worsening mitochondrial dysfunction [1, 9, 10, 14]. The report of the 258th European Neuromuscular Centre (ENMC) workshop, dedicated to Leigh spectrum, summarises the situation explicitly: for most spectrum diseases there are no effective treatments, and only three nuclear groups have established specific therapy — biotin-thiamine-responsive basal ganglia disease, biotinidase deficiency and coenzyme Q10 biosynthesis deficiencies — alongside ketogenic diet in pyruvate dehydrogenase complex deficiency [47].

Treatment should be conducted by a centre experienced in inherited metabolic or neurometabolic diseases, in liaison with local care; the 2026 international consensus for PDHc deficiency recommends that the team be coordinated by an inherited metabolic disease specialist or neurologist experienced in the disease and include a dietitian (for patients on ketogenic diet), a neurologist and speech and language, physical and occupational therapists, with neuropsychological follow-up and educational programmes; leadership by a rare disease specialist should be supported by a local team including, at minimum, a dietitian experienced in ketogenic diet [10]. The same consensus recommends early integration of palliative care and advance development of a care plan including goals, anticipated complications and family preferences [10]. These recommendations, formulated for PDHc deficiency, can be applied in their organisational logic to the spectrum as a whole, although specific guidelines for LSS overall do not yet exist — a gap patient associations brought to the ENMC workshop, proposing development of care standards similar to those for Duchenne muscular dystrophy [47].

7.2. Forms with targeted therapy: doses and sources

The doses in Table 7.1 derive exclusively from guidelines, consensuses and chapters actually consulted and identified in each row. None of these drugs has marketing authorisation for these indications in the USA or EU, so all are used off-label, except where the vitamin is used as a nutritional supplement [11, 46].

Four principles govern application of this table. First, speed: in BTBGD, manifestations typically resolve within a few days with biotin and thiamine, treatment initiated within the first month after onset significantly improves outcome, and late or absent treatment may cause death or chronic sequelae (dystonia, tetraparesis, epilepsy, mild intellectual disability) [9]. Second, persistence: treatment is lifelong in all these diseases and should continue during pregnancy in women affected by BTBGD [9]. Third, the early form: most infants with BTBGD beginning before 3 months do not respond and die despite supplementation, but GeneReviews considers a trial always worthwhile [9]. Fourth, thiamine alone versus combination: in a controlled study comparing biotin plus thiamine with thiamine alone, all patients diagnosed and treated early had good outcomes, without a difference in long-term sequelae between regimens but with slightly faster recovery from acute crises in the group receiving both vitamins; thiamine treatment, alone or combined, shows better results than biotin alone [9, 35].

In practice, with subacute encephalopathy and symmetrical basal ganglia lesions, thiamine and biotin should be initiated while awaiting results, alongside measurement of biotinidase activity, and continued until biochemical and/or genetic exclusion [1, 9].

7.3. Ketogenic diet

Ketogenic diet is the reference therapy in primary PDHc deficiency: patients do not efficiently metabolise carbohydrates, and ketone bodies cross the blood-brain barrier, supplying acetyl-CoA downstream of the block [10, 29, 33]. The 2026 consensus recommends classical ketogenic diet, initiated as soon as possible after genetic confirmation regardless of age, allowing a modified ketogenic diet in selected patients (milder disease, later onset or intolerance); the diet may be administered orally, enterally or parenterally according to clinical status; and should not be delayed in favour of a thiamine trial, particularly in patients with seizures and severe lactic acidosis [10]. The same consensus extends the recommendation to PDHB, DLAT, PDHX and PDP1 forms because they share the pathophysiology, although evidence concerns mainly PDHA1 [10].

Biochemical targets differ between sources and this divergence is clinically relevant. GeneReviews associates the best outcomes with maintaining plasma or serum beta-hydroxybutyrate at approximately 3.0–3.5 mEq/L and describes fat-to-carbohydrate-plus-protein ratios between 1:1 (modified diet) and 4:1, with fat proportions between approximately 55% and 80%, without studies demonstrating superiority of any ratio [33]. The more recent 2026 consensus states that there is no specific ketone target in PDHc deficiency: the target is individualised and corresponds to the level at which each patient is clinically stable [10]. Nyhan's Atlas describes that diets with 50% or more of calories as fat and 20% as carbohydrates may normalise blood lactate, but emphasises that patients with established neurological abnormalities do not improve neurologically, and that the only adverse effect observed in reviewed patients was hyperuricaemia [29]. Half of patients with PDHc deficiency-associated epilepsy achieve remission in the first year of ketogenic diet, and developmental benefit is greater in milder, post-neonatal-onset presentations, with better prior functioning and no brain malformations [33]. With infantile spasms or hypsarrhythmia in an infant with PDHc deficiency, the consensus recommends ketogenic diet first line, before corticosteroids [10].

Outside PDHc deficiency, evidence does not support ketogenic diet. GeneReviews explicitly states that there is no evidence of efficacy in other forms of Leigh spectrum, although the diet may be indicated in drug-resistant epilepsy [1, 2]. A systematic review concluded that the diet may be considered in patients with mitochondrial disease and refractory epilepsy, except mtDNA deletion myopathy and pyruvate carboxylase deficiency; a prospective controlled study documented reduced seizure frequency in 76% of patients with mitochondrial disease, with greater efficacy in MELAS [2].

The diet has its own complications requiring surveillance: free carnitine deficiency (supplement if plasma or dried blood spot levels are low), metabolic acidosis (bicarbonate and/or citrate), and renal stones, whose incidence in patients on ketogenic diet is estimated at approximately 5.9% and may be higher in PDHc deficiency because of underlying acidosis; prevention with adequate hydration, potassium citrate to alkalinise urine, regular urinalysis and ultrasound when indicated [10]. Many patients require gastrostomy to achieve and maintain the diet [33].

7.4. Supportive treatment of manifestations

Treatment of manifestations is shared by nuclear and mtDNA forms (Table 7.2) [1, 2].

For seizures, the 2024 InterERNs consensus, developed by 24 experts from five European reference networks using Delphi methodology, achieved high consensus on the safety of 14 of 25 reviewed drugs and endorsed NICE seizure treatment guidelines with modifications; exceptions identified were valproate in POLG disease, vigabatrin in GABA-transaminase deficiency and topiramate in patients at risk of renal tubular acidosis [15]. In PDHc deficiency, no antiseizure medication is absolutely contraindicated, but topiramate should not be first line because of the risk of precipitating acidosis, and caution is required with zonisamide for the same reason; any antiseizure medication should be prescribed in a carbohydrate-free formulation in patients on ketogenic diet [10].

7.5. Acute decompensation, surgery and anaesthesia

Prevention of catabolism is the overarching principle. The Mitochondrial Medicine Society recommends that patients have an emergency document, updated at every consultation, containing diagnosis, needs, medication, allergies and contraindicated therapies, as well as advance directives for critical illness; that they consider a medical alert bracelet; that they avoid prolonged fasting and receive dextrose-containing fluids before, during and after procedures and surgery — with the explicit qualification that dextrose should not be administered, or should be limited, in suspected or confirmed pyruvate metabolism defects, in patients on ketogenic diet and in those with a previous adverse response to high glucose intake [13, 14]. Nutritional support should begin as soon as safe, preferably within the first 48 hours of hospitalisation, to avoid prolonged catabolism; with vasopressor-refractory hypotension, stress-dose corticosteroids should be considered, collecting cortisol and adrenocorticotropic hormone beforehand [14].

In PDHc deficiency, the 2026 consensus defines a specific emergency protocol, with several recommendations reversing usual practice in other mitochondrial diseases [10]:

• Episodes may be precipitated by fasting, insufficient food intake, inappropriately high carbohydrate intake, vomiting, diarrhoea, intense exercise, fever, intercurrent illness, anaesthesia and surgery.
• With clinical suspicion, collect blood before starting fluids: blood gas analysis, glucose, electrolytes, bicarbonate, liver and renal function, lactate, ketones, full blood count, C-reactive protein, blood cultures if febrile, and urinalysis.
• High-dose glucose infusions (dextrose >5%) should be avoided because carbohydrates may precipitate lactic acidosis; use age-appropriate saline fluids and, if there is a risk of hypoglycaemia or prolonged use, add low-concentration dextrose (for example, a final concentration of 2.5%), maintaining glucose ≥2.5 mmol/L (45 mg/dL).
• Ringer's lactate (Hartmann's solution) is contraindicated.
• With metabolic acidosis with pH <7.20 and/or bicarbonate <10 mmol/L and/or base deficit greater than 10 mmol/L, intravenous or enteral bicarbonate (or citrate) may be considered.
• If ketosis exceeds 5 mmol/L during decompensation, avoid intravenous glucose and hydrate with normal saline or balanced crystalloids without dextrose.
• During fasting for procedures or when feeding is not tolerated, monitor glucose, ketones and blood gases every 4 hours; for fasting exceeding 24 hours, institute parenteral nutrition with glucose ≤5%, lipids and amino acids.
• Severe decompensations require intensive care; electroencephalography and magnetic resonance imaging are useful for assessing neurological manifestations.

Perioperatively, slow titration of volatile and parenteral anaesthetics, caution with prolonged propofol use because of lactic acidosis risk, avoidance of neuromuscular blockers in patients with muscle disease (or use under strict monitoring), prevention of body temperature changes and catabolism, and monitoring of acid-base balance, glucose, ketones and lactate are recommended [1, 10]. Anaesthesia may worsen respiratory manifestations and precipitate respiratory failure, so its indication and surveillance before, during and afterwards require careful consideration — one reason to reserve muscle biopsy [1, 2].

7.6. Supplements, “cocktails” and divergence between schools

Prescription of vitamin and cofactor combinations (riboflavin, thiamine, creatine, biotin, CoQ10, carnitine, alpha-lipoic acid) is common practice, but without international consistency or an established evidence base in Leigh spectrum [2]. Saudubray's chapter is explicit: some substances have a rational basis, such as riboflavin in complex I deficiency or thiamine in PDHc deficiency, but many are non-specific antioxidants and no evidence supports or refutes their use, which is why the authors do not even list components [6]. The Mitochondrial Medicine Society takes the opposite position: it recommends offering CoQ10 to most patients with mitochondrial disease, not only primary deficiency, that ubiquinol is the most bioavailable form (with corresponding dose adjustment), that plasma or leucocyte levels help monitor absorption and adherence, that alpha-lipoic acid and riboflavin be offered, that folinic acid be considered in patients with central nervous system manifestations and routinely administered with documented CSF deficiency, and that L-carnitine be administered with documented deficiency [13]. This divergence between American and European practice should be explained to families: the society itself acknowledges that few trials have explored clinical effects of these treatments and that recommendations derive from consensus, and a Cochrane review found little evidence favouring any vitamin or cofactor intervention [6, 13]. In PDHc deficiency, the 2026 consensus is brief: beyond thiamine, evidence supporting other vitamins and cofactors remains limited [10]. Lipoic acid is theoretically useful and has very low toxicity, but studies suggest exogenous lipoic acid is not used for mitochondrial lipoylation [33].

N-acetylcysteine may be considered when blood glutathione measurement reveals deficiency, and sugar-free L-carnitine is used to protect against secondary deficiency associated with prolonged ketogenic diet use [33].

7.7. Drugs and circumstances to avoid

Table 7.3 brings together agents to avoid, with the source of each recommendation. The reference basis is the 2020 international Delphi consensus on drug safety in primary mitochondrial disease, which reviewed 46 drugs with a panel of 16 experts from six countries and emphasises that decisions should be individualised [48].

The Brazilian textbook list — avoid propofol, valproic acid, tetracycline and barbiturates — is consistent with these sources; the same textbook's recommendation to avoid a carbohydrate-rich diet correctly applies to PDHc deficiency, but should not be generalised to the whole spectrum, where prevention of catabolism requires adequate energy provision [5, 10, 13]. It should also be emphasised that the Mitochondrial Medicine Society recommends vaccination of all patients with mitochondrial disease, including influenza vaccination, because the infection-associated worsening risk exceeds the theoretical vaccination risk [14].

7.8. Regulatory situation in the USA and European Union

On 6 October 2026, consultation of official databases confirms that no drug has marketing authorisation for Leigh syndrome or pyruvate dehydrogenase complex deficiency in the United States or European Union (Table 7.4) [11, 46, 49]. Orphan drug designations exist, which are not authorisations: designation recognises disease rarity and grants development incentives but does not attest to efficacy or safety.

The last two rows are included precisely to prevent frequent confusion: these are the first drugs authorised for specific mitochondrial diseases, marking a change in context, but neither has an indication in Leigh spectrum. Automated access to the FDA orphan drug designation database was not possible, so this article does not claim the existence or absence of American designations for Leigh syndrome [46].

7.9. Clinical trials and research

Table 7.5 brings together records consulted on ClinicalTrials.gov on 6 October 2026 [51].

Several points require cautious interpretation. The MIT-E trial did not meet its primary objective and the observed difference was not statistically significant; termination is attributed to sponsor decision rather than a safety signal, but registry death and serious adverse event counts should not be read as adjusted comparative rates [51]. The first two vatiquinone trials specifically targeting Leigh syndrome finished years ago without publication of results, as GeneReviews itself notes [2, 51]. The paediatric sonlicromanol programme is suspended, although the substance received European orphan designation in July 2025 for inherited oxidative phosphorylation defects — illustrating the difference between regulatory status and active clinical development [11, 51].

Among preclinical or exploratory approaches, the ENMC workshop recorded: screening of existing drugs in patient fibroblasts, which identified apomorphine as protective against oxidative stress with greater efficacy than idebenone (EC50 ≈ 100 nM), with prolonged survival and improved rotarod performance in Ndufs4−/− mice (p<0.01) and a patent application, with juvenile animal toxicity studies necessary for progression to paediatric trials under discussion with the Japanese regulatory agency; and screening in induced pluripotent stem cells from patients with MT-ATP6 variants, which identified phosphodiesterase-5 inhibitors as a potential strategy, with five patients apparently showing clinical improvement on compassionate use — with the qualification, explicitly stated by the researchers themselves, that the unpredictable natural history of LSS prevents robust conclusions from compassionate-use data [47]. Mitochondrial genome editing therapies (mitochondria-targeted zinc-finger nucleases and TALENs, cytidine deaminases) aim to shift heteroplasmy towards normal mtDNA, but at the time of the GeneReviews review, gene therapy had not progressed to human clinical trials in Leigh spectrum [2]. Other investigational strategies include agents stimulating mitochondrial biogenesis (bezafibrate, resveratrol, nicotinamide analogues, KL1333) and decanoic acid, a putative active component of ketogenic diet [2].

Finally, the 2026 consensus explicitly recommends informing families about available local and international clinical trials, with links to official registries (ClinicalTrials.gov in the USA, euclinicaltrials.eu in the EU or equivalent) and about patient organisations [10].

8. Reproductive Options and Family Counselling

8.1. Determine mode of transmission before calculating risks

Counselling in Leigh spectrum depends entirely on molecular diagnosis, because the same clinical presentation may reflect autosomal recessive, X-linked, autosomal dominant de novo or mitochondrial DNA maternal inheritance, with recurrence risks ranging from near zero to 25% or more [1–3]. The Oxford Desk Reference describes the problem directly: diagnosis is frequently difficult to establish, affected relatives may already have died, recurrence risks are potentially high, and the variety of transmission patterns adds difficulty, so the first priority is to confirm that the proband truly has a presentation compatible with Leigh encephalopathy [3]. That source also notes that without molecular diagnosis, empirical risks are used: the 33% risk estimated by van Erven and colleagues in 1987, which the Oxford authors themselves consider probably overestimated because it was not corrected for unascertained sibships, and an empirical range between 1 in 8 and 1 in 4 in the largest Leigh disease study available at the time — with the qualification that, if the mother carries an unrecognised mtDNA variant, risk may be far higher [3, 28]. These values should be presented as historical fallback estimates, not as the risk for a specific family with an established molecular diagnosis.

Table 8.1 summarises risks by transmission mode.

In pyruvate dehydrogenase complex deficiency, the distribution of de novo variants markedly influences counselling: approximately 60–63% of males and 85–95% of females with PDHA1 disease have a de novo variant, so the 2026 consensus recommends offering targeted genetic testing to all mothers of patients with a confirmed PDHA1 variant to determine whether the variant is maternal or arose de novo (recommendation G3, class I, level B); in recessive forms (PDHB, DLAT, PDP1, PDHX), parental segregation should be confirmed genetically, and prenatal or preimplantation diagnosis may be discussed with families once causal variants are identified (G4) [10, 33]. If the proband's variant is not detected in the mother, recurrence risk is low but above the general population because of possible germline mosaicism [10]. An additional complexity: because typical variants differ between male and female probands (insertions/deletions in girls, missense in boys), predicting manifestations in a sibling of a different sex is difficult — theoretically, sisters of a male proband with a missense variant may be asymptomatic or more mildly affected, whereas an insertion/deletion variant identified in a female proband may be incompatible with life in a hemizygous male fetus; conclusive data are lacking [1, 33].

Two technical cautions apply to recessive forms. If only one parent is a carrier, with biological maternity and paternity confirmed, consider that one variant arose de novo in the proband or in a parent with postzygotic mosaicism; and if the proband appears homozygous, this may result from a single- or multiexon deletion undetected by sequencing, or isodisomic uniparental disomy [1, 9].

8.2. Heteroplasmy, bottleneck and limits of prediction

In mtDNA forms, counselling confronts four cumulative phenomena: the mitochondrial bottleneck, which varies the proportion of mutated molecules transmitted between generations; variant-specific selection; the threshold effect; and the inherent difficulty of predicting phenotype from mutant load [2]. Families should be explicitly told that specific predictions about an individual's clinical outcome or that of their offspring are not possible [2].

In practice, testing the proband's mother for the identified variant is recommended, preferably in at least two different tissues (for example blood, urinary sediment, buccal cells) [2, 42]. In most cases the mother has far lower heteroplasmy than the child and remains asymptomatic or develops only mild manifestations such as gait disturbance or retinopathy; occasionally she has a substantial load and develops severe disease in adulthood [2]. Because mtDNA variants may be lost from the leucocyte population with age, in adults with mild manifestations and older asymptomatic maternal relatives the variant may be detectable only in hair follicles, urinary sediment or skeletal muscle [2, 3]. If the proband's variant is not identified in maternal tissues, possibilities are a de novo variant and maternal germline mosaicism; sibling risk is then low but not zero [2, 52].

The partial exception is the pair of MT-ATP6 variants at position 8993. For m.8993T>G and m.8993T>C, there is a strong positive relationship between maternal heteroplasmy and predicted recurrence risk, quantified by White and colleagues, although wide 95% confidence intervals limit its usefulness for individual counselling [2, 53]. Furthermore, variant m.8993T>G shows skewed segregation through positive selection, with oocytes and embryos distributed at the extremes, near 0% or approximately 100% [2]. For other variants (m.3243A>G, m.8344A>G, m.9176T>C, m.14459G>A, m.14487T>C, m.10158T>C, m.10191T>C, m.13513G>A), heteroplasmy has predictive value only near 0% or 100%, and empirical risks from retrospective studies are used [2].

It should also be explained that testing an at-risk asymptomatic relative is technically possible once the variant has been identified in the proband, but interpreting the result is extremely difficult: phenotype cannot be predicted and absence of the variant in one tissue, such as blood, does not guarantee absence from other tissues [2]. The ideal time to determine genetic risk and discuss availability of prenatal and preimplantation testing is before pregnancy [2].

8.3. Prenatal diagnosis

In nuclear forms, once causal variants have been identified in the family, prenatal diagnosis and preimplantation genetic testing are possible and follow the principles of other Mendelian diseases [1, 6]. The Oxford Desk Reference states that, once genetic diagnosis is established, chorionic villus sampling is relatively straightforward in families with SURF1 and pyruvate dehydrogenase complex gene variants, and also in nuclear causes of complex I deficiency, although with less availability [3].

In mtDNA forms, prenatal diagnosis consists of quantifying heteroplasmy in fetal tissue obtained by chorionic villus sampling (usually at 11–12 weeks) or amniocentesis (usually at 15–18 weeks) [2]. Technical and interpretative aspects are crucial:

• Analysis should be performed on direct villi or amniocytes, not cultured cells, because heteroplasmy may change during culture [2].
• Available evidence suggests that the proportion of abnormal mtDNA is similar in extraembryonic and embryonic tissues and does not vary substantially during pregnancy; for several variants (m.8993T>G, m.8993T>C and m.9176T>C in MT-ATP6, m.3243A>G in MT-TL1, m.10191T>C in MT-ND3 and m.13513G>A in MT-ND5), the proportion remains relatively stable during embryofetal development and in postnatal tissues. However, variation between placenta and embryonic or postnatal tissues, and between multiple placental samples, has also been described [2].
• There is no consensus between centres on preference for villi or amniocentesis, although some studies support amniocentesis given intraplacental variation described with chorionic villus sampling [2].
• Thresholds used vary in clinical practice. A systematic review suggested a generic threshold of 18%, based on the observation that individuals with muscle heteroplasmy below approximately 18% had a 95% or greater probability of being unaffected, irrespective of variant. Other studies applied low (<30%), intermediate (30–60%) and high (>60%) prenatal risk groups, with the intermediate range constituting a “grey zone” of difficult or impossible interpretation. This stratification is feasible for some variants but may not suit others, such as m.3243A>G [2].
• The main limitation is therefore difficulty predicting clinical outcome from the fetal result; reliable prediction requires close correlation between load and severity, uniform tissue distribution and stable heteroplasmy over time [2].

Prenatal diagnosis is the preferred option when the variant appears to have arisen de novo in the proband and recurrence risk is low, and also in women at low risk because their heteroplasmy is non-zero but low, for example below 20% [2, 52]. For m.8993T>G, despite skewed segregation, prenatal diagnosis remains considered appropriate in women with low heteroplasmy because reliable prediction is possible at most load levels [2].

8.4. Preimplantation genetic testing

Preimplantation genetic testing (PGT) quantifies variant load to select the most suitable embryo for transfer [2]. It is considered the appropriate reproductive option for women with familial heteroplasmic mtDNA variants and probably the best available option for those with moderate recurrence risk; it may be unsuccessful in these women and is not recommended in those with homoplasmic or near-homoplasmic levels [2, 54]. Biopsy was traditionally performed on one or two blastomeres on day three or four, a stage for which robust data correlate blastomere and embryo load; most centres now use trophectoderm biopsy, with limited data in this context, and one study suggesting poorer suitability for estimating embryonic heteroplasmy, an unresolved issue [2].

Embryos should be considered suitable for implantation only if they have a very low proportion of abnormal mtDNA, preferably 0%, although thresholds of 18% may imply low recurrence risk; because disease thresholds and transmission patterns vary between variants, the decision is individual and should explore the residual risk the couple is willing to accept, also considering heteroplasmy loads in affected and unaffected relatives [2, 54]. In women with a high proportion of abnormal mtDNA in blood or urinary sediment, many oocytes may have a high load, and several ovarian stimulation cycles may not produce a suitable embryo; nevertheless, PGT provides valuable information for future reproductive planning [2]. If most embryos have a substantial proportion of abnormal mtDNA, oocyte donation is probably the only option to guarantee an unaffected embryo; conversely, if most have no detectable abnormal mtDNA, parents may opt for prenatal diagnosis in subsequent spontaneous pregnancies [2].

8.5. Oocyte donation and mitochondrial donation

Oocyte donation followed by in vitro fertilisation using the partner's sperm eliminates maternal mtDNA transmission. A point frequently overlooked should be explicit: a maternal relative should not serve as donor, because her oocytes may contain a high proportion of abnormal mtDNA even if leucocytes do not reveal it [2].

Mitochondrial donation (mitochondrial replacement therapy) consists of transferring the nucleus removed from a zygote (pronuclear transfer) or oocyte (maternal spindle transfer) into an enucleated cell at the same stage from an unaffected donor [2]. The legal and clinical framework has evolved differently by jurisdiction: after extensive scientific, ethical and public consultation, the UK government authorised use of the technique in 2016 to prevent transmission of severe mitochondrial disease caused by specific mtDNA variants, regulated by the Human Fertilisation and Embryology Authority, limited to families at substantial transmission risk where PGT is considered unsuitable or likely ineffective, with a five-year clinical trial conducted at Newcastle Fertility Centre; in 2022, the Australian parliament passed the mitochondrial donation reform law, known as Maeve's Law, allowing phased implementation beginning with a clinical trial [2]. The first birth after maternal spindle transfer was reported in 2017, in a child of a woman with the MT-ATP6 m.8993T>G variant; the child was well at seven months, but long-term neurodevelopmental follow-up was not published [2, 55]. Saudubray's chapter adds that concerns remain about long-term efficacy and safety of these techniques [6].

Results from the British programme were published in July 2025 and constitute the most relevant development in this area. In an integrated programme offering PGT to women with heteroplasmy and pronuclear transfer to women with homoplasmy or high heteroplasmy, clinical pregnancies were confirmed in 8 of 22 patients (36%) undergoing intracytoplasmic sperm injection for pronuclear transfer and in 16 of 39 (41%) for PGT; pronuclear transfer resulted in 8 births and one ongoing pregnancy, and PGT in 18 births. Blood heteroplasmy in the 8 newborns in the pronuclear transfer group ranged from undetectable to 16%, with reduction of the maternal pathogenic variant by 95–100% in 6 and 77–88% in 2 compared with corresponding enucleated zygotes; in the 10 infants in the PGT group with known heteroplasmy, values ranged from undetectable to 7% [56]. The clinical report from the same programme describes that 22 women began or completed pronuclear transfer, with 8 births; all children were healthy at birth, with absent or low blood heteroplasmy; one child, whose mother had hyperlipidaemia during pregnancy, developed hyperlipidaemia and cardiac arrhythmia, both responding to treatment, and another developed infantile myoclonic epilepsy with spontaneous remission; at reporting, all had normal development [57]. These data should be presented as initial results from a regulated programme, with few children and short follow-up, rather than demonstrating long-term safety.

The international European Neuromuscular Centre workshops, particularly the 243rd, dedicated to reproductive options for families with maternally transmitted mtDNA disease, form the basis of guidance in this area [2, 54].

8.6. At-risk relatives, carrier detection and pregnancy

In treatable diseases, early identification of relatives has immediate therapeutic consequences. In biotin-thiamine-responsive basal ganglia disease, clarifying the genetic status of apparently asymptomatic older and younger at-risk siblings is appropriate to identify as early as possible those who will benefit from immediate biotin and thiamine initiation and avoidance of sodium valproate and ACTH [9]. Carrier testing in at-risk relatives requires prior identification of familial variants; it should be considered in partners of BTBGD patients and carriers, particularly if consanguinity is possible or both share a geographical origin, given a founder variant in the Saudi population [9].

In nuclear forms without confirmed molecular diagnosis, storing the proband's DNA should be considered because testing methodology and knowledge of genes and mechanisms will continue to improve [1].

Regarding pregnancy, two distinct recommendations apply. In women affected by BTBGD, biotin and thiamine should continue throughout pregnancy; no information is available on fetal risk associated with an affected mother [9]. In women with mitochondrial disease generally, the Mitochondrial Medicine Society recommends closer specialist surveillance, obstetric follow-up experienced in high-risk pregnancy (because of increased preterm birth and pre-eclampsia risk), counselling about potentially increased gestational diabetes risk with early and late glucose tolerance testing, review of ongoing medication (antiseizure medications, antihypertensives), preconception genetic counselling with prenatal testing offered where feasible, and closer fetal surveillance when prenatal manifestations are possible; the same source notes that premature ovarian failure characterises several mitochondrial diseases and that affected women wishing to have children should be referred for assisted reproductive techniques [14]. Available quantitative data derive mainly from carriers of the MELAS-associated m.3243A>G variant: in a retrospective analysis of 96 pregnancies, 25% ended in preterm birth (5.5% extremely preterm), 12% had pre-eclampsia and 11% gestational diabetes, 13% of children had birthweight below the 5th centile and four fetal deaths were reported [14]. These numbers concern a specific variant and carrier population, not the whole Leigh spectrum, and should not be extrapolated without this qualification.

Finally, genetic counselling should be offered to all affected or at-risk young adults, including general discussion of offspring risks and reproductive options, and all families with a genetic diagnosis of PDHc deficiency should be referred for genetic counselling, discussing mode of transmission and implications for other relatives and future pregnancies [2, 10].

9. Special Considerations: Population Diversity and Epidemiology

9.1. Frequency of Leigh spectrum

Leigh spectrum is the phenotype most frequently observed among paediatric-onset mitochondrial diseases [2]. Available frequency estimates derive from few population studies, with different denominators and definitions, and should be read as orders of magnitude (Table 9.1).

None of these estimates is gene-specific. The principle applies generally: known proportions (for example, SURF1 causing approximately 50% of Leigh syndrome with COX deficiency and approximately 30% of nuclear forms, or MT-ATP6 causing approximately 40% of mtDNA forms) allow genetic investigation to be prioritised but do not permit conversion of a proportion into a gene-specific population rate, because validated denominators are lacking [1, 2, 4]. For broader context, the InterERNs consensus estimates primary mitochondrial disease prevalence at approximately 1 in 4300 [15].

National and international registries compiled at the ENMC workshop allow cohort comparisons and show notable agreement on the most frequent genes: SURF1 and complex I-related genes predominate among nuclear genes, and MT-ATP6, MT-ND3 and MT-ND5 among mitochondrial genes; in Italy (122 patients), the United Kingdom (72), China (209), North America (97, North American Mitochondrial Disease Consortium, NAMDC registry), northern Europe (130) and the GENOMIT registry (715 genetically confirmed patients), the most frequent nuclear gene is SURF1 and mitochondrial gene MT-ATP6 [47]. The proportion of patients with elevated blood lactate ranges from 58% to 84% depending on cohort, and elevated CSF lactate from 11% to 95%, a spread reflecting mainly differences in selection of tested patients rather than biological differences between populations [47].

9.2. Founder effects

Some populations have local frequencies far above average for specific genetic forms. Table 9.2 brings together founder effects documented in the sources consulted.

The practical consequence is twofold. First, in a child with a compatible presentation and origin from one of these populations, targeted founder variant testing may be faster and cheaper than broad sequencing, particularly when it identifies a treatable form such as BTBGD [9]. Second, local frequency does not transfer to other populations: Leigh spectrum is pan-ethnic and BTBGD has been described on all continents [9].

9.3. Data from Portugal and Brazil

In Portugal, a national cohort published in 2023 assembled 40 patients with molecularly confirmed Leigh spectrum, 28 with mitochondrial DNA variants and 12 with nuclear variants, including novel variants in a heterogeneous set of genes [60]. This is a national reference laboratory series, not a population prevalence study, and does not permit estimation of disease frequency in Portugal. In 2024, based on the 2023 criteria, the same group proposed a diagnostic approach in which biochemical screening (amino acids, acylcarnitines and urinary organic acids) runs in parallel with genetic investigation; in their series, this approach characterised 80% of cases and enabled specific intervention in 10% [41].

In Brazil, consulted sources include a national reference textbook chapter and two detailed published cases — a child followed in Rio de Janeiro and a 2-year-old child with an MTRFR (C12orf65) variant followed in São Paulo, described in section 10 — but contain no population prevalence data [5, 18, 61]. Absence of data in consulted sources does not demonstrate absence of published disease frequency data in Portugal or Brazil; it means only that this article did not find them in its sources and that no targeted geographical searches were performed.

9.4. Reference networks, registries and patient resources

Leigh spectrum is a rare multisystem disease whose follow-up depends on experienced centres. In Europe, the relevant European reference network is MetabERN, whose PM-MD subnetwork explicitly covers pyruvate metabolism disorders, Krebs cycle defects, mitochondrial oxidative phosphorylation defects and thiamine transport and metabolism disorders — that is, all spectrum causes discussed here [16]. The 2024 consensus on epilepsy in mitochondrial disease was developed by the Mito InterERN Working Group, bringing together experts from five European reference networks (EpiCARE, ERN EYE, ERN-RND, EURO-NMD and MetabERN), a model of inter-network liaison reflecting the multisystem nature of these diseases [15].

Several international registries and consortia are active: GENOMIT, a European Joint Programme on Rare Diseases-funded network with eight partners (Germany, Austria, Italy, France, United Kingdom, Japan and China), whose registry and sequencing database covers 2958 paediatric patients, including 623 with Leigh spectrum, 415 of whom have a genetic diagnosis; the North American Mitochondrial Disease Consortium, a National Institutes of Health-funded collaborative effort encompassing 17 North American centres and maintaining a patient contact registry; the Leigh Syndrome Roadmap Project, an international initiative of the International Leigh Syndrome Consortium (collaboration between UMDF, the Lily Foundation, Mitocon, the Australian Mito Foundation and People Against Leigh Syndrome), which funded eight preclinical trials and one natural history study; the prospective LION study (Leigh syndrome: Investigating Outcome measures and Natural history), designed to quantify disease burden and progression and explore outcome measures that could serve as trial endpoints; and MitoSHARE, a global database promoted by patient associations [47].

Patient organisations referenced by consulted sources include the United Mitochondrial Disease Foundation and MitoAction (United States), Association Contre Les Maladies Mitochondriales (France), Deutsche Gesellschaft für Muskelkranke (Germany), Mitocon (Italy), Mito Foundation (Australia), MitoCanada (Canada), Metabolic Support UK, the Lily Foundation and Climb (United Kingdom), International Mito Patients, People Against Leigh Syndrome, the Freya Foundation (specifically dedicated to pyruvate dehydrogenase complex deficiency), the Muscular Dystrophy Association, Retina International and NORD, as well as registries such as the Massachusetts General Hospital Mitochondrial Disease Registry and Tissue Bank and eyeGENE [1–3, 28, 33]. The 2026 consensus explicitly recommends informing families about local and international patient organisations and offering psychosocial support after diagnosis [10]. This article did not search for Portuguese or Brazilian associations and therefore does not list them — an absence reflecting search methodology rather than non-existence of these organisations.

Patient associations have also assumed a role in setting research priorities. At the ENMC workshop, Lily Foundation representatives presented the example of international Duchenne muscular dystrophy care standards, accredited by NICE, associated with improved care standards and survival rates and a framework enabling global multicentre trials, and called for similar standards for Leigh spectrum [47].

10. Illustrative Clinical Case

10.1. A 2-year-old child with Leigh spectrum due to an MTRFR (C12orf65) variant, São Paulo

At a university centre in São Paulo, Perrone and colleagues described a 2-year-old girl admitted with respiratory failure and drowsiness [61]. She was the first child of healthy consanguineous parents, without relevant family history. Delivery was preterm by caesarean section for fetal distress; birthweight was 1.720 kg, length 41.5 cm and head circumference 29 cm, with Apgar scores of 3 at one minute and 8 at five minutes, requiring neonatal resuscitation. She remained in neonatal care for the first three months of life because of episodes of respiratory distress. Karyotype was normal and initial abdominal and ophthalmological examinations revealed no abnormalities.

Incompletely documented developmental milestones included social smiling and head control at 3 months, unsupported sitting at 10 months, crawling at 11 months, and supported standing and first words at 12 months. After the first year, failure to thrive developed and neurological regression was described: she stopped grasping toys and pointing to objects and people. At 2 years she began having diarrhoea and drowsiness episodes, with respiratory failure at home prompting emergency attendance and, because it did not resolve, paediatric intensive care admission [61].

Examination showed global hypotonia, developmental delay and no dysmorphic features. Serum lactate remained repeatedly elevated throughout admission, with values of 21 mg/dL at admission, 19 mg/dL on day seven, 54 mg/dL on day fourteen, 69 mg/dL on day nineteen and 30 mg/dL on day twenty-four, against a reference range of 4.5 to 14.4 mg/dL. Tandem mass spectrometry and urinary organic acids were normal — an important finding demonstrating that normal metabolic screening does not exclude diagnosis. Ophthalmological examination revealed optic atrophy; echocardiography and abdominal and Doppler ultrasounds were normal. EEG suggested parieto-occipital epileptiform discharges [61].

Brain MRI showed bilateral T2 and FLAIR hyperintense foci in periventricular and peritrigonal white matter, the posterior limb of the internal capsule and peripheral thalamus, accompanied by hyperintensities in the cerebral peduncle, midbrain tegmentum, dorsal pons and medulla, with some areas of restricted diffusion. On reassessment at 2 years and 9 months, prominent sulci and cisterns, compensatory ectasia of the supratentorial and infratentorial ventricular system and diploic thickening suggested progressive cerebral atrophy; T2-weighted sequences showed symmetrical cerebellar hemisphere hyperintensity and bilateral symmetrical hyperintensity of the substantia nigra, cerebellar peduncle, midbrain segment, posterior pons, superior cerebellar peduncle and medulla. Images are reproduced in figure 1 of the open-access article (figure 1 of Perrone and colleagues, 2020) [61].

A combination of vitamins and cofactors — thiamine, riboflavin, coenzyme Q10, biotin and L-carnitine — was initiated empirically without clinically detectable improvement, illustrating in practice the lack of an evidence base discussed in section 7.6. The course was unfavourable, with oxygen dependence, spastic tetraplegia and severe developmental delay [61].

Clinical exome sequencing, performed with the parents' informed consent, identified a novel homozygous 14-base-pair deletion in exon 2 of C12orf65: NM_152269.5:c.207_220del p.(Pro70AsnfsTer28) (MANE Select transcript; the article gives NM_152269 without a version), confirmed by Sanger sequencing and absent from the population databases consulted. The predicted truncated protein would have 96 amino acids instead of the 166 of the normal protein, losing the GGQ domain and the coiled-coil region, which suggests loss of function; the variant was classified as pathogenic according to the American College of Medical Genetics and Genomics criteria [61].

The authors' literature review assembled 27 previously described patients with pathogenic variants in this gene, with phenotypes ranging from spastic paraplegia type 55 to Leigh syndrome. Among these 27 patients, 92.3% had optic atrophy, 80% peripheral neuropathy, 60% spasticity, 52.1% cognitive impairment, 42.8% nystagmus, 38.4% ophthalmoplegia and 20% bulbar dysfunction. The patient described had optic atrophy, bulbar dysfunction (dysphagia and respiratory failure) and neurodevelopmental delay; nystagmus and ophthalmoplegia could not be assessed and spasticity was not present on initial examination. The 11 variants described were predominantly nonsense, splice-site and frameshift variants, all considered null, including a synonymous variant (NM_152269.5:c.282G>A p.(Lys94=), published as p.Lys94Lys) predicted to cause loss of a splice site with retention of intron 2. Variants associated with severe phenotypes more frequently affected the GGQ domain of the protein, which supports a genotype–phenotype correlation [61].

Two study findings have direct care implications. First, median age at symptom onset in reviewed patients was 4 years and median diagnostic age 17 years, a median diagnostic delay of 13 years; in the described patient, first symptoms arose shortly after birth and diagnosis was made two years later. The authors attribute usual delay to variable expressivity, disease rarity, lack of awareness and difficult access to molecular confirmation, and faster diagnosis in their case to classical MRI signs, the elevated serum biomarker and severity prompting active aetiological investigation, as well as rapid exome sequencing access [61].

10.2. A child with Leigh syndrome and optic atrophy, Rio de Janeiro

At the ophthalmology service of Hospital Universitário Clementino Fraga Filho, Universidade Federal do Rio de Janeiro, Roma, Pereira and Dantas described a 10-year-old girl referred by paediatric neurology [18]. Development was normal until 1 year and 3 months, after which she stopped acquiring skills and her mother reported regression of motor and intellectual learning, seizures and sleep changes. There was no family history; her older brother was healthy.

Neurological examination showed spasticity of all four limbs and severe muscle hypotonia, with sustained hip flexion and knee extension, predominantly choreoathetotic movements and dystonia, deep reflexes with lower limb hypertonia, bilateral extensor plantar responses and IX and X cranial nerve palsy; physical appearance is documented in figure 1 of the article (figure 1 of Roma and colleagues, 2008). MRI showed periventricular, periaqueductal and basal ganglia T2 hyperintensity, and non-enhancing T1 hypointense lesions involving the lenticular nuclei and globus pallidus bilaterally, with a symmetrical appearance; CT showed demyelination and cortical atrophy. The plasma aminogram showed increased ammonia, cystine, ornithine, asparagine, citrulline and glutamine, with aminoaciduria, and blood lactate and pyruvate were elevated [18].

Ophthalmological examination showed low-amplitude nystagmus, preserved ocular motility, preserved direct and consensual pupillary light reflexes, normal biomicroscopy and intraocular pressure; direct and indirect ophthalmoscopy showed pale optic discs with clear contours and regular margins, reduced foveolar reflex, choroidosis and mildly reduced vascular calibre, documented on retinal photography reproduced in figure 2 of the article. Visual evoked potentials showed increased P100 wave latency and reduced amplitude, confirming optic atrophy. Visual acuity could not be assessed because of lack of cooperation [18].

The authors emphasise that the patient had the usual syndrome findings — dystonia, choreoathetosis, ataxia, hypotonia, intellectual and psychomotor delay — except cardiac abnormalities, and that although nystagmus is the most common ophthalmological finding, optic atrophy is less frequent and warrants attention to different ocular manifestations. They also note that normal lactate levels do not exclude Leigh syndrome. Diagnosis was established by combining clinical presentation with characteristic neuroimaging, elevated lactate and pyruvate and the aminogram, without molecular confirmation, and the authors state that, because the disease has no cure, treatment consists of palliative measures — anticonvulsants, management of endocrine dysfunction, dietary management and surgical procedures when necessary — emphasising multidisciplinary follow-up by paediatrics, neurology, cardiology, physiotherapy and ophthalmology; the article does not describe the patient's course under treatment [18].

Published in 2008, this case documents practice before widespread next-generation sequencing and allows comparison with the previous case: the same clinical entity, but without molecular diagnosis, specific genetic counselling or identification of a possible treatable form. It is precisely this difference that justifies the current recommendation for comprehensive genetic investigation in all patients with a compatible presentation [1, 10]. It should also be noted that the 2008 article mentions carbohydrate loading-induced hyperlactataemia among suggestive investigations — a test the 2026 international consensus for PDHc deficiency considers unnecessary and recommends avoiding, allowing it only where genetic testing is inaccessible or prohibitively costly [10, 18].

11. Prognosis and Long-Term Follow-Up

11.1. Natural history and mortality

Prognosis is generally poor. Most patients alternate episodic deterioration with stabilisation plateaux, with stepwise progressive neurological decline, and up to 50% die before 3 years, most frequently from respiratory or cardiac failure, neurological deterioration or sepsis; in patients still undiagnosed, death may appear sudden and unexpected [1, 2]. Progression is nevertheless variable: some patients have a slow course, with stable periods lasting years between decompensations, and survival into adulthood occasionally occurs [1].

Quantitative data derive from retrospective hospital cohorts with referral bias rather than prospective population studies; the ENMC workshop's observation that prospective natural history studies in Leigh spectrum are lacking, motivating the LION study, remains applicable [47]. Table 11.1 summarises the main series.

Three consistent conclusions emerge from these cohorts. Early onset is the most robust poor-prognosis predictor, followed by brainstem involvement, particularly apnoeas [1, 12, 47, 62]. Survival into the second or third decade is documented in specific diseases, including MECR and MTFMT forms [1]. And early identification of a treatable form radically changes trajectory, as demonstrated by 100% 3-year survival in ECHS1 and SLC19A3 forms in the Beijing cohort [63].

One point requires caution in counselling: reports on prognosis of specific genetic causes are contradictory, complicating appropriate family counselling. For SURF1, Ogawa and colleagues and Wedatilake and colleagues described better outcomes, whereas Stenton and colleagues and Lim and colleagues described poorer prognosis [1]. The 2026 consensus states the rule explicitly for PDHc deficiency, but it is generalisable: prognosis should not be based on genotype alone (recommendation D7, class I, level B) [10].

The Brazilian reference textbook presents its own estimates, which should be read as such rather than cohort data: it states that patients with mitochondrial complex IV deficiency and pyruvate dehydrogenase deficiency tend to have poorer prognosis and die within a few years, that children die 3 to 10 years after symptom onset, that patients with partial deficiencies may live to 6–7 years and some to mid-adolescence, that thiamine-responsive forms may have longer survival, that most develop seizures and dysphagia and become wheelchair-dependent, and that death generally occurs from pneumonia or severe infection [5]. The last statement, placing pneumonia-related death in the first year of life, is more pessimistic than international cohorts and should be interpreted in light of Table 11.1. The Oxford Desk Reference concordantly describes progressive neurodegenerative disease punctuated by regression episodes, with brainstem dysfunction (dysphagia, aspiration pneumonia, hypopnoea), dystonia, seizures and recurrent metabolic acidosis representing substantial disease burden and markedly shortened life, with respiratory failure a common outcome [3].

11.2. Surveillance

Patients should be assessed at regular intervals, typically every 6 or 12 months, to monitor progression of known manifestations and emergence of new ones; surveillance may be guided by knowledge of gene-associated phenotypes [1]. Table 11.2 brings together surveillance plans from the two GeneReviews reviews, with additions from the Mitochondrial Medicine Society care standards and the 2026 PDHc deficiency consensus.

In pyruvate dehydrogenase complex deficiency, the 2026 consensus defines specific laboratory intervals adjustable to each patient: in newborns and infants, every 3 months or at least every 6 months depending on metabolic stability; in children under 12 years, every 6 to 12 months; in adolescents and adults, annually. Core parameters include full blood count with differential, electrolytes, transaminases and liver function, renal function (urea, creatinine, cystatin C and cystatin-estimated glomerular filtration rate in patients with stones, reduced renal function or reduced muscle mass), lactate, complete blood gas analysis, plasma amino acids, ferritin and micronutrients; in patients on ketogenic diet add beta-hydroxybutyrate, lipase (and amylase), lipid profile, free and total carnitine and urinalysis [10]. GeneReviews for PDHc deficiency recommends neurological examination and developmental assessment every 6 to 12 months, repeating MRI and EEG only with new symptoms, annual ophthalmological examination, and serum bicarbonate and lactate measurement whenever there is intercurrent illness or metabolic stress and when initiating or modifying ketogenic diet; in patients with DLD variants it adds hepatic surveillance, including physical examination, imaging and liver function tests, noting a theoretical hepatocellular carcinoma risk for which annual alpha-fetoprotein and ultrasound may be considered, although no cases have been described in this disease [33]. With developmental regression at any age, the 2026 consensus recommends full assessment for a potentially treatable complication, as well as evaluation of progression and treatment review, including repeat neuroimaging and EEG when subclinical seizures are suspected [10].

11.3. Transition to adulthood and palliative care

Prolonged survival in some forms makes an organised transition to adult care necessary. The 2026 consensus recommends that patients reaching adulthood be followed by an adult team, preferably experienced in the disease or in neurometabolic diseases, including a dietitian familiar with the ketogenic diet; that transition planning begin at least two years before the expected transfer; that there be a transition plan with goals and emergency instructions, reviewed at least annually; that patients be progressively involved in decisions, according to their abilities, and receive practical training to increase independence (for example, medication reminder apps, pill boxes, guidance on ketogenic diet-compatible shopping); that there be, whenever possible, at least one joint paediatric and adult consultation; that no patient be discharged from paediatric services before being seen by adult services; that the service responsible for metabolic emergencies be designated; and that, in the period immediately after transition, paediatricians and adult physicians consult each other on acutely ill patients, with responsibility for acute management formally assigned to the adult team. For patients at the end of life, continuity within paediatric services may be considered, to minimise additional stress for the patient and family [10].

Palliative care should be integrated early in the clinical course, not only terminally: the consensus recommends an advance plan including care goals, anticipated complications and patient and family preferences, and decisions on invasive ventilation through tracheostomy — sometimes necessary in patients with recurrent aspiration pneumonia, respiratory failure or recurrent central apnoeas — should be discussed beforehand and incorporated into that plan, including defining intervention limits [10]. GeneReviews similarly recommends ongoing assessment of the need for palliative care and home nursing involvement [1, 2].

12. Summary and Final Considerations

Leigh syndrome has evolved from a neuropathological entity into a spectrum defined by a neuroradiological and biochemical pattern shared by more than 115 nuclear genes and 15 mitochondrial genes, united by failure of mitochondrial energy production [1, 2, 8]. Diagnosis rests on explicit criteria — typical neuroimaging, at least one neurological criterion and at least one biochemical or mitochondrial abnormality — and is completed by identification of causal variant(s), preferably through genome sequencing analysing both genomes simultaneously [1, 2, 8]. Approximately 75% of patients develop manifestations in the first 24 months, with stepwise regression triggered by infection, surgery, anaesthesia or fasting, signs of brainstem and basal ganglia dysfunction, and variable lactate elevation [1, 2].

The most important clinical message is rapid identification of treatable forms, because this is the only decision substantially altering trajectory. With subacute encephalopathy and symmetrical basal ganglia lesions, thiamine and biotin should be initiated while awaiting results, and biotinidase activity measured in every case: in biotin-thiamine-responsive basal ganglia disease, manifestations typically resolve within days and treatment within the first month significantly improves outcome; in pyruvate dehydrogenase complex deficiency, ketogenic diet is the reference therapy and should not be delayed in favour of a thiamine trial; in primary coenzyme Q10 deficiencies, early supplementation may limit progression [1, 9, 10, 47]. The Beijing cohort quantifies the difference: 3-year survival of 100% in ECHS1 and SLC19A3 forms versus 50% or less in MT-ND5, SURF1, ALDH5A1 forms and two MT-ATP6 variants [63].

For other forms, treatment is supportive, multidisciplinary and organised around three principles: prevent catabolism, actively treat organ-specific manifestations and avoid agents worsening mitochondrial dysfunction, foremost valproate and lactate-containing solutions [1, 2, 13, 48]. Practice differs between schools regarding supplements: the Mitochondrial Medicine Society recommends offering coenzyme Q10, alpha-lipoic acid and riboflavin to most patients, while European sources emphasise absence of an evidence base; this divergence should be explained to families rather than omitted [2, 6, 13]. On 6 October 2026 no drug is authorised for Leigh syndrome in the United States or European Union; recent European orphan designations exist for biotin in biotin-thiamine-responsive basal ganglia disease (September 2025) and sonlicromanol in inherited oxidative phosphorylation defects (July 2025), but designation is not authorisation, and the paediatric sonlicromanol programme is suspended [11, 51]. The largest completed randomised trial in the field, with vatiquinone in patients with mitochondrial disease-associated epilepsy, did not meet its primary objective [51].

Genetic counselling depends entirely on the identified transmission mode and requires recognition of predictive limits. In autosomal recessive forms, risk is 25% per pregnancy; in PDHA1 X-linked forms, most variants are de novo (60–63% in males, 85–95% in females), with residual risk because of possible germline mosaicism; in reported autosomal dominant forms, all variants were de novo, with estimated recurrence risk of 1%; and in mitochondrial DNA forms approximately one quarter of probands have a de novo variant, and the mother should be tested in at least two tissues [1, 2, 10, 33, 52]. Prenatal diagnosis uses direct cells, never culture; preimplantation testing is suitable for familial heteroplasmy but not homoplasmic variants; oocyte donation should not use a maternal relative; and mitochondrial donation, legally established in the United Kingdom since 2016 and Australia since 2022, produced the first published results of a regulated programme in 2025, with eight births after pronuclear transfer and reductions of 77% to 100% in maternal variant load [2, 56, 57].

Prognosis remains guarded: up to 50% of patients die before 3 years, early onset and brainstem involvement are the most consistent poor-outcome predictors, and progression documented by Newcastle scales is rapid, with doubling of the proportion of children with severe disease burden in just over two and a half years [1, 12, 37]. Nevertheless, survival to adulthood occurs in specific forms, and reports on prognosis of particular genes are contradictory, so counselling should not rely on genotype alone [1, 10]. Regular surveillance every 6 to 12 months, with neurological, nutritional, ophthalmological, hearing, cardiac, hepatic, renal and endocrine assessment, and early integration of palliative care and transition planning, constitute current care standards [1, 2, 10, 14].

Two structural gaps remain and should be acknowledged. First is the absence of international care standards for Leigh spectrum overall: guidelines exist for pyruvate dehydrogenase complex deficiency (2026), epilepsy in mitochondrial disease (2024), drug safety (2020) and care standards in primary mitochondrial disease (2017), but not LSS as an entity, a gap patient associations formally raised at the ENMC workshop [10, 14, 15, 47, 48]. Second is the scarcity of prospective natural history studies and validated outcome measures, remaining the main obstacle to clinical trial design; international registries and ongoing prospective studies aim precisely to fill it [47]. Until then, the essentials offered to a family are what this article has sought to systematise: rapid comprehensive diagnosis, active exclusion of treatable forms, competent multidisciplinary support, informed avoidance of risk agents, genetic counselling precise about what is known and what cannot be predicted, and timely access to palliative care.

13. References

1. Rahman S, Thorburn DR, Ball M. Nuclear Gene-Encoded Leigh Syndrome Spectrum Overview. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Published on 1 October 2015; full update on 1 May 2025. NBK320989. Accessed on 6 October 2026.
2. Ball M, Thorburn DR, Rahman S. Mitochondrial DNA-Associated Leigh Syndrome Spectrum. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Published in 2003; full update on 9 May 2024. NBK1173. Accessed on 6 October 2026.
3. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Leigh encephalopathy, p. 474–475.
4. GeneGuidance. Database export of 26 September 2026 for Leigh spectrum-associated genes (SURF1, MT-ATP6, PDHA1, SLC19A3, LRPPRC, NDUFS4, NDUFV1 and others) and associated conditions. Accessed on 6 October 2026.
5. Carneiro AG, Sampaio MV, Ribeiro EM. Síndrome de Leigh. In: Ribeiro EM, coordinator. Síndromes Genéticas: Reconhecer para Tratar. Fortaleza: EdUnichristus; 2021. Chapter 6.4, p. 130–133.
6. Rahman S, Mayr JA. Disorders of Oxidative Phosphorylation. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 10, sections 10.1–10.5.6.
7. Orphanet. Leigh syndrome. ORPHA:506 [Internet]. Paris: INSERM. Epidemiology record: birth prevalence of 2.8/100 000 in Europe, point prevalence of 2.0/100 000 in Europe (European Medicines Agency, 2011) and annual incidence of 1.3/100 000 in Australia. Disease entry. Accessed on 6 October 2026.
8. McCormick EM, Keller K, Taylor JP, Coffey AJ, Shen L, Krotoski D, et al. Expert panel curation of 113 primary mitochondrial disease genes for the Leigh syndrome spectrum. Ann Neurol. 2023;94(4):696–712. doi:10.1002/ana.26716. PMID:37255483.
9. Tabarki B, Al-Hashem A, Ortigoza-Escobar JD, Alsharhan H, Alfadhel M. Biotin-Thiamine-Responsive Basal Ganglia Disease. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Published in 2013; updated on 9 January 2025. NBK169615. Accessed on 6 October 2026.
10. Keshavan N, Neugebauer J, Bellusci M, Bertini E, Brown G, Darin N, et al. Consensus guidelines for diagnosis and management of pyruvate dehydrogenase complex deficiency. J Inherit Metab Dis. 2026;49(6):e70226. doi:10.1002/jimd.70226. PMID:42823914. PMCID:PMC13631685. Published on 2 October 2026; full text accessed on 6 October 2026.
11. European Medicines Agency. Orphan drug designation report (official dataset, extraction of 6 October 2026): EU/3/25/3172 (biotin, biotin-thiamine-responsive basal ganglia disease, designation on 12 September 2025), EU/3/25/3083 (sonlicromanol hydrochloride, 18 July 2025), EU/3/23/2800 (cannabidiol, 25 July 2023), EU/3/23/2831 (sildenafil citrate, 13 October 2023), EU/3/23/2867 (methyl 4-(2-acetamidoethylsulfanyl)-4-oxobutanoate, 13 December 2023), EU/3/15/1581 (sodium phenylbutyrate, 11 November 2015), EU/3/11/937, EU/3/14/1336 and EU/3/21/2531 (withdrawn). EMA orphan designations. Accessed on 6 October 2026.
12. Sofou K, De Coo IF, Isohanni P, Ostergaard E, Naess K, De Meirleir L, et al. A multicenter study on Leigh syndrome: disease course and predictors of survival. Orphanet J Rare Dis. 2014;9:52. doi:10.1186/1750-1172-9-52. PMID:24731534.
13. Parikh S, Goldstein A, Koenig MK, Scaglia F, Enns GM, Saneto R, et al. Diagnosis and management of mitochondrial disease: a consensus statement from the Mitochondrial Medicine Society. Genet Med. 2015;17(9):689–701. doi:10.1038/gim.2014.177. PMID:25503498.
14. Parikh S, Goldstein A, Karaa A, Koenig MK, Anselm I, Brunel-Guitton C, et al. Patient care standards for primary mitochondrial disease: a consensus statement from the Mitochondrial Medicine Society. Genet Med. 2017;19(12). doi:10.1038/gim.2017.107. PMID:28749475.
15. Mancuso M, Papadopoulou MT, Ng YS, Ardissone A, Bellusci M, Bertini E, et al. Management of seizures in patients with primary mitochondrial diseases: consensus statement from the InterERNs Mitochondrial Working Group. Eur J Neurol. 2024;31(7):e16275. doi:10.1111/ene.16275. PMID:38576261.
16. MetabERN — European Reference Network for Hereditary Metabolic Disorders. PM-MD subnetwork: pyruvate metabolism disorders, Krebs cycle defects, mitochondrial oxidative phosphorylation disorders and thiamine transport and metabolism disorders. metab.ern-net.eu. Accessed on 6 October 2026.
17. Leigh D. Subacute necrotizing encephalomyelopathy in an infant. J Neurol Neurosurg Psychiatry. 1951;14(3):216–221. doi:10.1136/jnnp.14.3.216. PMID:14874135.
18. Roma AC, Pereira PRAA, Dantas AM. Síndrome de Leigh: relato de caso. Arq Bras Oftalmol. 2008;71(1):118–121. doi:10.1590/S0004-27492008000100026. Full text and figures 1 and 2 accessed on 6 October 2026.
19. Nyhan WL, Hoffmann GF, Al-Aqeel AI, Barshop BA. Atlas of Inherited Metabolic Diseases. 4th ed. Boca Raton: CRC Press; 2020. Chapter 47 — Introduction to lactic acidemias, p. 357–359.
20. Holt IJ, Harding AE, Petty RK, Morgan-Hughes JA. A new mitochondrial disease associated with mitochondrial DNA heteroplasmy. Am J Hum Genet. 1990;46(3):428–433. PMID:2137962.
21. Nyhan WL, Hoffmann GF, Al-Aqeel AI, Barshop BA. Atlas of Inherited Metabolic Diseases. 4th ed. Boca Raton: CRC Press; 2020. Chapter 53 — Neurodegeneration, ataxia and retinitis pigmentosa (NARP), p. 399–403.
22. Tatuch Y, Christodoulou J, Feigenbaum A, Clarke JT, Wherret J, Smith C, et al. Heteroplasmic mtDNA mutation (T→G) at 8993 can cause Leigh disease when the percentage of abnormal mtDNA is high. Am J Hum Genet. 1992;50(4):852–858. PMID:1550128.
23. Santorelli FM, Shanske S, Macaya A, DeVivo DC, DiMauro S. The mutation at nt 8993 of mitochondrial DNA is a common cause of Leigh's syndrome. Ann Neurol. 1993;34(6):827–834. doi:10.1002/ana.410340612. PMID:8250532.
24. Zhu Z, Yao J, Johns T, Fu K, De Bie I, Macmillan C, et al. SURF1, encoding a factor involved in the biogenesis of cytochrome c oxidase, is mutated in Leigh syndrome. Nat Genet. 1998;20(4):337–343. doi:10.1038/3804. PMID:9843204.
25. Tiranti V, Hoertnagel K, Carrozzo R, Galimberti C, Munaro M, Granatiero M, et al. Mutations of SURF-1 in Leigh disease associated with cytochrome c oxidase deficiency. Am J Hum Genet. 1998;63(6):1609–1621. doi:10.1086/302150. PMID:9837813.
26. Mootha VK, Lepage P, Miller K, Bunkenborg J, Reich M, Hjerrild M, et al. Identification of a gene causing human cytochrome c oxidase deficiency by integrative genomics. Proc Natl Acad Sci U S A. 2003;100(2):605–610. doi:10.1073/pnas.242716699. PMID:12529507.
27. Lake NJ, Compton AG, Rahman S, Thorburn DR. Leigh syndrome: one disorder, more than 75 monogenic causes. Ann Neurol. 2016;79(2):190–203. doi:10.1002/ana.24551. PMID:26506407.
28. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Mitochondrial DNA diseases, p. 490–494.
29. Nyhan WL, Hoffmann GF, Al-Aqeel AI, Barshop BA. Atlas of Inherited Metabolic Diseases. 4th ed. Boca Raton: CRC Press; 2020. Chapter 50 — Deficiency of the pyruvate dehydrogenase complex, p. 376–384.
30. Brivet M, Gaignard P, Schiff M. Pyruvate Dehydrogenase Complex (PDHC) Deficiency. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 11.3, sections 11.3.1–11.3.5.
31. National Center for Biotechnology Information. ClinVar: variants in MT-ATP6, MT-ND3, MT-ND5, MT-ND6, MT-TK (NC_012920.1), SURF1 (NM_003172.4), SLC19A3 (NM_025243.4), LRPPRC (NM_133259.4), PDHX (NM_003477.3) and DLD (NM_000108.5). ClinVar. Accessed through E-utilities on 6 October 2026.
32. Ng YS, Martikainen MH, Gorman GS, Blain A, Bugiardini E, Bunting A, et al. Pathogenic variants in MT-ATP6: a United Kingdom-based mitochondrial disease cohort study. Ann Neurol. 2019;86(2):310–315. doi:10.1002/ana.25525. PMID:31187502.
33. Ganetzky R, McCormick EM, Falk MJ. Primary Pyruvate Dehydrogenase Complex Deficiency Overview. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Published on 17 June 2021. NBK571223. Accessed on 6 October 2026.
34. Wedatilake Y, Brown RM, McFarland R, Yaplito-Lee J, Morris AA, Champion M, et al. SURF1 deficiency: a multi-centre natural history study. Orphanet J Rare Dis. 2013;8:96. doi:10.1186/1750-1172-8-96. PMID:23829769.
35. Brown G, Plecko B. Disorders of Thiamine (vitamin B1) Metabolism. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 29.1, sections 29.1.1–29.1.7.
36. Blank AC, Breur JMPJ, Fuchs SA, Koop K, Baas AF. Mitochondrial Cardiomyopathies. In: Baars HF, van der Smagt JJ, Doevendans PAFM, van Tintelen JP, editors. Clinical Cardiogenetics. 3rd ed. Cham: Springer; 2020. Chapter 11, section 11.3, p. 179–183.
37. Lim AZ, Ng YS, Blain A, Jiminez-Moreno C, Alston CL, Nesbitt V, et al. Natural history of Leigh syndrome: a study of disease burden and progression. Ann Neurol. 2022;91(1):117–130. doi:10.1002/ana.26260. PMID:34716721.
38. Carli S, Levarlet A, Diodato D, Bertini ES, Martinelli D, Malandrini A, et al. Natural history of patients with mitochondrial ATPase deficiency due to pathogenic variants of MT-ATP6 and MT-ATP8. Neurology. 2025;104(7):e213462. doi:10.1212/WNL.0000000000213462. PMID:40112238.
39. Debray FG, Morin C, Janvier A, Villeneuve J, Maranda B, Laframboise R, et al. LRPPRC mutations cause a phenotypically distinct form of Leigh syndrome with cytochrome c oxidase deficiency. J Med Genet. 2011;48(3):183–189. doi:10.1136/jmg.2010.081976. PMID:21266382.
40. Larson AA, Balasubramaniam S, Christodoulou J, Burrage LC, Marom R, Graham BH, et al. Biochemical signatures mimicking multiple carboxylase deficiency in children with mutations in MT-ATP6. Mitochondrion. 2019;44:58–64. doi:10.1016/j.mito.2018.01.001. PMID:29307858.
41. Baldo MS, Azevedo L, Coelho MP, Martins E, Vilarinho L. A comprehensive approach to the diagnosis of Leigh syndrome spectrum. Diagnostics (Basel). 2024;14(19):2133. doi:10.3390/diagnostics14192133. PMID:39410537.
42. Mavraki E, Labrum R, Sergeant K, Alston CL, Woodward C, Smith C, et al. Genetic testing for mitochondrial disease: the United Kingdom best practice guidelines. Eur J Hum Genet. 2023;31(2):148–163. doi:10.1038/s41431-022-01249-w. PMID:36513735.
43. National Center for Biotechnology Information and European Bioinformatics Institute. MANE Select, version 1.5 (file MANE.GRCh38.v1.5.summary). Reference transcripts of Leigh spectrum nuclear genes. Accessed on 6 October 2026.
44. Bedoyan JK, Hage R, Shin HK, Linard S, Ferren E, Ducich N, et al. Utility of specific amino acid ratios in screening for pyruvate dehydrogenase complex deficiencies and other mitochondrial disorders associated with congenital lactic acidosis and newborn screening prospects. JIMD Rep. 2020;56(1):70–81. doi:10.1002/jmd2.12153. PMID:33204598.
45. Peretz RH, Ah Mew N, Vernon HJ, Ganetzky RD. Prospective diagnosis of MT-ATP6-related mitochondrial disease by newborn screening. Mol Genet Metab. 2021;134(1–2):37–42. doi:10.1016/j.ymgme.2021.06.007. PMID:34176718.
46. U.S. Food and Drug Administration. openFDA databases (Drugs@FDA and labels): NDA 215244, Forzinity (elamipretide), priority approval on 19 September 2025, indicated in Barth syndrome; NDA 219792, Kygevvi (doxecitine and doxribtimine), priority approval on 3 November 2025, indicated in thymidine kinase 2 deficiency. Search by indication “Leigh” without results. api.fda.gov. Accessed on 6 October 2026.
47. Diodato D, Schiff M, Cohen BH, Bertini E, Rahman S; Workshop participants. 258th ENMC international workshop: Leigh syndrome spectrum — genetic causes, natural history and preparing for clinical trials, 25–27 March 2022, Hoofddorp, Amsterdam, The Netherlands. Neuromuscul Disord. 2023;33(8):700–709. doi:10.1016/j.nmd.2023.06.002. PMID:37541860. Full text accessed on 6 October 2026.
48. De Vries MC, Brown DA, Allen ME, Bindoff L, Gorman GS, Karaa A, et al. Safety of drug use in patients with a primary mitochondrial disease: an international Delphi-based consensus. J Inherit Metab Dis. 2020;43(4):800–818. doi:10.1002/jimd.12196. PMID:32030781.
49. European Medicines Agency. Human medicines dataset (extraction of 6 October 2026). Kygevvi (doxecitine and doxribtimine), opinion adopted on 29 January 2026 and marketing authorisation on 26 March 2026, marketing authorisation holder UCB Pharma, for thymidine kinase 2 deficiency. EMA medicine data. Accessed on 6 October 2026.
50. PTC Therapeutics, Inc. PTC Therapeutics receives complete response letter for vatiquinone NDA. Press release of 19 August 2025, filed with the U.S. Securities and Exchange Commission as exhibit 99.1 to form 8-K. SEC EDGAR. Accessed on 6 October 2026.
51. ClinicalTrials.gov, application programming interface v2. Records NCT01721733, NCT02352896, NCT02616484, NCT03747328, NCT04378075, NCT04846036, NCT05218655, NCT06843811 and NCT06990984, including posted results for NCT04378075 and NCT05218655. clinicaltrials.gov. Accessed on 6 October 2026.
52. Sallevelt SCEH, de Die-Smulders CEM, Hendrickx ATM, Hellebrekers DMEI, de Coo IFM, Alston CL, et al. De novo mtDNA point mutations are common and have a low recurrence risk. J Med Genet. 2017;54(2):73–83. doi:10.1136/jmedgenet-2016-103876. PMID:27450679.
53. White SL, Collins VR, Wolfe R, Cleary MA, Shanske S, DiMauro S, et al. Genetic counseling and prenatal diagnosis for the mitochondrial DNA mutations at nucleotide 8993. Am J Hum Genet. 1999;65(2):474–482. doi:10.1086/302488. PMID:10417290.
54. Poulton J, Steffann J, Burgstaller J, McFarland R; workshop participants. 243rd ENMC international workshop: developing guidelines for management of reproductive options for families with maternally inherited mtDNA disease, Amsterdam, the Netherlands, 22–24 March 2019. Neuromuscul Disord. 2019;29(9):725–733. doi:10.1016/j.nmd.2019.08.004. PMID:31501000.
55. Zhang J, Liu H, Luo S, Lu Z, Chávez-Badiola A, Liu Z, et al. Live birth derived from oocyte spindle transfer to prevent mitochondrial disease. Reprod Biomed Online. 2017;34(4):361–368. doi:10.1016/j.rbmo.2017.01.013. PMID:28385334.
56. Hyslop LA, Blakely EL, Aushev M, Marley J, Takeda Y, Pyle A, et al. Mitochondrial donation and preimplantation genetic testing for mtDNA disease. N Engl J Med. 2025;393(5):438–449. doi:10.1056/NEJMoa2415539. PMID:40673696.
57. McFarland R, Hyslop LA, Feeney C, Pillai RN, Blakely EL, Moody E, et al. Mitochondrial donation in a reproductive care pathway for mtDNA disease. N Engl J Med. 2025;393(5):461–468. doi:10.1056/NEJMoa2503658. PMID:40689593.
58. Lee N, Daly MJ, Delmonte T, Lander ES, Xu F, Hudson TJ, et al. A genomewide linkage-disequilibrium scan localizes the Saguenay-Lac-Saint-Jean cytochrome oxidase deficiency to 2p16. Am J Hum Genet. 2001;68(2):397–409. doi:10.1086/318197. PMID:11156535.
59. Piro E, Serra G, Antona V, Giuffrè M, Giorgio E, Sirchia F, et al. Novel LRPPRC compound heterozygous mutation in a child with early-onset Leigh syndrome French-Canadian type: case report of an Italian patient. Ital J Pediatr. 2020;46(1):140. doi:10.1186/s13052-020-00903-7. PMID:32972427.
60. Baldo MS, Nogueira C, Pereira C, Janeiro P, Ferreira S, Lourenço CM, et al. Leigh syndrome spectrum: a Portuguese population cohort in an evolutionary genetic era. Genes (Basel). 2023;14(8):1536. doi:10.3390/genes14081536. PMID:37628588.
61. Perrone E, Cavole TR, Oliveira MG, Virmond LDA, Silva MFB, Soares MFF, et al. Leigh syndrome in a patient with a novel C12orf65 pathogenic variant: case report and literature review. Genet Mol Biol. 2020;43(2):e20180271. doi:10.1590/1678-4685-GMB-2018-0271. PMID:32478789. PMCID:PMC7263430. Full text and figure 1 accessed on 6 October 2026.
62. Ogawa E, Fushimi T, Ogawa-Tominaga M, Shimura M, Tajika M, Ichimoto K, et al. Mortality of Japanese patients with Leigh syndrome: effects of age at onset and genetic diagnosis. J Inherit Metab Dis. 2020;43(4):819–826. doi:10.1002/jimd.12218. PMID:31967322.
63. Stenton SL, Zou Y, Cheng H, Liu Z, Wang J, Shen D, et al. Leigh syndrome: a study of 209 patients at the Beijing Children's Hospital. Ann Neurol. 2022;91(4):466–482. doi:10.1002/ana.26313. PMID:35094435.