Primary Coenzyme Q10 Deficiency

Class of disorders of ubiquinone biosynthesis: COQ10D1–COQ10D9, nephrotic syndrome type 9, spastic ataxia 10 and COQ7-related distal motor neuronopathy

Literature and regulatory information updated: 1 November 2026.

0. Initial Reference Table

Entities of the class

Sources for the table: [1, 2, 6, 10–22, 24, 25, 31–34].

1. Introduction

1.1. Coenzyme Q10 and the concept of primary deficiency

Coenzyme Q10, also called ubiquinone, is a lipid of the inner mitochondrial membrane formed by a benzoquinone ring and a tail of ten isoprene units. In the respiratory chain it acts as a mobile electron carrier: it receives electrons from complexes I (NADH-ubiquinone oxidoreductase) and II (succinate-ubiquinone oxidoreductase), from electron transfer flavoprotein (ETF) and from dihydroorotate dehydrogenase, and delivers them to complex III, from where they pass via cytochrome c to complex IV [3, 35, 41]. Outside the respiratory chain, CoQ10 is an antioxidant in membranes and lipoproteins, links energy production to pyrimidine synthesis and takes part in the regulation of the permeability transition pore and of thermogenesis [35]. All cells depend on endogenous synthesis, because dietary CoQ10 does not effectively replace the cell's own production [42].

GeneReviews defines primary CoQ10 deficiency as the group of conditions with reduced CoQ10 in tissues or cultured cells associated with biallelic variants in one of the genes involved in the biosynthesis of the molecule, and classifies it among the mitochondrial respiratory chain disorders [1]. Secondary deficiency, by contrast, results from variants in genes that do not belong directly to the synthetic pathway, such as respiratory chain defects, multiple acyl-CoA dehydrogenase deficiency (MADD) and ataxia with oculomotor apraxia type 1 (AOA1), or from external factors such as statins [1, 42]. The distinction matters in practice: only molecular testing separates the two situations, and primary deficiency is considered the most readily treatable subgroup of mitochondrial diseases [1, 35].

1.2. Historical milestones

The first description is by Ogasahara and colleagues, in 1989: two sisters, aged 14 and 12 years, with progressive muscle weakness, abnormal fatigability, central nervous system dysfunction since childhood and lactic acidaemia. Isolated complexes I to IV were normal or increased, but the combined I-III and II-III activities, which depend on CoQ10, were low, and mitochondrial CoQ10 content in muscle was 3.7% of the mean of ten controls, with normal values in serum and fibroblasts; the authors concluded that this was a tissue-specific defect of biosynthesis [43]. In 2000, Rötig and colleagues described siblings with severe encephalomyopathy and renal failure, with generalised CoQ10 deficiency in muscle, lymphocytes and fibroblasts, and documented substantial improvement over three years with oral ubidecarenone at a dose of 5 mg/kg/day [44]. The 2012 consensus of the 176th workshop of the European Neuromuscular Centre (ENMC) summarised the pre-molecular era: six cases from four families with recurrent rhabdomyolysis, seizures and mental retardation, with residual muscle CoQ10 of 4% to 16% of controls, and multisystem infantile cases, usually lethal without treatment [35].

The first molecular cause was identified in 2006 by Quinzii and colleagues: a homozygous missense variant in COQ2 in two siblings of consanguineous parents with the infantile form, confirmed by radioisotope biosynthesis assays [45]. In the same year, López and colleagues described PDSS2 variants in an infant with Leigh syndrome and nephropathy [46]; in 2007, Mollet and colleagues added PDSS1 and a truncating COQ2 variant [47]. There followed ADCK3 (now COQ8A) in 2008, in a recessive ataxia with CoQ10 deficiency [48], COQ9 in 2009 [49], COQ6 in 2011, in a nephrotic syndrome with deafness [50], ADCK4 (COQ8B) in 2013, in an SRNS [51], COQ4 in 2015 [52], COQ7 in 2015 [16] and COQ5 in 2018 [53]. More recently, late-onset allelic forms have been recognised: COQ7-related distal motor neuronopathy in 2023 [54], COQ4-related spastic ataxia in 2022 [19] and COQ9-related spastic paraplegia in 2025 [4]. In 2021 the 4-hydroxymandelate pathway, produced by the HPDL protein, was described as a source of the aromatic ring of CoQ10, which links disorders due to HPDL variants to this class [5, 6].

1.3. Nomenclature and classification

Historical nomenclature reflected the evolution of knowledge. The gene for the atypical kinase that regulates synthesis was first called CABC1, because it was thought to be required for complex bc1 activity; the ENMC consensus asked HUGO to change it to ADCK3, and the current designation is COQ8A [1, 35]. Its homologue ADCK4 became COQ8B [1]. OMIM numbers the phenotypes from COQ10D1 to COQ10D9 in order of discovery, but the COQ8B phenotype was registered as nephrotic syndrome type 9 (NPHS9) and not as a numbered CoQ10 deficiency [2, 18]. The COQ8A phenotype has two equivalent designations: COQ10D4 and autosomal recessive spinocerebellar ataxia 9 (SCAR9), also referred to as ARCA2 [1, 12].

Traditionally, primary deficiency was classified into five phenotypes: encephalomyopathy, cerebellar ataxia, severe infantile multisystem disease, SRNS and isolated myopathy [1]. GeneReviews considers this classification outdated, because the spectrum is broader, the presentations overlap and there are no patients with molecularly confirmed primary deficiency with isolated myopathy; most patients with predominantly muscular disease have secondary deficiency, in particular due to MADD [1, 35]. The classification adopted is genetic, gene by gene, with the allelic phenotypes of each gene and their distribution by organ (sections 2 and 3). Saudubray groups coenzyme Q biosynthesis defects among the mitochondrial defects and, in the table of oxidative phosphorylation disease genes, places the ten genes in the cofactor subgroup, with typical deficiency of the combined I+III and II+III activities [3, 55].

HPDL-related disease has its own nomenclature history. The consanguineous Pakistani family described in 1997 and 1999 as autosomal recessive symmetrical spastic cerebral palsy gave rise to the entry «cerebral palsy, spastic quadriplegic, 1» (CPSQ1), attributed to a GAD1 variant; when a new branch of the family was studied, that variant did not segregate with the disease, and autozygosity linkage followed by exome sequencing identified the homozygous HPDL variant NM_032756.4:c.527T>C p.(Leu176Pro), present in both branches [21, 56]. Between 2020 and 2021, three independent series (Husain and colleagues, 17 patients from 13 families; Wiessner and colleagues, 34 patients from 25 families; Ghosh and colleagues, eight families) established HPDL as a disease gene [57–59]. OMIM records two phenotypes: neurodevelopmental disorder with progressive spasticity and brain white matter abnormalities (neurodevelopmental disorder with progressive spasticity and brain white matter abnormalities, NEDSWMA, 619026), which absorbed the former CPSQ1, and autosomal recessive spastic paraplegia 83 (SPG83, 619027) [21, 22]. Orphanet uses the codes ORPHA:641353 and ORPHA:631076 for the same entities [24, 25]. The 2025 reviews propose three clinical classes instead of two, severe, intermediate and mild, because about one fifth to one third of patients fit neither infantile epileptic encephalopathy nor juvenile spastic paraplegia [32, 33]. The link to CoQ10 was established only in 2021, when Banh and colleagues showed that HPDL produces 4-hydroxymandelate, an intermediate in the synthesis of the CoQ10 ring [7].

1.4. Definition of the entities in scope

This article covers the whole class of primary CoQ10 deficiencies due to variants in the genes PDSS1, PDSS2, COQ2, COQ4, COQ5, COQ6, COQ7, COQ8A, COQ8B and COQ9, with the late-onset allelic phenotypes (COQ4-related spastic ataxia, COQ7-related distal motor neuronopathy, COQ9-related spastic paraplegia) and the COQ2-associated susceptibility to multiple system atrophy. The HPDL defect is included as a related entity, because it impairs the synthesis of the CoQ10 ring, although it is not part of the OMIM numbered series [5, 26]. No patients with COQ3 variants have been described, nor disease linked to COQ10A or COQ10B, chaperone proteins that do not take part directly in synthesis [5, 42]. Secondary deficiencies enter only the differential diagnosis (section 6).

2. Molecular Genetics and Pathophysiology

2.1. The biosynthetic pathway

The benzoquinone ring of CoQ10 is formed from tyrosine and phenylalanine, via 4-hydroxybenzoate (4-HB); the polyprenyl tail is formed from acetyl-CoA by the mevalonate pathway and is elongated to ten units by decaprenyl-diphosphate synthase, a heterotetramer encoded by PDSS1 and PDSS2 [35]. COQ2 condenses 4-HB with the decaprenyl tail, and the ring then undergoes decarboxylation, hydroxylations and methylations by enzymes assembled in a multiprotein complex on the matrix face of the inner mitochondrial membrane (complex Q or «CoQ synthome», with the proteins COQ3 to COQ9) [5, 35, 42]. In this complex, COQ5 is a C-methyltransferase, COQ6 and COQ7 are hydroxylases (COQ7 converts 6-demethoxy-CoQ10, DMQ10, in a late step), COQ3 is an O-methyltransferase acting at two steps, and COQ4, COQ8A, COQ8B and COQ9 have structural or regulatory functions; COQ8A and COQ8B are atypical kinases homologous to yeast Coq8p, which stabilises Coq3p [4, 35, 42]. HPDL produces 4-hydroxymandelate (4-HMA), a precursor of 4-HB, which explains the involvement of this gene in the synthesis of the head of the molecule [5, 6]. The mevalonate pathway also links CoQ10 to cholesterol and dolichol: in mevalonic aciduria, plasma ubiquinone was reduced in four of six patients, and lovastatin lowered it by a further 20% in two patients, with severe crisis and rhabdomyolysis [60].

2.2. Genes of the class

Sources: MANE transcripts confirmed in VariantValidator; patient counts from Wang and Hekimi (literature up to May 2022) and from Mantle and colleagues (2023, according to Mantle 2026) [2, 5, 19–22, 42, 61]. GeneGuidance lists RefSeq transcripts for PDSS1 and COQ6 that are not the MANE ones (NM_001321978 and NM_001425255) [4].

The counts reflect the way patients were sought. Among 303 patients published up to 2022, 37.0% had COQ8A variants and 29.0% COQ8B variants; the authors attribute part of this proportion to targeted screening, two in patients with ataxia (69 patients with COQ8A) and three in patients with kidney disease (63 with COQ8B) [42]. No patient with COQ3 variants had been described up to 2026, possibly because the enzyme acts at two steps and loss of function would be more severe [5, 42].

2.3. Mechanisms of injury

CoQ10 deficiency impairs electron transfer from complexes I and II to complex III, which translates biochemically into reduced combined I+III and II+III activities with normal isolated complexes [3, 43]. In fibroblasts, the severity of the deficiency determines the type of injury: with 12% residual CoQ10 (PDSS2), ATP synthesis falls markedly without an increase in reactive oxygen species; with 30% (COQ2), the ATP deficit is mild, but there is an increase in reactive oxygen species, lipid and protein oxidation and cell death. Residual levels of 10–15% or above 60% are not associated with significant reactive species production, whereas 30–50% are associated with maximal oxidative stress and cell death [35]. Fibroblasts with COQ2 variants require uridine to grow, because dihydroorotate dehydrogenase of pyrimidine synthesis depends on ubiquinone [35]. In COQ7 and COQ9 deficiency the intermediate DMQ10 accumulates, which is toxic to mitochondrial function; in the Coq9 R239X mouse, the truncated protein destabilises complex Q and the phenotype is more severe than in the Coq9 Q95X mouse, which has no measurable protein [5, 30].

Organ specificity has partly known bases. In the kd/kd mouse (V117M variant of Pdss2), kidney disease appears after failure of the physiological rise in renal CoQ9 and CoQ10 at around 40 days of life, and conditional deletion of Pdss2 in podocytes only reproduces the proteinuria, unlike deletion in tubules, monocytes or hepatocytes [35]. COQ8B has relatively enriched expression in podocytes, which is consistent with the predominantly renal phenotype [42]. The absence of nephropathy in patients with PDSS1 and the cerebral involvement in PDSS2 and COQ9 deficiencies are unexplained; the ENMC consensus suggests differences in tissue CoQ10 levels, in reactive oxygen species production and in gene expression [35]. Exogenous CoQ10 crosses the blood–brain barrier poorly, which contributes to the poorer response of neurological manifestations [5, 35].

2.4. Inheritance

Inheritance is autosomal recessive for all genes of the class [1, 4]. A de novo contiguous gene deletion at 9q34, with COQ4 haploinsufficiency, caused CoQ10 deficiency in one patient, and the systematic review by Wang and Hekimi included a patient with a monoallelic COQ4 deletion who improved with CoQ10 [8, 42]. COQ2-associated susceptibility to multiple system atrophy appears in OMIM with autosomal dominant and recessive inheritance and is considered controversial (section 2.7) [9].

2.5. Variant spectrum and nomenclature

COQ2 nomenclature has a frequent pitfall. Desbats and colleagues showed that the main functional transcript is shorter than initially described [62], and the MANE transcript (NM_001358921.2) starts 150 nucleotides, or 50 amino acids, after the long transcript NM_015697.9 used in the first articles [38, 61]. The following table converts the most cited variants.

Sources: [31, 34, 38, 41, 45, 47, 50, 54, 61, 63–67]. The Drovandi renal series described COQ2 variants on the long transcript NM_015697.8 and the proteins on NP_001345850.1, which corresponds to the MANE transcript; the MANE descriptions in the table were obtained in VariantValidator from the genomic coordinate [31, 61].

Nyhan's Atlas transcribes the COQ4 variant as «c.190C>T (p.Pro645Ser)», which is impossible in a protein of about 265 amino acids; VariantValidator confirms p.(Pro64Ser) [41, 61]. In the 2017 analysis by Hughes and colleagues, the COQ4 variant p.Arg145Gly was classified as tolerated by SIFT, but caused death at four hours of life in homozygosity and did not complement the Δcoq4 yeast, which shows the limits of in silico prediction in these genes [30].

2.6. Genotype–phenotype correlations

In COQ2, the residual activity of mutant proteins measured in yeast correlates with phenotype, unlike in COQ8A [62]. Distelmaier and colleagues describe three phenotypes linked to residual activity: neonatal with encephalopathy, renal dysfunction and early death; intermediate childhood form, with delay, ataxia, hypotonia and SRNS evolving to neurodegeneration and renal failure; and an adult form resembling multiple system atrophy, described mainly in Asian populations; the variants of the severe forms cluster near the 4-HB transport tunnel in the enzyme [38]. In COQ4, among the patients of Yu and colleagues, those who did not respond to CoQ10 had a null variant and those who responded had two missense variants [1]. In COQ8A, multisystem involvement was more frequent with missense variants than with two loss-of-function variants (82–93% versus 53%; p = 0.029), and homozygosity for p.Arg271Cys (published protein designation) was associated with a very severe early presentation [34]. In COQ8B, biallelic truncating variants, present in one third of patients (47 of 140), were associated with a milder renal phenotype: patients without two truncating variants more often had onset before 10 years (OR 2.2; 95% CI 1.0–4.8) and end-stage kidney disease before 15 years (OR 7.7; 95% CI 2.9–20.6); missense variants in the GQα1/2 or GQα3 motifs were associated with end-stage kidney disease in 83% of patients during observation, versus 61% in the others [31]. In COQ2 and COQ6, almost all patients had at least one missense variant; the four exceptions had truncating variants near the 3′ region, compatible with residual activity. The associations of recurrent variants with severity (COQ2 p.(Ser96Asn) severe and p.(Asn178Ser) of later onset; COQ6 p.(Gly255Arg) severe, p.(Ala353Asp) later and p.(Arg360Trp) with more extrarenal involvement; COQ8B c.1339dup and c.1199dup slower) appear in the table in section 2.5 [31]. In COQ7, hypomorphic alleles, with partial rescue in yeast, are associated with the late-onset neuropathy form, and those with more complete loss of function with the neonatal multisystem form; the c.3G>T variant affects the main isoform, but preserves other isoforms, which maintains some synthesis [4, 20]. In all large renal series, intra- and interfamilial variability was considerable, which suggests genetic and non-genetic modifiers [31]. In the Drovandi series, the coefficient of variation of age at onset was 35% among relatives with the same variant, 64% among different families with the same variant and 106% among families with different variants [31].

2.7. Allelic disorders and related entities

Several genes of the class cause late-onset phenotypes that may not be recognised as CoQ10 deficiency. COQ4 causes, in addition to neonatal encephalocardiomyopathy, autosomal recessive spastic ataxia 10 (SPAX10), with onset from infancy to adulthood [19]. COQ7 causes autosomal recessive distal hereditary motor neuronopathy 9 (HMNR9), classified as a strong association by the ClinGen Charcot-Marie-Tooth gene curation expert panel on 31 March 2026 [4, 20]. COQ9 caused pure spastic paraplegia in one family with a homozygous splicing variant [4]. COQ8B has been associated with retinitis pigmentosa in Orphanet, and one patient with COQ8B-related SRNS had retinitis pigmentosa [1, 4].

The association between COQ2 and multiple system atrophy was proposed in 2013 on the basis of two Japanese multiplex families (M78V-V343A homozygosity, historical protein designations, and R337X/V343A compound heterozygosity) and of an association of the common variant V343A, exclusive to the Japanese population, with sporadic disease [64]. OMIM records this association as a controversial susceptibility: it was not replicated in 299 Korean patients and 365 controls (allele frequency of 2.7% versus 2.6%) or in other series, and two Japanese siblings with COQ2 variants had retinitis pigmentosa and low cerebellar CoQ10, as in primary deficiency [9]. In ClinVar, V343A is classified as benign or likely benign, whereas R337X is registered as a risk factor [63]. A heterozygous COQ6 variant was found in a family with autosomal dominant schwannomatosis, but its contribution was not confirmed and OMIM classifies it as of uncertain significance [14]. HPDL deficiencies range from neonatal encephalopathy to spastic paraplegia with onset in adolescence [5, 21, 22].

2.8. Founder variants and population frequencies

The COQ4 variant c.370G>A p.(Gly124Ser) was present in 10 of 11 patients from southern China (five homozygous), with a population frequency of 0.001118 and exclusive presence in East and Southeast Asian populations in gnomAD; homozygotes shared a haplotype of up to about 577 kb [65]. The COQ2 variant NM_015697.9:c.973A>G p.(Thr325Ala) (long transcript; = NM_001358921.2:c.823A>G p.(Thr275Ala) on the MANE transcript), hypomorphic in yeast, has a frequency of 0.00071 in East Asia in gnomAD and below 0.00001 globally [68]. In the Jewish population of Iraqi origin, the 9.6 kb duplication in COQ5 had three heterozygous carriers among 92 controls, with an allele frequency of about 1.5% [17]. In Korea, five of six children with SRNS and deafness due to COQ6 were compound heterozygous for p.Lys64del and p.Pro261Leu (protein designations published in OMIM) [14]. The COQ7 variant c.3G>T p.(Met1?) was described in a sibship of Portuguese parents and in several unrelated Brazilian families (section 9) [20, 66]. In the international renal series, the geographical concentration of recurrent variants suggests founder effects: COQ2 p.(Ser96Asn) in Turkey; COQ6 p.(Ala353Asp) in Kazakh, Turkish and Iranian populations, p.(Gly255Arg) in the Middle East and p.(Arg360Trp) in Central and Eastern Europe and China; COQ8B p.(Asp250His) in China and c.1339dup, c.1199dup and p.(Leu98Arg) in patients of Turkish or Kurdish ancestry [31].

2.9. The HPDL gene: structure, function, variants and correlations

Structure and function. HPDL (4-hydroxyphenylpyruvate dioxygenase-like) is located at 1p34.1 and has a single exon; the MANE transcript is NM_032756.4 and the protein has 371 amino acids (NP_116145.1), with an N-terminal mitochondrial targeting signal and two VOC (vicinal oxygen chelate) domains, between amino acids 7–135 and 160–328 [33, 61, 69]. Residues His163, His258 and Glu339 coordinate the iron of the catalytic centre [32, 33]. The gene is thought to have arisen by retrotransposition at the base of the tetrapod lineage and is a paralogue of HPD, the dioxygenase of tyrosine catabolism, but unlike the latter it does not alter tyrosine metabolites [59]. The submitochondrial localisation is described differently: Wiessner and colleagues localised the protein to the outer mitochondrial membrane, whereas Lee's group describes it as an intermembrane space protein [33, 69]. The function was identified by ¹⁸O₂ labelling in human cells: HPDL converts 4-hydroxyphenylpyruvate (4-HPPA), the first intermediate of tyrosine catabolism, into 4-hydroxymandelate (4-HMA), which gives rise to 4-HB, the immediate precursor of the CoQ10 ring [6, 7]. HPDL-related disease is therefore a defect upstream of COQ2, in the production of the aromatic head of the molecule, and not in the complex Q assembly pathway.

Mechanism. The evidence points to loss of function. In fibroblasts from the most severely affected patients, HPDL protein was reduced; as the gene has a single exon, truncating variants are not subject to nonsense-mediated mRNA decay and give rise to shortened proteins [57, 69]. Missense variants introduced at the homologous positions of HPPD reduced the conversion of 4-HPPA to homogentisate, with a greater effect for variants close to the catalytic centre, although the levels and localisation of mutant HPDL in HeLa cells were normal [58, 69]. The Hpdl−/− mouse has seizures, poor weight gain, a small brain with cortical apoptosis and death by day 15 of life, with reduced plasma 4-HMA [6, 59]. hpdl knockdown in zebrafish causes abnormal motor behaviour [58]. Patients' cells show reduced oxygen consumption; in a Chinese series of six patients, fibroblasts and immortalised lymphocytes had decreased mitochondrial respiration and reduced complex II activity [59, 70], and in two families citrate synthase and complex I activity were reduced [71]. In cortical neurons and organoids derived from patients' induced pluripotent stem cells there was premature neurogenesis with progenitor depletion, defective assembly of respiratory supercomplexes and slow organoid growth, which the authors relate to the microcephaly of the severe forms [72]. Patients' muscle biopsy showed variation in fibre calibre and respiratory chain activities sometimes increased and sometimes decreased, without a consistent signature of oxidative phosphorylation deficiency [21].

Inheritance. Inheritance is autosomal recessive, with biallelic variants in homozygosity or compound heterozygosity [21, 22]. In the 2026 review, consanguinity was present in 49.55% of published families [73]. No affected heterozygous carriers have been described.

Variant spectrum. In the systematic review by Alecu and colleagues (91 patients from 59 families), there were 69 distinct pathogenic variants; 65.6% of patients were homozygous and 34.4% compound heterozygous; the variants were missense in 65.0%, frameshift in 23.9% and nonsense in 8.3% [32]. Missense variants cluster in three segments, one in the VOC1 domain and two in VOC2, in the regions adjacent to the iron-binding residues [32]. The recurrent variant p.(Gly50Asp) was present in 15 patients from 9 families from Egypt, Sudan, Syria and Turkey, and haplotype analysis in the Middle Eastern families suggested a founder effect [32, 57]. The following table brings together the most cited variants, all confirmed in VariantValidator (GRCh38, MANE transcript).

Three nomenclature notes are useful when reading the sources. The 2020–2021 articles used versions NM_032756.2 and NM_032756.3 of the transcript, with the same c. numbering as the MANE version NM_032756.4 [61, 69]. OMIM describes p.(Ile266Thr) as «c.797C-T», but the substitution that produces threonine at codon 266 is c.797T>C, as VariantValidator confirms [61, 69]. In one paragraph by Lee and colleagues, p.(Gly50Asp) appears as «c.140C>G»; the correct description is c.149G>A [33, 61]. OMIM and the Wiessner article describe c.256del as «Ala86fs»; the full description is p.(Ala86ArgfsTer45) [61, 69].

Genotype–phenotype correlations. In the review by Lee and colleagues (99 patients), 75.0% of patients with two truncating variants had the severe form and none had the mild form, whereas only 36.4% of patients with two missense variants had the severe form; mean age at onset was 16.1 months with two truncating variants, 46.5 months with one truncating and one missense variant and 58.3 months with two missense variants (p = 0.014) [33]. Missense variants located less than 5 Å from the iron-binding residues (published protein designations NP_116145.1:p.Leu164Pro, p.Gly260Glu and p.Leu338Pro) and those in the C-terminal region were always associated with the severe form, as were the p.(Met1?) variants [33]. Alecu and colleagues classified the 2355 possible missense substitutions with 34 predictors; «moderate» missense variants in homozygosity were over-represented in the mild group (odds ratio 12.4; 95% CI 3.2–71.8) and almost absent in the severe group, and a single moderate allele, even in trans with a truncating variant, was enough to increase the probability of the mild form (odds ratio 13.6; 95% CI 4.4–49.1) and delay onset (median 11 years versus 6 months) [32]. Wiessner and colleagues had observed truncating variants in 47% of the alleles of severe patients and in 10% of mild ones [33, 58]. These associations are exploratory, are based on heterogeneously published data and do not predict individual course: two Japanese siblings homozygous for NM_032756.4:c.340G>T p.(Gly114Cys) had distinct phenotypes (the elder sister, pure spastic paraplegia, and the brother, a complicated form with early intellectual disability and MRI abnormalities) [75], and the 8-year-old patient treated with 4-HB had the same truncating genotype as his brother who died in the first year with encephalopathy [6].

3. Clinical Spectrum and Manifestations by System

3.1. Overview

Primary CoQ10 deficiency is extremely heterogeneous. Age at onset ranges from the prenatal period to the seventh decade, and the spectrum extends from fatal neonatal encephalopathy with hypotonia to a late, slowly progressive presentation resembling multiple system atrophy [1]. There are no pathognomonic signs [1]. The Brazilian textbook on genetic syndromes summarises the manifestations by system: neurological (fatal neonatal encephalopathy with hypotonia, cerebellar ataxia, pyramidal dysfunction, neurodegeneration with autonomic failure, intellectual disability, seizures, spasticity, dystonia, peripheral neuropathy, Leigh syndrome), auditory (sensorineural deafness), ophthalmological (retinopathy, optic atrophy), cardiac (hypertrophic cardiomyopathy, heart failure), renal (tubulopathy, SRNS progressing to end-stage kidney disease) and muscular (myopathy, weakness, exercise intolerance) [28]. Saudubray distinguishes the typical infantile presentation, with rapidly progressive nephropathy variably associated with sensorineural deafness, optic atrophy, ataxia, dystonia, weakness and stroke-like episodes, from the late-childhood forms with encephalomyopathy, seizures and recurrent myoglobinuria, cerebellar ataxia or, more rarely, myopathy [3]. Comparing ages, the childhood forms present as Leigh syndrome, myoglobinuria or encephalopathy, and the adult forms as cerebellar ataxia, seizures or myopathy [76].

The following table reproduces the distribution of manifestations by gene according to GeneReviews.

HCM: hypertrophic cardiomyopathy. Source: GeneReviews, table 1 [1]. Because patients with COQ6 and COQ8B were identified mainly through SRNS, GeneReviews allows that these genes may cause broader phenotypes [1]. Later OMIM data and the 2023–2026 series added deafness in PDSS1, COQ5 ataxia confirmed in one family and the late-onset forms of COQ4, COQ7 and COQ9 [4, 10, 17, 19, 20].

3.2. Kidney

SRNS, unusual in mitochondrial diseases, is the hallmark of primary deficiency; without high-dose CoQ10 it usually progresses to end-stage kidney disease [1]. It has been described with PDSS1, PDSS2, COQ2, COQ6 and COQ8B, and not with COQ4, COQ8A or COQ9, nor in the Coq9 mouse model [1]. It usually presents with proteinuria in infancy; many patients begin with SRNS that progresses to end-stage kidney disease, followed by encephalomyopathy with seizures and stroke-like episodes, severe neurological deficit and death, whereas others have only SRNS in the first or second decade, with slow progression and without extrarenal manifestations [1]. Biopsy usually shows focal segmental glomerulosclerosis (FSGS) or collapsing glomerulopathy; diffuse mesangial sclerosis has been described with COQ6 and PDSS2 [14, 35, 77].

The largest renal series, with 251 patients with variants in COQ2, COQ6 or COQ8B (173 published and 78 new, from the PodoNet, mitoNET and CCGKDD registries), defined the natural history by gene. First symptoms appeared at a median age of 1.0 year for COQ2, 1.2 years for COQ6 and 9.8 years for COQ8B, and the first renal manifestation at 1.0, 2.0 and 9.9 years, respectively; the latest ages at renal diagnosis were 11.6 years (COQ6), 27 years (COQ2) and 32.2 years (COQ8B), and 14.3% of all patients manifested kidney disease after 16 years. Renal involvement was present in 98% (246 of 251) and was isolated at diagnosis in 34% of patients with COQ2, 10.8% with COQ6 and 70.7% with COQ8B. Proteinuria was nephrotic in 85.7%, 86.1% and 71.7%; COQ2 typically presented with oedema and hypertension and normal renal function, about 40% of patients with COQ6 had stage 2–4 chronic kidney disease and 21% end-stage kidney disease at diagnosis, and one third of patients with COQ8B were already in end-stage kidney disease at diagnosis. Proteinuria was asymptomatic in 18.7% of patients with COQ6 and 23.6% with COQ8B, including younger siblings of index cases, and microhaematuria occurred in 6.8%, 8.5% and 18%. The probability of end-stage kidney disease at 5 years was 47.6% (COQ2), 47.8% (COQ6) and 3.1% (COQ8B), and at 18 years 80.8%, 72.3% and 74.2%, with median ages at end-stage kidney disease of 2.5, 3.4 and 13 years; in adulthood, renal survival was equally poor in the three genes, between 20% and 25%, and the risk persisted after 25 years [31]. These values mix treated and untreated patients and come from registries and literature review, with selection bias towards cases with kidney disease. On biopsy, performed in 155 patients, FSGS accounted for 74.2% of histological diagnoses, with the collapsing variant in 24% of FSGS in COQ2, 11.5% in COQ6 and 9.4% in COQ8B; electron microscopy showed dysmorphic mitochondria in 30%, 25% and 10% of biopsied patients, and in three cases this finding led to the diagnosis of mitochondrial disease before genetic testing [31]. In COQ6, Heeringa's original series had proteinuria at a median age of 1.2 years (0.2–6.4) and end-stage kidney disease at a median age of 1.7 years (0.4–9.3), with five deaths in childhood [14]; six Korean children with COQ6 all reached end-stage kidney disease at a median of 13 months (1–27) after onset of nephrotic syndrome, at 29 months on average [14]. In COQ8B, Ashraf's initial series included 15 patients from 8 families with typical onset between 10 and 20 years, with FSGS or collapsing glomerulopathy [18, 51]. COQ9 causes tubulopathy and not SRNS; in one of the two patients from one family, the tubulopathy appeared hours after birth [1].

3.3. Central and peripheral nervous system

Encephalopathy is the most frequent neurological manifestation and encompasses diverse clinical and imaging pictures; some patients have Leigh syndrome imaging and others MELAS-like stroke-like episodes [1]. Seizures, dystonia, spasticity and intellectual disability are frequent, but the number of patients is too small for generalisations [1]. Cerebellar atrophy is frequent in almost all subgroups, although non-specific [35]. Among the treatable progressive ataxias, Saudubray highlights CoQ10 deficiency, in particular the ataxia with oculomotor apraxia phenotype [55]. A 2026 review of patients with COQ4 variants found seizures in 48 of 69 (70%), with early-infantile-onset developmental and epileptic encephalopathy, epilepsy of infancy with migrating focal seizures and infantile spasms syndrome responsive to corticosteroids [78]. Adult-onset status epilepticus has been described in a series of patients with COQ8A [79].

Peripheral neuropathy has two patterns. In PDSS1, sensorimotor neuropathy was described in siblings with deafness and optic atrophy [1, 10]. In COQ7, the late-onset form is a distal motor neuronopathy: in a review of 28 patients, median onset was at 10 years (interquartile range 5–11), the dominant subtype was distal hereditary motor neuropathy (89%), there were foot deformities in 19 of 23 (83%) and upper motor neuron signs in 20 of 27 (74%) [67]. The phenotyping of HMNR9 includes distal weakness and wasting predominantly in the lower limbs, hand involvement, pes cavus, foot drop, inability to walk on heels or toes, mild sensory changes in some and pyramidal signs [20].

The multiple system atrophy-like phenotype, with autonomic failure and combinations of parkinsonism, cerebellar ataxia and pyramidal dysfunction, began in the seventh decade in two multiplex families with COQ2 [1, 64].

3.4. Sensory organs

Sensorineural deafness is common in COQ6: it was present in all nine tested patients of the original series and, in the 2022 renal series, in 73.9% (34 of 48), with a mean age at manifestation of three years; four patients had congenital deafness, in those with later onset the median age at diagnosis was 5 years (interquartile range 2.9–6), and in 24% deafness preceded renal symptoms [31, 35]. In the same series, deafness occurred in only one patient with COQ2 and in none with COQ8B [31]. It also occurs in some patients with COQ2, PDSS1 and PDSS2 [1]. Retinopathy, sometimes isolated in adults, has been described with PDSS1, COQ2, COQ4 and COQ5, and optic atrophy with PDSS1, PDSS2 and COQ2 [1]. Adult retinitis pigmentosa may be isolated or associated with late multiple system atrophy in COQ2 [1]. Rötig's three siblings had progressive visual impairment due to pigmentary retinopathy, optic atrophy and cataracts [35].

3.5. Heart

Hypertrophic cardiomyopathy has been described in neonatal-onset COQ2, in COQ4 with prenatal onset and in COQ9 with neonatal lactic acidosis and multisystem disease; in the latter, the cardiac disease worsened despite CoQ10 [1]. In neonatal COQ7, Kwong described intrauterine growth restriction, cardiomegaly and severe hypertrophic cardiomyopathy at 26 days of life [16]. A patient homozygous for COQ8A p.Arg301Trp (published protein designation) died at 17 years from severe dilated cardiomyopathy [34]. GeneReviews notes that, up to 2023, cardiomyopathy had been observed only in the most severe neonatal-onset phenotypes [1].

3.6. Muscle

Weakness and exercise intolerance are frequent; biopsy may show non-specific lipid accumulation and mitochondrial proliferation, but muscle involvement is almost always accompanied by extramuscular manifestations in primary deficiency [1]. The isolated myopathic form described before the molecular era, with exercise intolerance between 6 and 33 years, proximal weakness, creatine kinase up to 20 times the upper limit and lactate up to 10 times after exercise, proved in most cases to be secondary to MADD due to ETFDH variants [35].

3.7. Subtypes by gene

PDSS1 (COQ10D2). The first patients were two siblings of consanguineous Moroccan parents, aged 22 and 14 years, with early-onset deafness, optic atrophy, mild intellectual disability, peripheral neuropathy, obesity, livedo reticularis and cardiac valvulopathy, without nephropathy and with prolonged survival [10, 35]. Later, an infant with nephrotic syndrome and death at 16 months and a boy with poor weight gain at 5 months, deafness, pulmonary hypertension, livedo reticularis and white matter cysts were described [10]. Orphanet calls it deafness–encephaloneuropathy–obesity–valvulopathy (DEOV) syndrome [4].

PDSS2 (COQ10D3). The first patient, the son of unrelated parents, had neonatal pneumonia and hypotonia, refractory seizures at three months, Leigh syndrome and nephrotic syndrome, with a fatal outcome [11, 46]. Iványi and colleagues described a 7-month-old infant with nephrotic syndrome, encephalomyopathy, hypertrophic cardiomyopathy, suspected deafness and retinitis pigmentosa and raised lactate, who died four weeks after admission despite CoQ10; autopsy showed diffuse mesangial sclerosis; by 2018 six patients had been described [77]. The Leigh GeneReviews attributes less than 1% of Leigh cases of nuclear cause to PDSS2, and also includes in the Leigh spectrum COQ4 and COQ7 (less than 1% each), COQ9 (one person) and HPDL (less than 1%) [26].

COQ2 (COQ10D1). This is the gene with the broadest spectrum, from birth to the seventh decade [1]. The forms include isolated SRNS, SRNS with encephalomyopathy, severe multisystem disease, adult retinitis pigmentosa and the multiple system atrophy-like presentation [1]. A French family had two children who died soon after birth with anaemia, pancytopenia, liver failure and renal failure [2]. In the 2022 renal series, 22% of patients with COQ2 had the classic infantile multiorgan form, with death at a median age of 6 months, and 47% developed neurological symptoms (encephalopathy, ataxia, seizures, nystagmus, psychomotor delay), at a median age of 2.7 years and within the first three years in 87%; of the 22 patients with initially isolated kidney disease, nine developed extrarenal manifestations, six within three years and three up to 20 years later. Survival was 77.2% at 1 year and 69.9% at 10 years for COQ2, versus 90.8% for COQ6 and 99.2% for COQ8B at 10 years, and no patient with COQ2 died after 5 years [31]. Prenatal diagnosis has been described in a fetus with severe growth restriction and cardiac anomalies [68].

COQ4 (COQ10D7 and SPAX10). The classic form is a neonatal-onset encephalocardiomyopathy, with hypotonia, bradycardia, respiratory failure, seizures, basal ganglia lesions and cardiomyopathy, usually fatal [15, 65]. Nyhan's Atlas summarises the five patients of Brea-Calvo: four with prenatal or neonatal onset and death hours after birth, and one with more indolent disease, slowly progressive neurodegeneration from 10 months, spastic gait at 3 years, loss of walking at 6 years and seizures at 12, all with marked lactic acidaemia and reduced CoQ10 in muscle and fibroblasts [41]. Six girls from four families described by Chung died between 36 hours and 19 months [15]. In the southern Chinese series, five of eleven patients had the neonatal form and six onset in the first year, more heterogeneous, with dystonia in two [65]. At the less severe end, SPAX10 is a slowly progressive spastic paraparesis and/or cerebellar ataxia, with onset between 15 and 34 years in German and Iraqi patients and between 1 and 55 years in patients from China and Taiwan [19].

COQ5 (COQ10D9). Three sisters of Iraqi Jewish ancestry had cerebellar ataxia with non-progressive cerebellar atrophy in the first decade; the most affected, aged 17 years, had short stature, motor and cognitive delay, dysarthria, myoclonus, controlled tonic-clonic seizures, mild to moderate intellectual disability, nystagmus, slow saccades, gaze apraxia and mild spasticity [17]. The 2023 review defined the core spectrum (reduced CoQ10, intellectual disability, encephalopathy, ataxia, cerebellar atrophy, speech regression or dysarthria, short stature) and the common mechanism of altered splicing; in 2025 two siblings with profound delay and lesions compatible with strokes at various stages were described [4].

COQ6 (COQ10D6). The distinctive phenotype is infantile-onset SRNS with sensorineural deafness [14, 31]. In the original series, three patients had central nervous system involvement with seizures or ataxia [35]. In the 2022 renal series, almost a quarter of patients had seizures or cognitive impairment, and one in twelve had haematuria with proteinuria and deafness, a picture that can mimic Alport syndrome [31].

COQ7 (COQ10D8 and HMNR9). The severe neonatal form includes pulmonary hypoplasia, contractures, prenatal renal dysplasia with renal dysfunction, early infantile hypertension and cardiac hypertrophy; survivors develop progressive peripheral neuropathy, hypotonia and muscle wasting, mild psychomotor delay and hearing and visual deficits [4, 16]. There are intermediate forms with spasticity and mild deafness [16]. The late-onset form is the distal motor neuronopathy described in 3.3 [20, 67].

COQ8A (COQ10D4, SCAR9). The multicentre cohort of 59 patients by Traschütz is the largest description [34]. Cerebellar ataxia was present in all and was the initial symptom in 68%, with median onset at 7 years (interquartile range 4–15). Hyperkinetic movement disorders occurred in 41% (myoclonus 30%, dystonia 28%, head tremor 22%, postural or action tremor 20%), sometimes as the initial symptom; cognitive impairment in 49%, epilepsy in 32%, exercise intolerance or weakness in about 25% and neuropsychiatric symptoms (depression, anxiety, psychosis, impulsivity) in about 25%. Deafness, stroke-like episodes, cataracts, diabetes and optic atrophy were rare. Cerebellar atrophy was present in all patients on magnetic resonance imaging (MRI), and 28% had cerebral atrophy or T2 hyperintensity in the dentate nuclei and pons [34]. GeneReviews describes onset with weakness and exercise intolerance between 18 months and 3 years, followed by ataxia, with a progressive or apparently self-limiting course [1].

COQ8B (NPHS9). SRNS appears in the first and second decades and often progresses to end-stage kidney disease; four patients had mild intellectual disability, two occasional seizures and one retinitis pigmentosa [1]. It is the gene with the highest proportion of isolated renal involvement (70.7%); in the renal series of 140 patients, extrarenal manifestations occurred in 29.3%, mainly neurological (12.1%: moderate neurocognitive deficit, developmental delay, seizures), and mortality was 4.3%, with sepsis as the most frequent cause [31].

COQ9 (COQ10D5). The original form is neonatal: the seventh child of unrelated Pakistani parents presented on the first day with poor feeding, hypothermia and seizures, with lactic acidosis, mild left ventricular hypertrophy, tubulopathy and cerebral and cerebellar hypoplasia or atrophy [13, 35]. An infant of consanguineous Turkish parents was born at 36 weeks because of anhydramnios, with hypotonia, bradycardia and severe lactic acidosis [13]. In 2025, a homozygous splicing variant was described in two adult siblings with pure spastic paraplegia since childhood, with low CoQ10 in muscle and fibroblasts and accumulation of DMQ10 [4].

HPDL. The patients described were aged between 6 months and 39 years, with developmental delay, seizures and spasticity, in a spectrum ranging from neonatal encephalopathy to adolescent spastic paraplegia [5, 21, 22].

3.8. Natural history and treated disease

Pure natural history data are scarce, because almost all diagnosed patients receive CoQ10. The severe neonatal multisystem forms respond poorly and generally lead to death in the neonatal period or in the first year [1]. In COQ8A ataxia, longitudinal progression in seven untreated patients was 0.45 SARA points per year (95% CI 0.12–0.77), and dependence on a walking aid and on a wheelchair was estimated at about 55% and 35%, but not before 40 and 50–60 years [34]. In kidney disease due to COQ2, COQ6 and COQ8B, comparison with a matched untreated cohort showed kidney failure-free survival at five years of 62% in treated and 19% in untreated patients, with a median annual loss of estimated glomerular filtration rate of 0 and 18 ml/min/1.73 m², respectively (section 7) [37]. The differences between these series reflect the gene, age at diagnosis and timing of CoQ10 initiation, and do not allow direct comparisons.

3.9. HPDL-related disease

HPDL-related disease is mainly neurological: no renal, auditory or cardiac involvement has been described, unlike the other genes of the class [32, 33]. The continuous spectrum ranges from neonatal-onset epileptic encephalopathy to pure adolescent spastic paraplegia, and the two quantitative reviews of 2025 divide it into three classes, whose figures are summarised in the following table.

The two reviews use different criteria, clinical classes defined by the authors in Lee and unsupervised clustering of phenotypic terms in Alecu, so the groups do not coincide exactly; Alecu's intermediate group (16 patients) has onset in infancy and is more severe than Lee's [32, 33].

Severe form. Patients present in the first month or the first months with seizures, axial hypotonia, feeding difficulty and central apnoeas, and progress to spastic tetraparesis, axial dystonia, contractures, progressive microcephaly (down to −5 to −6 standard deviations) and profound global delay, often without eye contact, walking or speech [21, 32]. In the most severe group of the Alecu review, seizures appeared in 93.6% (median onset at 2.5 months), predominantly focal myoclonic and tonic, central apnoea in 22.6% (median onset at 1.6 months, usually associated with seizures), microcephaly in 74.2%, mostly secondary, poor weight gain in 38.7% and short stature in 32.3% [32]. Ghosh and colleagues described, in 16 patients from consanguineous families from Egypt, Pakistan and Iran, myoclonic or focal epilepsy with onset in the first days or months, generally refractory, extrapyramidal signs (dystonia, tremor, rigidity), cortical blindness and nystagmus; eight died between 2 months and 13 years, mainly from epilepsy [21, 59]. In the Husain series, five patients had episodes of acute respiratory failure requiring admission, and three had peripheral neuropathy on electrophysiological study [21, 57].

Intermediate form. Onset lies between the end of the first year and school age, with motor delay, spasticity predominantly in the lower limbs and, in some patients, initially normal development followed by regression [33]. In the Wiessner series, the nine intermediate patients began between 3 and 7 years and all achieved independent walking; seven had upper limb involvement, six ataxia, six oculomotor abnormalities and two epilepsy, and MRI showed cerebellar atrophy in five of eight, corpus callosum abnormalities in four and hyperintense lesions of the inferior olives in five [21, 58]. Many of these patients had previously been diagnosed with idiopathic cerebral palsy [33]. The family of the former CPSQ1 illustrates this course: onset in early childhood with delayed walking and lower limb spasticity, moderate to severe intellectual disability, without seizures, and wheelchair dependence in the second or third decade [21, 56].

Mild form (SPG83). In the Husain series, five patients from four families, aged 17 to 39 years, began at 14–15 years with unsteady gait, cramps, myalgia and reduced exercise tolerance, and had spastic paraparesis with hyperreflexia and Babinski sign; one lost the ability to walk, and dysarthria, dysphagia and mild urinary urgency were occasional; brain and spinal MRI was normal and none had developmental delay, cognitive impairment or epilepsy [22, 57]. In the Wiessner series, the ten mild patients began between 8 and 17 years, after normal motor development; three had pyramidal signs in the upper limbs, two cognitive delay, and MRI was normal in the seven studied [22, 58]. In Alecu's mild group, 25.0% lost the ability to walk, and prodromal symptoms included leg muscle stiffness (34.1%), frequent falls (15.9%), unsteady gait (13.6%) and exertional myalgia (11.4%); mild motor delay and mild intellectual disability each occurred in 27.3%, and weight, height and head circumference were normal [32]. In 2026 three cases of SPG83 were described in a Mexican family with a homozygous variant [80].

Neuroimaging. Husain and colleagues defined three MRI patterns: pattern I, with multifocal striatal T2 changes, reduced white matter volume, thin corpus callosum and deficient myelination, often with U-fibre involvement and without cerebellar involvement, associated with the most severe forms; pattern II, with white matter lesions of the thalami, brainstem and cerebellum; and pattern III, without defined pathology [21, 57]. The changes progressed in some patients and regressed in others [21]. In Alecu's severe group, MRI showed white matter abnormalities in 83.9%, thin corpus callosum in 64.5%, cortical atrophy in 41.9%, vermis hypoplasia in 9 of 12 and cerebellar atrophy in 8 of 12, and spectroscopy may show a lactate peak [32]. A Japanese infant developed, after an initial leukoencephalopathy, bilateral lesions of the putamina and brainstem similar to those of Leigh syndrome [74], and the GeneReviews on the Leigh spectrum includes HPDL among the nuclear genes responsible for less than 1% of cases [26]. Acute ophthalmoplegia with an MRI pattern similar to that of Wernicke encephalopathy has also been described [81].

Other systems and laboratory findings. Blood and cerebrospinal fluid lactate rise mainly during episodes of encephalopathy and normalise between them (medians of 2.1 and 1.81 times the upper limit of normal in nine severe patients) [32]. Acute episodes are frequently triggered by fever or infection [32]. Ophthalmological abnormalities such as nystagmus, ophthalmoplegia, visual impairment and strabismus are frequent in the severe and intermediate forms; in a 9-year-old child with SPG83, intermittent exotropia became constant (80 prism dioptres) within one year and was surgically corrected with a good result at 12 months, and there was altered pigmentation of the retinal pigment epithelium in one eye [33, 82]. In five children with mitochondrial diseases that began with presentations resembling autoimmune encephalitis or acquired demyelinating syndrome, one of them with HPDL variants, there was initial improvement with corticosteroids or immunoglobulin, followed by relapses and progression [83].

Natural history and treated disease. All outcome data for HPDL-related disease are natural history data, because no disease-modifying treatment is available outside individual use. In the Alecu review, 10 of 91 patients (11.0%) had died, all from the severe group, with an estimated median survival of 11 years in that group and a median age at death of 18 months; the causes were refractory status epilepticus and acute deterioration, often after fever or infection [32]. In the Lee review, deaths (12 of 48 severe patients) were due to seizures in nine, pneumonia in two and encephalopathy in one [33]. The data are cross-sectional, with short follow-up ages in the severe forms (median 4 years), and do not allow estimation of the lifetime risk of loss of walking or of death in the later-onset forms [32].

4. Clinical Recognition and Warning Signs

4.1. When to suspect

GeneReviews recommends suspecting primary CoQ10 deficiency in all individuals with clinical or neuroradiological manifestations suggestive of mitochondrial encephalomyopathy, with SRNS (especially when there are neurological manifestations), with retinitis pigmentosa or with unexplained ataxia, and also considering it in motor neuropathy without another evident cause [1]. The ENMC consensus stresses suspecting the infantile form in young children with nephropathy, especially SRNS, even without encephalopathy, and in unexplained cerebellar ataxia [35]. A family history compatible with autosomal recessive inheritance (affected siblings, consanguinity) strengthens the suspicion, but its absence does not exclude the diagnosis [1]. The importance of recognition stems from the possibility of treatment: primary deficiency should be considered in very diverse presentations because it responds, in some patients, to high-dose oral CoQ10 [1].

HPDL-related disease should be considered in four contexts: the newborn or infant with early refractory epilepsy, central apnoeas, progressive microcephaly and diffuse white matter abnormalities, with or without raised lactate; the child diagnosed with spastic cerebral palsy without perinatal insult, especially if there is progression, regression or consanguinity; the child or adolescent with spastic paraplegia beginning between 8 and 17 years and normal MRI; and the infant with Leigh-pattern lesions and normal metabolic screening [32, 33, 74, 84]. In a series of 210 patients with hereditary spastic paraplegia, diplegic cerebral palsy or moderate to severe neurodevelopmental disorders studied by exome sequencing, 14 patients with biallelic HPDL variants were identified [84].

4.2. Warning signs and urgent situations

The signs warranting urgent assessment are neonatal lactic acidosis with respiratory or cardiac failure, massive proteinuria with generalised oedema and acute encephalopathy with refractory seizures or stroke-like episodes [1, 11, 38]. A normal blood lactate does not exclude the diagnosis, and cerebrospinal fluid lactate, although it may be more sensitive, may also be normal [1]. MRI may show bilateral symmetrical lesions of the basal ganglia and brainstem (Leigh pattern), thalamic lesions, stroke-like lesions or cerebellar atrophy [1, 38, 76]. The ENMC consensus adds the triad of encephalopathy, myopathy with recurrent myoglobinuria and ragged-red fibres as characteristic of the encephalomyopathic form [35].

In mitochondrial diseases in general, anaesthesia, surgery and intercurrent illness may trigger metabolic decompensation, and the textbook chapter on mitochondrial cardiomyopathies recommends their prevention [85]. In children with nephrotic syndrome, urinalysis of siblings is recommended by the International Pediatric Nephrology Association (IPNA) even before the genetic result [36].

In HPDL-related disease, fever and infections trigger episodes of acute encephalopathy with seizures, apnoea and lactic acidosis, and refractory status epilepticus is the main cause of death in the severe forms; the five patients in the Husain series with acute respiratory failure required admission [21, 32]. Initial improvement with corticosteroids in a presentation resembling autoimmune encephalitis does not exclude a mitochondrial disease, and relapses with new signs should prompt genetic testing [83].

4.3. Newborn screening

None of the sources consulted describes newborn screening for primary CoQ10 deficiency, and presymptomatic diagnosis is made by evaluating the siblings of an index case (sections 8 and 11) [1].

5. Diagnosis: Criteria, Tests and Algorithms

5.1. Criteria

There are no formal clinical criteria. The diagnosis of primary deficiency is established by identifying biallelic pathogenic or likely pathogenic variants in a gene of the biosynthetic pathway, in a patient with a compatible phenotype; biochemistry supports the diagnosis when molecular testing is inconclusive or when the effect of variants of uncertain significance (VUS) needs confirmation [1]. CoQ10 measurement distinguishes CoQ10 deficiency from other mitochondrial diseases with a similar presentation, but does not distinguish primary from secondary deficiency; only molecular testing does so [1].

5.2. Molecular testing

GeneReviews proposes two strategies. The first is a phenotype-directed multigene panel (SRNS, nuclear mitochondrial diseases, retinitis pigmentosa, ataxia or peripheral neuropathy) that includes the pathway genes and the genes of the differential diagnosis; it reduces the identification of VUS and of findings unrelated to the phenotype, but composition and sensitivity vary between laboratories. The second is the exome, more widely used, or the genome, which do not require choosing the genes [1]. In paediatric SRNS, the IPNA recommends genetic testing in all children with primary SRNS when available (grade B, moderate recommendation), with priority for familial cases, those with extrarenal manifestations and those preparing for transplantation (grade C, weak); it recommends a next-generation panel with all known SRNS genes, including COQ2, COQ6, ADCK4 (COQ8B) and PDSS2, or the single gene when the phenotype is suggestive (grade B); it suggests testing before kidney biopsy in priority cases if the result is available within a few weeks (grade D, weak); and it advises against testing in secondary steroid resistance (grade C, moderate) [36]. Exome sequencing diagnoses a monogenic cause in 10% to 30% of children with SRNS [36]. The 2012 ENMC consensus proposed sequencing the biosynthesis genes in infants with encephalopathy and SRNS and ADCK3 and APTX in ataxic presentations [35]; this gene-by-gene strategy has since been replaced by panels and exome sequencing [1].

5.3. Biochemical tests

Muscle biopsy may show ragged-red fibres and lipid deposits, which are not universal [35]. The ENMC consensus recommended measuring CoQ10 directly in muscle in all biopsies of patients with suspected mitochondrial disease, and not only when I+III or II+III activities are reduced [35]. The 2023 GeneReviews notes that biochemistry has been largely abandoned as a first-line test, because of the scarcity of laboratories and false negatives [1]. The divergence between Saudubray, which values CoQ10 in mononuclear cells for diagnosis, and the Mitochondrial Medicine Society (MMS) and GeneReviews, which consider it insufficient for diagnosing primary defects, is resolved in practice by prioritising molecular testing and reserving tissue biochemistry for selected cases [1, 3, 86].

5.4. Algorithm

In ataxia with CoQ10 deficiency without a molecular diagnosis, the ENMC consensus advocated a therapeutic trial of CoQ10 to identify responders [35].

5.5. Limitations of the tests

Panel and exome sequencing may miss structural, deep intronic and regulatory region variants; the homozygous 9.6 kb duplication in COQ5 was identified only by detailed genetic analysis [1, 17]. More than 50 patients with biochemical CoQ10 deficiency remained undiagnosed after sequencing of all known genes in 2011, which may reflect still unknown genes, regulatory variants or secondary deficiencies [35]. The genes included in panels vary by laboratory and some genes, such as COQ7, have only recently entered neuropathy panels [1, 54]. In silico predictions are unreliable in these genes: the COQ4 variant p.Arg145Gly, predicted as tolerated, is lethal in homozygosity, and the synonymous COQ8A variant c.993C>T, published as pathogenic, is likely benign given its allele frequency of 1.6% [30, 34]. COQ2 nomenclature requires confirming the transcript (MANE NM_001358921.2), because the numbering of the long transcript differs by 150 nucleotides [38, 61].

5.6. Report content and interpretation

The report should state the gene, the reference transcript with version (for example NM_001358921.2 for COQ2), the HGVS description at cDNA and protein level, the zygosity, confirmation of phase by parental testing and the classification according to the criteria of the American College of Medical Genetics and Genomics (ACMG) and the Association for Molecular Pathology (AMP) [1, 36]. When only one parent is a carrier, a de novo variant in the proband, parental post-zygotic mosaicism and uniparental isodisomy should be considered [1]. The GeneReviews authors offer to review VUS in the pathway genes [1].

5.7. Prenatal and preimplantation diagnosis

Once the familial variants have been identified, prenatal diagnosis and preimplantation genetic testing are possible [1, 28]. In 2025, prenatal diagnosis by amniocentesis and trio exome sequencing was published in a fetus with severe growth restriction and cardiac anomalies, compound heterozygous for COQ2 NM_015697.9:c.779-2A>G (= NM_001358921.2:c.629-2A>G), which causes complete skipping of exon 5 with total loss of function, and NM_015697.9:c.973A>G p.(Thr325Ala) (= NM_001358921.2:c.823A>G p.(Thr275Ala)), hypomorphic, with confirmation by minigene assay and yeast complementation [68].

5.8. Diagnosis of HPDL-related disease

There are no formal clinical criteria; the diagnosis is established by identifying biallelic pathogenic or likely pathogenic HPDL variants in a patient with a compatible phenotype, with confirmation of phase by parental testing [21, 33]. All series were diagnosed by exome, genome or linkage study followed by exome sequencing, and older cerebral palsy, epilepsy or mitochondrial disease panels may not include the gene, which was only associated with disease in 2020 [57–59]. Unlike the other genes of the class, there is no routine biochemical marker: metabolic screening (urinary organic acids, plasma amino acids and acylcarnitines) was normal in the infant with the Leigh pattern, respiratory chain activities in that patient's fibroblasts were normal, and muscle biopsy shows non-specific changes [21, 74]. Lactate may be normal outside acute episodes [32]. The sources consulted do not describe measurement of CoQ10 in muscle or fibroblasts of patients with HPDL variants, so the diagnostic value of that test in this defect is unknown. Research markers include plasma 4-HMA, reduced in the Hpdl−/− mouse [6], and, in patients, reduced serum 4-hydroxyphenylacetic acid, a precursor of 4-HMA and 4-HB, and increased plasma glial fibrillary acidic protein [84]; none has been validated for clinical use.

The gene has a single exon, which facilitates exome coverage, but also means that a deletion covering the whole gene is detected only with copy number variation analysis or genome sequencing; the sources consulted do not describe patients with deletions [32, 69]. Missense variants are the majority and many are private, so ACMG/AMP classification depends on segregation, absence or rarity in gnomAD and position relative to the VOC domains and iron-binding residues; the list of predictions for all 2355 possible missense substitutions published by Alecu and colleagues helps to prioritise variants of uncertain significance, but does not replace functional studies [32, 33]. The report should use the MANE transcript NM_032756.4 [61]. Once the familial variants have been identified, prenatal diagnosis and preimplantation genetic testing are possible, as for the other genes of the class [1].

6. Differential Diagnosis

The differential depends on the dominant organ. Because primary deficiency shares manifestations with almost all mitochondrial diseases and with several hereditary nephropathies and ataxias, the diagnosis depends on laboratory tests and, ultimately, on molecular testing [1, 28].

6.1. Primary versus secondary deficiency

Secondary deficiency is the most important differential, because biochemistry does not separate it from primary deficiency [1]. Secondary deficiencies include respiratory chain and mtDNA defects: in a multicentre study of 76 patients with heterogeneous mitochondrial diseases, 28 (37%) had CoQ10 deficiency, nine of them with mtDNA variants, and five patients with MELAS had 40–60% of control values [35]. MADD due to ETFDH explains most cases formerly classified as the isolated myopathic form; the myopathy usually responds to riboflavin, sometimes with additional benefit from CoQ10, and about 15% of patients have early multisystem disease without response [35]. In AOA1, muscle CoQ10 may be low, normal or raised, and three siblings with muscle CoQ10 deficiency and a nonsense APTX variant improved their gait with up to 3000 mg/day [35]. Statins reduce serum and muscle CoQ10 by inhibiting mevalonate synthesis [42].

6.2. SRNS

In a child with SRNS, the presence of sensorineural deafness points to COQ6, neurological manifestations and onset in the first year to COQ2 or PDSS2, and onset in adolescence with isolated kidney disease to COQ8B [1, 31]. The combination of haematuria, proteinuria and deafness in about one in twelve patients with COQ6 may mimic Alport syndrome [31]. SRNS with intellectual disability and cerebellar atrophy also requires consideration of Galloway-Mowat syndrome, whose diagnosis is molecular [87]. The distinction has immediate therapeutic consequences, because in the monogenic forms immunosuppression has low remission rates (complete in 2.7–3.0%, partial in 10.8–16%) and the IPNA suggests stopping it in non-responders, considering CoQ10 in defects of the pathway [36].

6.3. Ataxia

In a child with cerebellar ataxia and cerebellar atrophy, CoQ10 deficiency belongs to the group of treatable ataxias, together with vitamin E deficiency, Refsum disease, cerebrotendinous xanthomatosis and Niemann-Pick disease type C [55]. The combination of exercise intolerance, epilepsy and hyperkinetic movement disorders suggests COQ8A [34]. Oculomotor apraxia with hypoalbuminaemia points to AOA1 [91].

6.4. HPDL-related disease

The differential of HPDL-related disease is that of infantile epileptic encephalopathies with leukoencephalopathy, of spastic cerebral palsy and of early-onset hereditary spastic paraplegias; the absence of SRNS and deafness separates it from most other primary CoQ10 deficiencies.

In a child diagnosed with spastic cerebral palsy, progression of spasticity, postnatal microcephaly, regression, early-onset epilepsy and consanguinity increase the probability of a monogenic cause, and HPDL-related disease belongs to the group of causes that mimic cerebral palsy [33, 56, 84]. In an adolescent with pure spastic paraplegia and normal MRI, HPDL-related disease cannot be distinguished clinically from the other recessive spastic paraplegias, and the diagnosis is molecular [22, 75].

7. Treatments and Risk Management Strategies

7.1. Principles

There is no cure and there are no specific clinical practice guidelines for primary CoQ10 deficiency; GeneReviews states this explicitly in the 2023 revision [1]. The available recommendations come from the consensus of ENMC workshop 176 (2012), the IPNA recommendations for paediatric SRNS (2020), the MMS opinion tables for mitochondrial disease in general (2009 and 2015) and observational series [35, 36, 86, 92]. In 2026, a Delphi panel of 36 experts from 18 European countries considered CoQ10 an essential nutritional product, in the sense of the World Health Organization definition of an essential medicine, for primary deficiency type 1 (COQ2) and type 9 (COQ5), with level 4 evidence on the Oxford Centre for Evidence-Based Medicine scale [94].

Treatment rests on three pillars: high-dose oral CoQ10, started as early as possible and continued for life; treatment directed at the affected organs (nephroprotection, kidney transplantation, antiseizure medications, rehabilitation, nutritional support, cardiology, audiology and ophthalmology); and early assessment of siblings in order to start supplementation before irreversible damage [1, 35, 37]. CoQ10 may limit progression and reverse some manifestations, but does not reverse severe established neurological or renal damage [1, 28, 35].

7.2. Ubidecarenone (coenzyme Q10)

Mechanism and pharmacokinetics. Exogenous CoQ10 is replacement therapy [42, 95]. It is highly lipophilic and practically insoluble in water; less than 5% of the oral dose reaches the plasma, tissue uptake is low and brain uptake is almost nil in rodent studies [35]. It is absorbed via the lymphatic route in micelles and chylomicrons, is distributed mainly to the liver, and is incorporated into membranes before reaching the inner mitochondrial membrane, which explains the slow clinical response [35, 42, 96]. According to the Portuguese summary of product characteristics (SmPC), peak concentration occurs between 5 and 10 hours (mean 6.5 hours) after 100 mg, the half-life is about 34 hours and about 60% of the dose is recovered unchanged in the faeces [96]. The MMS recorded no increase in plasma concentration with 3000 mg/day compared with 2400 mg/day in adults, which suggests saturation of absorption, and a return to baseline levels about two weeks after discontinuation [92]. Solubilised formulations appear more bioavailable, and absorption may improve with meals [1, 92].

Dose. The reference dose is that of the ENMC 176 consensus, which states that oral CoQ10 at 10–30 mg/kg/day in children and 1200–3000 mg/day in adults appears effective in patients with COQ2, especially for neurological and renal manifestations [35]. The other sources diverge, as the table shows.

The dose in the Portuguese SmPC (30 mg/day at any age) corresponds to a fraction of the reference dose; in a 20 kg child, 10 mg/kg/day would be 200 mg/day, that is, 20 capsules of 10 mg [35, 96]. In the 2022 systematic review, the doses used in patients ranged between 60 and 2100 mg/day or between 5 and 100 mg/kg/day, and there was no difference in dose or duration between responders and non-responders [42]. In a 2026 report, a child with COQ2 received 85 mg/kg/day combined with enalapril, with complete remission of proteinuria [98]. These doses above those of the consensus are outside any label and have no dose-comparison studies. In the international renal series, the proteinuria response did not depend on dose, and in univariate analysis higher doses and initiation at a younger age were associated with a higher risk of end-stage kidney disease on treatment (p = 0.021 and p = 0.003), in a retrospective analysis without adjustment for severity [37]. The ENMC consensus and GeneReviews do not state the number of doses; the MMS recommends dividing the daily dose into two or three doses with meals, and the renal series used two to three doses [37, 92].

Ubiquinol. The reduced form is considered by the MMS to be the most bioavailable, with dose adjustment [86]. GeneReviews classifies it among treatments without validated effect in primary deficiency, owing to the lack of dose and efficacy data [1]. In 2016 the EMA granted orphan designation to ubiquinol for primary deficiency (section 7.7) [29]. In the family with COQ5, the three sisters improved their ataxia and cognitive function with ubiquinol 15 mg/kg/day [99].

Adverse effects, precautions and interactions. The Portuguese SmPC mentions epigastric discomfort, nausea, anorexia, diarrhoea and rash in less than 1% of treated patients, and caution in biliary obstruction and hepatic impairment, because of possible accumulation [96]. The MMS mentions wakefulness and sleep disturbance [92]. In the renal series of 116 patients, adverse effects were rare and mild: gastrointestinal disturbances (loss of appetite, nausea, vomiting, altered stools) in five children (4%), with discontinuation after a few months in two [37, 99]; in the COQ8A cohort, two of 30 patients (7%) had anorexia or diarrhoea [34]. No serious adverse effect attributable to CoQ10 was described in the paediatric series reviewed in 2024 [99]. CoQ10 may reduce the anticoagulant effect of warfarin, because of its structural similarity to vitamin K2, which justifies monitoring anticoagulation when they are combined [92, 96].

Monitoring. The MMS suggests measuring the CoQ10 level before treatment and monitoring it in plasma or leucocytes to assess absorption and adherence [86, 92]. Plasma levels are not useful for diagnosis (section 5) [1]. In the renal series, leucocyte CoQ10 rose by 208–238% in three patients measured on treatment, but the authors note that serum and leucocyte levels do not correlate with disease course, and may be normal with active disease or low in remission, and they propose proteinuria as a pharmacodynamic marker for individualising the dose [37]. In the renal forms, response is assessed by proteinuria (urinary protein/creatinine ratio) and renal function [36, 37].

7.3. Response by gene and evidence of efficacy

Response is highly variable and depends on the gene, severity and unknown factors [1]. In COQ2, the severe forms with onset in the first months respond poorly, and neurological manifestations and SRNS respond well in the later-onset forms; response correlates with variant severity [1]. In the renal series, six of 32 treated patients with COQ2 died from progressive neurological deterioration after a median of 0.9 years of treatment (versus 10 of 31 untreated), five had refractory seizures with severe encephalopathy despite treatment, and five improved (tone, strength and motor development in three; disappearance of epilepsy or headaches in two) [37]. In COQ4, patients with a null variant did not respond and those with two missense variants had a partial response [1]; a 12-month-old infant treated with CoQ10 and thiamine had resolution of respiratory difficulty, crying spells and seizures [15]. In COQ6, homozygotes for NM_182476.3:c.763G>A p.(Gly255Arg) or c.1058C>A p.(Ala353Asp) responded well, and the renal response was the best among the genes studied [1]. In COQ8B, a homozygote for a truncating variant had resolution of oedema and significant improvement in proteinuria [1]. In PDSS2 and COQ9, the neonatal forms respond poorly, and the COQ9 cardiac disease worsened on CoQ10 [1, 35]. In COQ8A, GeneReviews describes a weak response in most, with three favourable responses (stabilisation of ataxia, lasting improvement in dystonia and myoclonus, improvement in tremor and drawing), and the OMIM summary, based on the initial descriptions, states absence of significant neurological improvement [1, 12]. The Traschütz cohort provided different data (below). In COQ5, supplementation improved the ICARS of the three sisters after six months at 15 mg/kg/day [5, 53]. In COQ7, CoQ10 added to the culture medium restored proliferation of patients' cells; in the Brazilian patients with neuropathy, low doses (100–500 mg/day, limited by cost) started late showed neither improvement nor worsening during short follow-up [54, 66].

Sources: [34, 35, 37, 42, 100].

These studies have important limitations. None is randomised in primary deficiency; the renal series compares treated patients with matched historical controls, which eliminates neither indication bias nor differences between eras; the COQ8A cohort depends on clinical reports and a small subgroup with longitudinal measurements, and the authors themselves note that natural variation in SARA may simulate improvement; and the systematic review classified patients with minimal effect as non-responders and excluded cases with ambiguous descriptions, which makes it conservative [34, 37, 42]. The relative reduction in proteinuria (88%) has to be read in the light of absolute values: among patients with stage 1–4 chronic kidney disease and continuous treatment, only 32–58% lowered the protein/creatinine ratio to less than 1.5 mg/mg and 23% achieved complete remission; proteinuria was measured by different local methods between patients, and 48 patients already in end-stage kidney disease when starting CoQ10 were excluded from this analysis [37]. By gene, the benefit in renal survival was significant in COQ6 (8 patients, all with preserved function, versus progression of all controls within 2.3 years; p = 0.0049) and only nominal in COQ2 (survival at 2 years of 78% versus 33%; p = 0.282) and COQ8B (55% versus 29.5% at 5 years; p = 0.58), possibly owing to late diagnosis in the latter; combination with an ACE inhibitor or ARB did not add further reduction in proteinuria [37]. The only randomised trial of CoQ10 in children with mitochondrial disease (NCT00432744), not specific to primary deficiency and at a low dose, was negative and underpowered [100]. The conclusions diverge: the authors of the renal series advocate early, lifelong CoQ10 for all patients, whereas Wang and Hekimi conclude that the evidence of efficacy is very weak [37, 42]. The two readings are not incompatible: the benefit appears greater for renal manifestations and late-onset forms, and small or absent in the severe neonatal forms and in established central nervous system damage [1, 35]. Cases of deterioration after discontinuation and improvement on resumption, including two patients with COQ8A with periods on and off treatment, are the most suggestive individual evidence of an effect [34, 42].

7.4. Quinone analogues: idebenone

The Oxford Desk Reference states that rare cases of primary deficiency may respond to ubiquinone or to idebenone, which crosses the blood–brain barrier [40]. The 2023 GeneReviews classifies short-chain quinone analogues, such as idebenone, among ineffective treatments or those without validated effect, and reports that they caused clinical deterioration in patients with CoQ10 deficiency [1]. The ENMC consensus showed in fibroblasts that short-tail analogues do not replace CoQ10 in the respiratory chain [35]. In the review by Wang and Hekimi, a patient with COQ8A worsened (severe exercise intolerance, vomiting, raised lactate) in the four months after CoQ10 was replaced by idebenone and recovered on resuming CoQ10 [42]. Idebenone has demonstrated benefit in Leber hereditary optic neuropathy, which is a different disease [85]. The GeneReviews recommendation, more recent and supported by cellular and clinical data, prevails over that of the Oxford Desk Reference: idebenone should not replace CoQ10 in primary deficiency [1, 40].

7.5. Organ-directed treatment

Kidney. The IPNA recommends starting an ACE inhibitor or an angiotensin receptor blocker (ARB) as soon as SRNS is diagnosed (grade B), aiming at the maximum approved tolerated dose (grade C), with caution in stage 4 chronic kidney disease and discontinuation in volume depletion, acute kidney injury, hyperkalaemia or frequent vomiting and diarrhoea (grade X, strong), preference for ramipril or an ARB, which do not accumulate in renal failure (grade D), and contraception in adolescent girls because of teratogenicity (grade X, strong) [36]. The main text of the IPNA recommendations does not give numerical ACE inhibitor doses; as a paediatric reference, the ESCAPE trial used ramipril at a fixed dose of 6 mg/m² body surface area per day in 385 children aged 3–18 years with chronic kidney disease, with an initial 50% reduction in proteinuria and gradual rebound under prolonged treatment; this is an extrapolation from another nephrological population [101]. In the renal series treated with CoQ10, 34% of patients were receiving renin-angiotensin-aldosterone system inhibitors (29 ACE inhibitor, 2 ARB, 8 combination), three stopped them after remission, and immunosuppression was stopped in 10 of 12 patients between 2 months and 3 years after starting CoQ10 [37]. GeneReviews and the Brazilian textbook state that ACE inhibitors may be combined with CoQ10 in patients with proteinuria [1, 28]. In the monogenic forms of SRNS, immunosuppression rarely induces remission, the genetic forms progress more often to end-stage kidney disease (71–74% versus 4–29%) and have shorter median renal survival (45–48 months versus 58–205 months); the IPNA suggests stopping immunosuppression in non-responders and considering CoQ10 in defects of the pathway [36]. Kidney transplantation is an option in end-stage kidney disease [1, 28]; the three siblings of Rötig's series included two transplant recipients [35]. Genetic testing also allows counselling on the risk of recurrence after transplantation [36]: in the international renal series, 68 patients received a kidney transplant, at a median age of 8.3 (COQ2), 6.4 (COQ6) and 15.1 years (COQ8B), and recurrence of disease in the graft was never observed [31].

Nervous system. Seizures are treated with antiseizure medications chosen according to semiology, by an experienced neurologist [1]. In mitochondrial diseases, valproic acid and aminoglycosides should be avoided [85]. In one patient with COQ4 and infantile spasms syndrome there was a favourable response to corticosteroids [78]. Ataxia, dystonia, spasticity and neuropathy follow usual care, with physiotherapy to maintain strength and mobility and prevent contractures, and positioning equipment [1]. In a 46-year-old woman with COQ8A, 16:8 intermittent fasting with medium-chain triglycerides reduced SARA from 8.5 to 6.0 without reducing myoclonus; this is an isolated report [102].

Other organs. Cardiomyopathy follows usual protocols [1]. Sensorineural deafness is managed by hearing specialists; in COQ6, deafness responded to CoQ10 in 50% of 12 patients in a Korean series treated with 30 mg/kg [1, 99], but in the international renal series, among 11 treated children with audiometry, one improved within 15 months, one worsened and in two deafness was diagnosed for the first time after 8 and 16 months of treatment [37]. Retinopathy is followed by ophthalmology and low vision services [1]. Respiratory failure may require tracheostomy and ventilation [1]. Feeding is supported by a dietitian and gastroenterologist, with nasogastric tube or gastrostomy in dysphagia or risk of aspiration [1].

Development and support. Children with developmental delay or intellectual disability should be referred for early intervention, special education and therapies (physiotherapy, occupational therapy, speech therapy), with planning of the transition to adult life [1]. GeneReviews recommends clinical ethics consultation for difficult decisions in severely affected children, genetic counselling and social and home support for families [1].

7.6. Agents to avoid, pregnancy, perioperative care and transition

Statins inhibit mevalonate synthesis and reduce serum and muscle CoQ10; in mevalonic aciduria, lovastatin reduced ubiquinone by a further 20% and precipitated a severe crisis with rhabdomyolysis [42, 60]. None of the sources consulted makes a specific recommendation on statins in primary deficiency. Idebenone and short-chain analogues should not replace CoQ10 (section 7.4) [1]. Valproic acid and aminoglycosides are not advised in mitochondrial diseases [85]. Combination with warfarin requires monitoring [96].

In pregnancy, animal studies did not show teratogenic effects of ubidecarenone, but there are no studies in pregnant women; the Portuguese SmPC recommends use in the first trimester only if the benefit justifies it and, during breastfeeding, only if strictly necessary, because excretion into milk is unknown [96]. In the renal series, a 27-year-old woman had a miscarriage on CoQ10 and stopped treatment in her next pregnancy until the end of breastfeeding, a period during which proteinuria increased; this is an isolated report that does not establish causality [37]. In patients with SRNS on an ACE inhibitor or ARB, contraception is mandatory because of the teratogenicity of these drugs [36]. In mitochondrial diseases, anaesthesia, surgery and intercurrent illness may cause decompensation, which should be prevented [85]. Transition to adult medicine should ensure continuity of lifelong CoQ10, renal, neurological, auditory and ophthalmological surveillance and reproductive counselling [1, 37].

7.7. Regulatory status

Sources: [4, 29, 38, 96, 103–109].

CoQ10 is, in most countries, a product regulated as a food, which limits access, reimbursement and assurance of supply; the 2026 European Delphi panel proposed that essential nutritional products be recognised as essential medicines and, in some cases, formally authorised as medicines [94]. Cost constrained the doses used in Brazilian patients with COQ7 neuropathy (100–500 mg/day) [66].

7.8. Clinical trials and developments in the last 6–12 months

On 31 October 2026, ClinicalTrials.gov had no interventional trial registered specifically for primary CoQ10 deficiency, for COQ8A ataxia or for the precursors 4-HB and 2,4-dihydroxybenzoate; the existing records are observational (global mitochondrial disease registry with 6000 participants planned up to 2040, the North American NAMDC consortium, the Sanford registry) or bioavailability studies of CoQ10 products in volunteers [100]. The COQ8A cohort calculated that at least 48 patients per arm would be needed to demonstrate a 50% reduction in annual SARA progression [34].

Substrate therapy with 4-HBA in COQ2 deficiency (journal publication and single-patient reports). 4-HB is the natural substrate of COQ2. In the mouse with the Coq2 A252V variant, 4-HBA supplementation completely rescued multisystem disease and perinatal lethality, unlike CoQ10 [110]. In 2025–2026, Distelmaier and colleagues showed that supplementation has to be lifelong (discontinuation in the adult mouse led to neurological decline) and treated the first patient, a 3-year-old child described in section 10, with 75 mg/kg/day and then 150 mg/kg/day of oral 4-HBA in three doses, with remission of proteinuria within three weeks, progressive normalisation of lactate and motor and cognitive improvement, without adverse effects over six months [38]. A second patient, with the severe neonatal form and no benefit from CoQ10, normalised severe lactic acidosis and recovered renal function; according to the Mantle review, 4-HBA was tolerated up to 166 mg/kg/day in divided doses in a newborn [5, 111]. The authors restrict treatment to cases with confirmed in vitro response, because 4-HBA acts upstream and does not bypass distal blocks in the pathway, and may worsen DMQ accumulation in the Coq9 R239X model; long-term safety, including carcinogenic potential, has not been assessed, and the powder is irritant to skin and eyes [38]. In healthy adults, oral 4-HB was administered at up to 5 g every 6 hours for 24 hours without significant adverse effects [5].

4-HB and 4-HMA in HPDL deficiency (preclinical and single patient). In the Hpdl−/− mouse, 4-HMA and 4-HB were incorporated into brain CoQ9 and CoQ10 and allowed 90–100% of animals to reach adulthood; 4-HB stabilised and improved the progressive spasticity of a patient with biallelic HPDL variants [6]. In mice, supplemental CoQ10 did not increase brain pools [6].

Other precursors and combinations (preclinical). 2,4-Dihydroxybenzoate (β-resorcylic acid) bypasses COQ7 and COQ9 deficiencies in cells and mice, prevents renal pathology in the Adck4 (COQ8B) mouse when given before severe symptoms and prevents FSGS in the mouse with podocyte deletion of Coq6; vanillic acid bypasses COQ6 and COQ9 deficiencies; and the combination of CoQ10 with vanillic acid prolonged survival and improved motor function in the Coq9 R239X mouse more than either compound alone [5, 112]. In cells from patients with COQ7, the response to 2,4-dihydroxybenzoate depended on the variant: it restored synthesis in the severe variant p.Val141Glu, but not in the milder p.Leu111Pro (published protein designations) [5]. There are no randomised trials with these compounds in any indication [5].

BPM31510 (preclinical and press release). In the Coq4 F147C mouse, the lipid nanoparticle formulation of CoQ10 increased CoQ10 in the brain, including the cerebellum, in the myocardium and in the renal cortex [113]. The FDA orphan designations database records the designation of 13 January 2026; the sponsor announced «preliminary clinical data from multiple compassionate use cases», without published figures [104, 109].

Regulatory designations of 4-HBA (designation, not authorisation). On 16 April 2025 the EMA granted orphan designation EU/3/25/3052 to para-hydroxybenzoic acid for primary ubiquinone deficiency, at the request of the University of Granada, and between January and February 2026 the FDA granted three orphan drug designations to 4-HBA in primary CoQ10 deficiency, to academic groups in San Diego, Granada and New York [104, 105].

Other 2025–2026 publications. The review of 28 patients with COQ7 neuropathy proposed DMQ10 and CoQ10 in fibroblasts as biomarkers [67]; the series of 69 patients with COQ4 broadened the epileptic spectrum [78]; the Brazilian COQ7 series associated the variant NM_016138.5:c.3G>T p.(Met1?) with earlier onset and Brazilian origin with cerebellar manifestations [66]; and the European Delphi panel classified CoQ10 as an essential product [94].

7.9. Summary of recommendations and divergences

7.10. HPDL-related disease

Supportive treatment. There is no authorised treatment or specific guideline, and usual treatment is multidisciplinary and symptomatic: control of epilepsy, which in the severe forms generally requires several antiseizure medications and is often refractory; treatment of spasticity and contractures with physiotherapy, orthoses and the drugs and procedures used in other spastic paraplegias; nutritional support, with gastrostomy when there is dysphagia and aspiration; and respiratory support during episodes of apnoea or respiratory failure [21, 32, 33]. As fever and infections trigger decompensations, prevention and early treatment of infections and a written plan for intercurrent illness follow the general principles of mitochondrial diseases, including avoiding valproic acid [32, 85]. Strabismus can be corrected surgically, as in a child with SPG83 and constant exotropia, without recurrence one year later [82].

Ubidecarenone (CoQ10). There are no efficacy data for CoQ10 in HPDL-related disease. In the Hpdl−/− mouse, CoQ9, CoQ10 and related compounds did not improve pup survival, and supplemental CoQ10 did not increase brain pools [6]. The only patient with published data worsened during six weeks of oral CoQ10, and CoQ10 was stopped a few months after 4-HB was started [6, 21]. If CoQ10 is used, the reference dose is that of primary deficiency in general (10–30 mg/kg/day in children and 1200–3000 mg/day in adults, according to the ENMC consensus; section 7.2), applied by extrapolation from the other genes of the class; the mechanism of the defect, upstream of COQ2, and the mouse data do not support an expectation of neurological benefit [6, 35].

Substrate therapy with 4-HB and 4-HMA. The strategy is a metabolic bypass therapy: providing orally the product of the missing reaction (4-HMA) or the next compound in the pathway (4-HB), small molecules of less than 200 Da that cross the blood–brain barrier and are incorporated into cellular CoQ [6]. The preclinical methodology of Shi and colleagues combined isotopic labelling (¹³C₆-4-HMA and ¹³C₆-4-HB) to demonstrate incorporation into brain CoQ9 and CoQ10, survival curves, histology, electron microscopy of mitochondria, Purkinje cell electrophysiology and motor tests [6]. With 10 mg/kg/day orally started between days 3 and 7 of life and continued until day 30, 90–100% of Hpdl−/− mice reached adulthood, versus death of all untreated mice by day 15; 4-HMA improved mitochondrial and cerebellar morphology, electron transport chain activity and the functional maturation of Purkinje cells and reduced seizures, but reduced grip strength and altered respiratory quotient persisted in males [6]. Initiation after day 10 did not improve survival, which the authors interpret as a critical window of brain development [6]. In the hpdl-F0 zebrafish and in patients' cells, 4-HMA corrected the phenotypes and 4-HB did so only partially in the fish [84], and in cortical progenitors derived from patients, antioxidants and CoQ10 intermediates partially corrected the premature neurogenesis, with a genotype-dependent response [72]. The clinical evidence is limited to one patient treated with 4-HB under an FDA single-patient investigational new drug authorisation (IND 170119), described in section 10, who stabilised and then improved in spasticity and motor function over 250 days, with no toxicity detected on electrocardiograms and laboratory tests; the only adverse effect was vomiting with the large volume of solution during the 100 mg/kg phase [6]. The authors of the clinical treatment wrote that part of the article independently, under an institutional conflict of interest management plan, because other authors are co-inventors on patents for the use of 4-HMA and 4-HB [6]. The risks have not yet been characterised: there are no published human pharmacokinetic data, long-term safety data, optimal doses or efficacy data in the neonatal forms, in which the critical window observed in the mouse may already have passed at the time of diagnosis.

These doses come from a single patient and not from a guideline, consensus or trial; 4-HB has no marketing authorisation in any jurisdiction and therefore there is no label (EMA, FDA, INFARMED or ANVISA) with which to compare [6, 104, 105].

Ketogenic diet. A Japanese infant with the variants NM_032756.4:c.149_151del p.(Gly50del) and c.537G>A p.(Trp179Ter), refractory cyanotic apnoeas and Leigh-pattern lesions started a ketogenic diet at 10 months; within three months lactate and pyruvate normalised, and head control, social smiling and visual tracking were regained, and at 18 months the putaminal and brainstem lesions had disappeared, with reduced lactate on spectroscopy [74]. This is an isolated report, without comparison, and the diet is not an established treatment.

Immunomodulation. The improvement with corticosteroids or immunoglobulin described in a patient with a presentation resembling autoimmune encephalitis was transient, followed by relapses, and does not constitute treatment of the disease [83].

Regulatory status and trials. There is no authorised medicine for HPDL-related disease. The orphan designations of 4-HB granted by the EMA (EU/3/25/3052, 16 April 2025) and by the FDA (three designations between 16 January and 15 February 2026, one of them to Gleeson's laboratory at the University of California San Diego and another to New York University) refer to primary CoQ10 deficiency in general and not to HPDL-related disease in particular; an orphan designation is not an authorisation (section 7.7) [104, 105]. On 1 November 2026, ClinicalTrials.gov had a single dedicated study, the observational natural history registry NCT05848271 of the University of California San Diego, recruiting since 18 May 2023, with 50 participants planned and last updated on 30 March 2025, created to prepare a future trial; there was no registered interventional trial with 4-HB or 4-HMA in this disease [33, 114].

Developments in the last 12 months. Between November 2025 and October 2026, the studies in patient-derived cortical organoids (preclinical study) [72], a review with a new Iranian variant and calculation of the consanguinity rate (case series and review) [73] and three cases of SPG83 in a Mexican family (case series) [80] were published; the article by Shi and colleagues, with the first treated patient, was published online in July 2025 and in print in September 2025 (single patient) [6]. No trial results, opinions or authorisations for this disease were found in the sources consulted.

8. Reproductive Options and Family Counselling

8.1. Recurrence risks

Primary CoQ10 deficiency is, as a rule, transmitted in an autosomal recessive manner [1]. The parents of a patient with a molecular diagnosis are presumed to be heterozygous, and parental testing is recommended to confirm that both are carriers and to allow reliable risk calculation [1]. If only one parent is a carrier, after maternity and paternity have been confirmed, a de novo variant in the proband, a post-zygotic variant in a mosaic parent and uniparental isodisomy should be considered [1]. When both parents are heterozygous, each sibling has, at conception, a 25% chance of being affected, 50% of being an asymptomatic carrier and 25% of being neither affected nor a carrier [1]. The Brazilian textbook on genetic syndromes presents these risks with wording that mixes the categories and describes the inheritance as «autosomal recessive linked to mitochondrial DNA»; the genes of the class are all nuclear, and the correct risks are those of GeneReviews [1, 28]. Heterozygotes are asymptomatic and are not at risk of developing the disease [1]. The children of a patient are obligate carriers, and each sibling of the parents of a patient has a 50% chance of being a carrier [1]. The risk to the offspring of a patient depends on the partner's status; consanguinity and populations with founder variants (for example COQ4 c.370G>A in southern China or the COQ5 duplication in the Iraqi Jewish population) increase the probability that the partner is a carrier [17, 30, 65].

COQ4 haploinsufficiency due to a de novo contiguous gene deletion is an exception to recessive inheritance; in that scenario, the recurrence risk for siblings depends on confirmation that the deletion is absent in the parents [1, 8]. COQ2-related susceptibility to multiple system atrophy has no established recurrence risks, because the association is controversial [9].

8.2. Cascade testing, siblings and carriers

Given the importance of starting CoQ10 early, GeneReviews recommends evaluating the siblings of a patient as early as possible, by molecular testing when the familial variants are known, or by measuring CoQ10 in fibroblasts when the diagnosis is only biochemical [1]. The IPNA suggests offering urinalysis to siblings of patients with SRNS even before genetic testing (grade C) [36]. Carrier testing in relatives requires prior identification of the variants in the family [1]. When a molecular diagnosis has not been reached, DNA banking of the proband for future reanalysis should be considered [1].

Testing of asymptomatic minor siblings has, in this disease, direct medical benefit, because it allows treatment to be started before organ damage [1].

8.3. Prenatal and preimplantation diagnosis

Once the familial pathogenic variants have been identified, prenatal diagnosis and preimplantation genetic testing are possible; perspectives on their use may vary among professionals and within families [1]. The Brazilian textbook stresses that the best time to discuss these risks is during family and pregnancy planning [28]. A 2025 case shows prenatal diagnosis of COQ2 by trio exome sequencing in a fetus with severe growth restriction and cardiac anomalies, with functional characterisation of the variants [68]. In the forms that manifest prenatally, ultrasound signs (growth restriction, oligohydramnios or anhydramnios, fetal hypertrophic cardiomyopathy) may lead to diagnosis in the first affected pregnancy [1, 13, 16].

8.4. Counselling in HPDL-related disease

HPDL-related disease is autosomal recessive: when both parents are heterozygous, each pregnancy has a 25% chance of an affected child, 50% of an asymptomatic carrier and 25% of a non-carrier child [1, 21, 22]. Consanguinity is present in about half of the families described (49.55%), which makes carrier testing relevant in relatives planning unions within the family [73]. The genotype guides, but does not reliably predict, severity: homozygosity for p.(Gly50Asp) was associated with the mild form in 13 of 14 patients and biallelic truncating variants with the severe form in 75% [33], but siblings with the same genotype had pure spastic paraplegia and a complicated form [75], and a boy with symptoms at 8 years had the same genotype as his brother who died in the first year with encephalopathy [6]. Counselling should therefore present the possible spectrum within the same family. Testing of asymptomatic minor siblings does not, in 2026, have an associated authorised treatment, but the critical window observed in the mouse and the existence of a natural history registry and of individual use of 4-HB give it potential clinical utility, which should be discussed with the family [6, 114]. Once the familial variants have been identified, prenatal diagnosis and preimplantation genetic testing are possible [1].

9. Special Considerations: Population Diversity and Epidemiology

9.1. Prevalence

There is no prevalence estimate validated by population studies. MedlinePlus estimates prevalence at less than 1 per 100,000 people [27], the Brazilian textbook gives an incidence below 1 per 100,000 [28], and the 2016 EMA orphan designation used about 0.1 per 10,000 in the European Union, equivalent to approximately 5000 people in a population of 513.7 million [29]. Orphanet classifies the prevalence of the group as unknown [4]. GeneReviews has no prevalence section and documents rarity by the number of families described per gene [1]. By May 2022, 303 patients had been published [42], and the 2023 review counted 63 patients with COQ2, 44 with COQ4, 45 with COQ6, 123 with COQ8A, 99 with COQ8B, 7 with PDSS2, 7 with COQ9 and 6 with PDSS1 [5]. In paediatric nephrology, primary deficiency explains 1–2.7% of SRNS and up to 10% of SRNS with an identified genetic cause; the registries cited by the renal series attribute to it about 0.5% of mitochondrial diseases and 2% of SRNS or persistent proteinuria in children, and a Chinese multicentre study identified COQ8B as the most frequent cause of SRNS in its region (5.7% of cases) [31]. These proportions measure relative frequencies in cohorts of referred patients, not population prevalence.

The only genomic estimate is that of Hughes and colleagues (2017), who calculated the birth prevalence of homozygotes and compound heterozygotes from gnomAD allele frequencies (138,632 individuals without severe paediatric disease), assuming Hardy-Weinberg equilibrium [30]. With the known pathogenic variants, they estimated 1665 affected people worldwide and 192 in a population with the size and composition of the USA; with variants predicted to be pathogenic (high-confidence loss of function and missense variants considered deleterious by two algorithms), 123,789 worldwide and 1462 in the USA; COQ8A and COQ8B contributed 46,654 of the 123,789 [30]. The estimate measures genotypes at birth and not diagnosed disease; it may overestimate (embryonic lethality of some loss-of-function combinations, missense variants predicted deleterious but benign) and underestimate (COQ4 haploinsufficiency, consanguinity, variants not yet described) [30]. The ENMC consensus considered the deficiency probably underdiagnosed [35].

9.2. Differences between populations

In the Hughes analysis, inclusion of the predicted variants multiplied the estimated number of patients 3.5-fold in Western populations (Ashkenazi Jews, Finns and other Europeans) and 122-fold in non-Western populations (South and East Asians, Latin Americans and Africans), which the authors interpret as underdiagnosis in populations with less access to clinical sequencing [30]. The founder populations described are those of southern China and Taiwan (COQ4 c.370G>A, present in 10 of 11 patients) [65], the Jewish population of Iraqi origin (COQ5 duplication, allele frequency of about 1.5% in 92 controls) [17] and Korea (COQ6 p.Lys64del with p.Pro261Leu in five of six patients) [14]. The common COQ2 variant V343A was observed exclusively in the Japanese population in the 2013 multiple system atrophy study [64]. The COQ2 variant NM_015697.9:c.973A>G is more frequent in East Asia [68]. The first patients with PDSS1 and COQ9 were from Moroccan, Pakistani and Turkish families, several of them consanguineous [10, 13]. The international renal series observed a higher frequency of pathogenic variants in the CoQ10 genes in Asian populations, under-represented in European-based registries, and recurrent variants with regional distribution in Turkey, the Middle East, Kazakhstan, Iran, Central and Eastern Europe and China (section 2.8) [31].

In the USA, the genomic estimate for a population with the North American composition was 192 to 1462 people [30]. In Europe, the renal series integrated the PodoNet and mitoNET registries, and the COQ8A cohort brought together 15 centres in Europe and North America, with patients from 16 countries [31, 34].

Portugal. COQ7-related distal motor neuronopathy was first described in three siblings born to unrelated Portuguese parents, homozygous for c.3G>T p.(Met1?) [20, 54]. In Portugal, ubidecarenone is authorised as a medicine (Q10® 10 mg), with an indication in mitochondrial cytopathies [96].

Brazil. A 2026 series from São José do Rio Preto described five patients from three unrelated families with the same COQ7 variant c.3G>T p.(Met1?) in homozygosity, adding to five Brazilian families already described with this variant; the variant was associated with earlier onset, cognitive impairment and proximal lower limb weakness, and Brazilian origin with cerebellar manifestations, in exploratory analyses with wide confidence intervals [66]. The sources consulted did not study whether the variant has a common origin in the Portuguese and Brazilian families. In Brazil, no medicine containing ubidecarenone has an active ANVISA registration [107]. The Brazilian textbook on genetic syndromes devotes a subchapter to the disease [28].

HPDL-related disease. Orphanet classifies the prevalence of the two entities as below 1 per 1,000,000, with 32 cases recorded for the neurodevelopmental form and 16 for SPG83 in the sources it used [24, 25]; the 2025 reviews brought together 91 and 99 published patients [32, 33]. Shi and colleagues estimated, from gnomAD allele frequencies and assuming Hardy-Weinberg equilibrium, an incidence of genotypes homozygous for variants classified as pathogenic of about 1 in 155 million, and for likely pathogenic variants between 1 in 86,319 and 1 in 19.3 million depending on the assumption; these are model estimates, which do not count compound heterozygotes or variants not yet described, and do not measure prevalence [6]. Published patients are concentrated in populations with frequent consanguinity: Egypt, Sudan, Syria, Turkey and Iran, with a founder effect of p.(Gly50Asp) in the Middle East, Pakistan (family of the former CPSQ1) and Algeria [32, 56, 57, 59, 73]. In Europe, the series included patients from Germany and other European countries, and NM_032756.4:c.232G>A p.(Ala78Thr) was described in an Irish patient [32, 57, 58]. Outside these regions there are patients from China, Japan and Mexico [70, 74, 75, 80]. In the USA, the natural history registry of the University of California San Diego recruits patients from several countries [33, 114]. The sources consulted do not describe patients from Portugal or Brazil. The relevant European networks are ERN-RND (spastic paraplegias and rare neurological diseases) and MetabERN (mitochondrial diseases) [39].

The sources consulted do not provide prevalence estimates specific to Portugal, Brazil or individual European countries.

9.3. Special populations

Patients with primary deficiency diagnosed in adulthood (retinitis pigmentosa, ataxia, motor neuropathy, spastic paraparesis, multiple system atrophy-like phenotype) benefit from the same screening for renal, auditory and ophthalmological involvement as children [1]. In women of childbearing age, pregnancy requires review of teratogenic nephroprotective medication (ACE inhibitor, ARB) [36] and consideration of the limited safety data on ubidecarenone in pregnancy [96]. The authors of the international renal series recommend lifelong CoQ10 in all diagnosed patients, to delay kidney disease and prevent damage to other organs [37].

9.4. Reference networks and patient associations

The relevant European reference networks are MetabERN, for inherited metabolic diseases, including mitochondrial diseases; ERKNet, for rare kidney diseases and SRNS; ERN-RND, for ataxias and rare neurological diseases; and ERN EURO-NMD, for neuromuscular diseases [39]. Patient associations include the United Mitochondrial Disease Foundation (USA), The Lily Foundation and the Children's Mitochondrial Disease Network (United Kingdom; cited by the Oxford Desk Reference) and International Mito Patients [39, 40]. MedlinePlus has a dedicated page [27], and the online platform of Alcázar-Fabra and colleagues brings together the published clinical manifestations by gene and is updated periodically [115].

10. Illustrative Clinical Case

The case was published by Distelmaier and colleagues in the journal Brain (published online in September 2025; 2026 issue) and is the first patient treated with 4-hydroxybenzoic acid [38]. Figure 4 of the article shows the MRI, the enzyme model, the laboratory data and the clinical course: https://pmc.ncbi.nlm.nih.gov/articles/PMC13140534/figure/awaf334-F4/.

The patient is the first child of healthy, non-consanguineous parents of German ancestry, with uneventful pregnancy and delivery. In the first year developmental delay and hypotonia became evident: he sat unsupported at 12 months, crawled at 17 and, at 2 years and 6 months, said about 15 words. He had recurrent vomiting, poor weight gain and persistent lactic acidosis. At 2 years and 8 months generalised oedema appeared and nephrotic syndrome was diagnosed, without response to corticosteroids [38].

At 3 years, exome sequencing identified two compound heterozygous COQ2 variants: NM_001358921.2:c.421G>C p.(Val141Leu), of maternal origin, absent from gnomAD, classified as likely pathogenic (old nomenclature NM_015697.9:c.571G>C p.Val191Leu), and NM_001358921.2:c.542G>T p.(Ser181Ile), of paternal origin, already described as pathogenic (old c.692G>T p.Ser231Ile); the younger brother had neither variant [38]. Both descriptions were confirmed in VariantValidator [61]. MRI at 3 years and 3 months showed symmetrical thalamic lesions, in a Leigh-like pattern, and a moderate lactate peak on spectroscopy (figure 4A). The diagnosis was primary CoQ10 deficiency type 1 [38].

Oral CoQ10 was started at a dose of 30 mg/kg/day, which is within the range of the ENMC consensus, with ramipril for proteinuria [35, 38]. Within three months, the parents noticed slight improvements (more activity, less vomiting, less apathy), without major clinical change, and renal hyperfiltration decreased slightly [38]. The patient's fibroblasts had CoQ10 deficiency, which increased markedly with 4-HBA in vitro (figure 4C). With approval from the ethics committee of Düsseldorf University Hospital, review of the protocol by the hospital pharmacy and written consent from the family, an individual therapeutic trial was started [38].

At baseline, lactate was 5–6 mmol/l, base deficit about −10 mmol/l, serum CoQ10 and IgG were low, the urinary albumin/creatinine ratio was 922.8 mg/mmol and the protein/creatinine ratio 1086.1 mg/mmol, with raised glomerular filtration rate; abdominal ultrasound and echocardiogram were normal [38]. CoQ10 was stopped and 4-HBA, encapsulated by the hospital pharmacy and dissolved in milk to attenuate the acidity, was started at 3 years and 3 months at a dose of 250 mg three times a day (75 mg/kg/day), increased after one week to 500 mg three times a day (150 mg/kg/day) [38]. After one month, lactate had fallen to about 3 mmol/l; albuminuria and proteinuria decreased markedly within three weeks, which allowed ramipril to be stopped; serum CoQ10 rose progressively above 1 µmol/l and IgG tended to normalise [38]. There were no adverse effects over six months, with normal full blood count, electrolytes, liver function and coagulation [38].

Clinically, strength, mobility, exercise tolerance and fine motor skills improved; vomiting stopped and the child gained 2 kg in four months; after four months he walked unsupported over long distances, expanded his vocabulary to 70–80 words with two-word sentences, and his score on the Newcastle Paediatric Mitochondrial Disease Scale decreased [38]. Treatment was continuing at nine months and was planned for life [38].

The case illustrates the typical course of COQ2 deficiency of intermediate severity (delay, hypotonia, lactic acidosis, Leigh lesions and SRNS in early childhood) and the limited response to conventional CoQ10 when the central nervous system is involved [1, 38]. 4-HBA is not routine treatment: it is an unauthorised substance, used in a single patient, with short follow-up, whose long-term safety has not been established, and the authors restrict it to upstream defects of the pathway with confirmed in vitro response [38]. The standard of care remains high-dose oral CoQ10, nephroprotection and supportive treatment [1, 35, 36].

Complementary case: HPDL-related disease treated with 4-HB. In 2025 Shi and colleagues described, in the journal Nature, a boy with normal development up to 8 years who was running and playing sport in August 2023, developed bilateral ankle clonus in September and could no longer run in October [6]. He started CoQ10 in mid-October; in November he could stand unsupported, but had difficulty standing with his feet together, and at the end of that month he was falling frequently and the physiotherapists advised a wheelchair [6]. Two younger siblings had died in the first year of life with rapidly progressive encephalopathy and seizures; the second had the variants NM_032756.4:c.3G>C p.(Met1?) and c.789del p.(Pro264LeufsTer51), both classified as pathogenic in ClinVar, and the proband had the same genotype [6, 61]. Initial MRI showed irregular T2 hyperintensity of the corticospinal tract, of uncertain significance, without major cerebellar changes [6].

Given the rapid progression during 1.5 months of CoQ10, the brother's phenotype and the availability of pure 4-HB, the team obtained an FDA single-patient investigational new drug authorisation and approval from the New York University ethics committee, with parental consent and the child's assent [6]. 4-HB was given as a 5 mg/ml aqueous solution, starting with 30 mg on the first day and increasing over four days to 100 mg/kg, maintained until day 30; this was followed by 50 mg/kg for 95 days and then 25 mg/kg maintenance, and CoQ10 was stopped after a few months [6]. The sum of the modified Ashworth scale fell from 10 to 6 (day 30) and the Spastic Paraplegia Rating Scale (SPRS) score from 13 to 7 (day 250); the nine-hole peg test stabilised at age-appropriate times, and the 10-metre walk test varied, partly because the child compensated for imbalance by walking faster [6]. Falls went from several a day to rare, he was running again by around day 54, cycling on day 124 and, at 8 months, moving sideways to catch a ball; MRI was unchanged at 30 days and ankle clonus persisted [6]. Vomiting during the 600 ml solution phase was the only adverse effect, and electrocardiograms and laboratory tests showed no abnormalities [6].

The case shows the variability within the family (onset at 8 years in a boy with the genotype of a brother who died in infancy) and the first clinical evidence that ring intermediates can modify the course, but 4-HB is not routine treatment: it is an unauthorised substance, used in a single patient, without a control group, with 250 days of follow-up, and interpretation of the scales in a context of recent progression has limitations [6].

11. Prognosis and Long-Term Follow-Up

11.1. Prognosis

Prognostic data are limited by the small number of patients. The disease is progressive, with speed and organs involved dependent on the gene and the severity of the deficiency [1]. Children with severe multisystem disease respond poorly to treatment and generally die in the neonatal period or in the first year of life [1, 28]; in the southern Chinese COQ4 series, five of eleven patients had the classic neonatal form [65], and in the Chung series the six patients died between 36 hours and 19 months [15]. Patients with later onset respond better to CoQ10, which in many cases changes the natural history by halting renal progression and preventing neurological manifestations in COQ2, COQ6, COQ8B and PDSS2 deficiencies [1].

In kidney disease, the probability of end-stage kidney disease at five years was 47.6% for COQ2, 47.8% for COQ6 and 3.1% for COQ8B, and at 18 years 80.8%, 72.3% and 74.2%; renal survival in adulthood was 20–25% in the three genes, and there was no graft recurrence in 68 transplant recipients [31]. Overall survival at 10 years was 69.9% for COQ2, 90.8% for COQ6 and 99.2% for COQ8B; in COQ2, deaths were concentrated in the first year due to multiorgan failure [31]. CoQ10 was associated with kidney failure-free survival at five years of 62%, versus 19% in a matched untreated cohort [37]. In COQ8A ataxia, progression is mild to moderate (about 0.45–0.47 SARA points per year), slower than in Friedreich ataxia (0.8 per year); 82% of patients had mild to moderate ataxia and 11% depended on two walking aids or a wheelchair, most after 50–60 years, but some patients had a severe course with death in youth [34]. In COQ7 neuronopathy, progression is slow; one patient lost independent walking at 41 years [20]. Prognosis also depends on the timing of diagnosis: treatment started before severe damage has better results, and presymptomatic diagnosis of siblings is the main opportunity for prevention [1, 35].

In HPDL-related disease, prognosis depends on the clinical class: all published deaths occurred in the severe form (10 of 91 patients in the Alecu review; 12 of 48 severe patients in the Lee review), with a median age at death of 18 months and an estimated median survival of 11 years in that group, mainly due to status epilepticus and acute encephalopathy after fever or infection [32, 33]. In Alecu's intermediate group (median onset at 8.5 months), no patient walked unsupported at last follow-up, whereas the nine intermediate patients in the Wiessner series, with later onset (3–7 years), all achieved independent walking; the difference reflects different classification criteria [21, 32]. In the mild form there were no deaths, but 25.0% lost independent walking [32]. These figures come from cross-sectional data and not from prospective follow-up, and the NCT05848271 registry was created to obtain longitudinal data [32, 114].

11.2. Surveillance table

The sources consulted do not define numerical intervals for renal, ophthalmological and auditory surveillance; the intervals depend on the gene, the manifestations and the response to treatment [1].

12. Summary and Final Considerations

Primary CoQ10 deficiency is a class of autosomal recessive disorders of ubiquinone biosynthesis, caused by biallelic variants in PDSS1, PDSS2, COQ2, COQ4, COQ5, COQ6, COQ7, COQ8A, COQ8B or COQ9, with the HPDL defect as a related entity. The estimated prevalence is below 1 per 100,000, but analysis of genomic databases suggests underdiagnosis, especially outside European populations. The spectrum ranges from neonatal multisystem disease, with lactic acidosis, hypertrophic cardiomyopathy and encephalopathy, to SRNS of childhood and adolescence, cerebellar ataxia, distal motor neuronopathy, spastic paraparesis and adult retinitis pigmentosa. SRNS with deafness suggests COQ6; isolated SRNS of adolescence, COQ8B; ataxia with exercise intolerance, epilepsy and dystonia, COQ8A; distal motor neuropathy with pyramidal signs, COQ7.

The diagnosis is molecular, by phenotype-directed panel or by exome sequencing; muscle CoQ10 and the combined activities of complexes I+III and II+III serve to support variants of uncertain significance, but do not distinguish primary from secondary deficiency, and normal values do not exclude the disease. Plasma CoQ10 has no diagnostic value. In COQ2 nomenclature it is important to use the MANE transcript, because the old numbering differs by 50 amino acids.

The reference treatment is high-dose oral ubidecarenone, 10–30 mg/kg/day in children and 1200–3000 mg/day in adults according to the ENMC 176 consensus, started as early as possible and continued for life, with an ACE inhibitor or ARB for proteinuria and transplantation in end-stage kidney disease. The response is partial and depends on the gene and the time of initiation: it is more consistent for renal manifestations and in the late-onset forms, and small in the neonatal forms and in established neurological damage. Idebenone should not replace CoQ10. The siblings of a patient should be tested soon after diagnosis of the index case. Substrate therapies with ring precursors, in particular 4-hydroxybenzoic acid in COQ2 and HPDL deficiency, have shown responses in single patients and in animal models, but have not yet been evaluated in clinical trials; 4-HBA received orphan designation in the European Union in 2025 and in the USA in 2026, without marketing authorisation.

HPDL-related disease completes the class upstream of COQ2: HPDL produces the 4-hydroxymandelate that gives rise to 4-HB, and biallelic variants cause an exclusively neurological spectrum, from neonatal epileptic encephalopathy with microcephaly and leukoencephalopathy, with mortality of about one quarter in the severe form, to an intermediate form that mimics cerebral palsy and pure adolescent spastic paraplegia (SPG83); biallelic truncating variants are associated with the severe forms and homozygous p.(Gly50Asp) with the mild form. There is no routine biochemical marker and the diagnosis is made by exome sequencing. 4-HMA and 4-HB bypass the block in the mouse, within an early window, and 4-HB improved an 8-year-old patient, but they remain experimental, without a registered clinical trial.

13. References

1. Salviati L, Trevisson E, Agosto C, Doimo M, Navas P. Primary coenzyme Q10 deficiency overview. 2017 Jan 26 [updated 2023 Jun 8]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK410087/
2. OMIM® Online Mendelian Inheritance in Man. #607426 COENZYME Q10 DEFICIENCY, PRIMARY, 1; COQ10D1 [Internet]. Baltimore (MD): Johns Hopkins University; last edited 13 Mar 2025. Available from: https://omim.org/entry/607426
3. 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.
4. GeneGuidance. Records for the genes COQ2, COQ4, COQ5, COQ6, COQ7, COQ8A, COQ8B, COQ9, PDSS1 and PDSS2 [database]. Export of 2026-09-26.
5. Mantle D, Cufflin N, Hargreaves IP. Bypass Treatments for Primary Coenzyme Q10 Deficiency: An Update. Int J Mol Sci. 2026;27(8):3526. doi:10.3390/ijms27083526
6. Shi G, Miller C, Kuno S, Rey Hipolito AG, El Nagar S, Riboldi GM, et al. Coenzyme Q headgroup intermediates can ameliorate a mitochondrial encephalopathy. Nature. 2025;645(8080):466-474. doi:10.1038/s41586-025-09246-x
7. Banh RS, Kim ES, Spillier Q, Biancur DE, Yamamoto K, Sohn ASW, et al. The polar oxy-metabolome reveals the 4-hydroxymandelate CoQ10 synthesis pathway. Nature. 2021;597(7876):420-425. doi:10.1038/s41586-021-03865-w
8. Salviati L, Trevisson E, Rodriguez Hernandez MA, Casarin A, Pertegato V, Doimo M, et al. Haploinsufficiency of COQ4 causes coenzyme Q10 deficiency. J Med Genet. 2012;49(3):187-191. doi:10.1136/jmedgenet-2011-100394
9. OMIM® Online Mendelian Inheritance in Man. #146500 MULTIPLE SYSTEM ATROPHY 1, SUSCEPTIBILITY TO; MSA1 [Internet]. Baltimore (MD): Johns Hopkins University; last edited 23 Jun 2020. Available from: https://omim.org/entry/146500
10. OMIM® Online Mendelian Inheritance in Man. #614651 COENZYME Q10 DEFICIENCY, PRIMARY, 2; COQ10D2 [Internet]. Baltimore (MD): Johns Hopkins University; last edited 19 Apr 2023. Available from: https://omim.org/entry/614651
11. OMIM® Online Mendelian Inheritance in Man. #614652 COENZYME Q10 DEFICIENCY, PRIMARY, 3; COQ10D3 [Internet]. Baltimore (MD): Johns Hopkins University; last edited 27 Mar 2015. Available from: https://omim.org/entry/614652
12. OMIM® Online Mendelian Inheritance in Man. #612016 COENZYME Q10 DEFICIENCY, PRIMARY, 4; COQ10D4 [Internet]. Baltimore (MD): Johns Hopkins University; last edited 25 Sep 2020. Available from: https://omim.org/entry/612016
13. OMIM® Online Mendelian Inheritance in Man. #614654 COENZYME Q10 DEFICIENCY, PRIMARY, 5; COQ10D5 [Internet]. Baltimore (MD): Johns Hopkins University; last edited 11 Oct 2016. Available from: https://omim.org/entry/614654
14. OMIM® Online Mendelian Inheritance in Man. #614650 COENZYME Q10 DEFICIENCY, PRIMARY, 6; COQ10D6 [Internet]. Baltimore (MD): Johns Hopkins University; last edited 25 Sep 2023. Available from: https://omim.org/entry/614650
15. OMIM® Online Mendelian Inheritance in Man. #616276 COENZYME Q10 DEFICIENCY, PRIMARY, 7; COQ10D7 [Internet]. Baltimore (MD): Johns Hopkins University; last edited 23 Jan 2024. Available from: https://omim.org/entry/616276
16. OMIM® Online Mendelian Inheritance in Man. #616733 COENZYME Q10 DEFICIENCY, PRIMARY, 8; COQ10D8 [Internet]. Baltimore (MD): Johns Hopkins University; last edited 30 Jan 2024. Available from: https://omim.org/entry/616733
17. OMIM® Online Mendelian Inheritance in Man. #619028 COENZYME Q10 DEFICIENCY, PRIMARY, 9; COQ10D9 [Internet]. Baltimore (MD): Johns Hopkins University; last edited 28 Sep 2020. Available from: https://omim.org/entry/619028
18. OMIM® Online Mendelian Inheritance in Man. #615573 NEPHROTIC SYNDROME, TYPE 9; NPHS9 [Internet]. Baltimore (MD): Johns Hopkins University; last edited 7 Nov 2016. Available from: https://omim.org/entry/615573
19. OMIM® Online Mendelian Inheritance in Man. #620666 SPASTIC ATAXIA 10, AUTOSOMAL RECESSIVE; SPAX10 [Internet]. Baltimore (MD): Johns Hopkins University; last edited 7 May 2024. Available from: https://omim.org/entry/620666
20. OMIM® Online Mendelian Inheritance in Man. #620402 NEURONOPATHY, DISTAL HEREDITARY MOTOR, AUTOSOMAL RECESSIVE 9; HMNR9 [Internet]. Baltimore (MD): Johns Hopkins University; last edited 30 Jan 2024. Available from: https://omim.org/entry/620402
21. OMIM® Online Mendelian Inheritance in Man. #619026 NEURODEVELOPMENTAL DISORDER WITH PROGRESSIVE SPASTICITY AND BRAIN WHITE MATTER ABNORMALITIES; NEDSWMA [Internet]. Baltimore (MD): Johns Hopkins University; last edited 5 Feb 2026. Available from: https://omim.org/entry/619026
22. OMIM® Online Mendelian Inheritance in Man. #619027 SPASTIC PARAPLEGIA 83, AUTOSOMAL RECESSIVE; SPG83 [Internet]. Baltimore (MD): Johns Hopkins University; last edited 22 Nov 2021. Available from: https://omim.org/entry/619027
23. Orphanet. Coenzyme Q10 deficiency (ORPHA:35656) [Internet]. Paris: INSERM; 2026. Available from: https://www.orpha.net/en/disease/detail/35656
24. Orphanet. ORPHA:641353 Infantile neurodegeneration-progressive spasticity-intellectual disability-white matter lesions syndrome [Internet]. Paris: INSERM; epidemiology product of 23 Jun 2026 [cited 2026 Nov 1]. Available from: https://www.orpha.net/en/disease/detail/641353
25. Orphanet. ORPHA:631076 Autosomal recessive spastic paraplegia type 83 [Internet]. Paris: INSERM; epidemiology product of 23 Jun 2026 [cited 2026 Nov 1]. Available from: https://www.orpha.net/en/disease/detail/631076
26. Rahman S, Thorburn DR, Ball M. Nuclear gene-encoded Leigh syndrome spectrum overview. 2015 Oct 1 [updated 2025 May 1]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK320989/
27. MedlinePlus Genetics. Primary coenzyme Q10 deficiency [Internet]. Bethesda (MD): National Library of Medicine; 2017 [updated 2017 Apr 1; cited 2026 Oct 31]. Available from: https://medlineplus.gov/genetics/condition/primary-coenzyme-q10-deficiency/
28. Maia Filho JM, Medeiros LO, Ribeiro EM. Deficiência de coenzima Q10 (CoQ10). In: Ribeiro EM, coord. Síndromes Genéticas: Reconhecer para Tratar. Fortaleza: EdUnichristus; 2021. p. 127–129.
29. European Medicines Agency. EU/3/16/1765: orphan designation for the treatment of primary coenzyme Q10 deficiency syndrome (ubiquinol) [Internet]. Amsterdam: EMA; 2016 [updated 2022]. Available from: https://www.ema.europa.eu/en/medicines/human/orphan-designations/eu-3-16-1765
30. Hughes BG, Harrison PM, Hekimi S. Estimating the occurrence of primary ubiquinone deficiency by analysis of large-scale sequencing data. Sci Rep. 2017;7(1):17744. doi:10.1038/s41598-017-17564-y
31. Drovandi S, Lipska-Ziętkiewicz BS, Ozaltin F, Emma F, Gulhan B, Boyer O, et al. Variation of the clinical spectrum and genotype-phenotype associations in Coenzyme Q10 deficiency associated glomerulopathy. Kidney Int. 2022;102(3):592-603. doi:10.1016/j.kint.2022.02.040
32. Alecu JE, Tam A, Richter S, Quiroz V, Schierbaum L, Saffari A, et al. Quantitative natural history modeling of HPDL-related disease based on cross-sectional data reveals genotype-phenotype correlations. Genet Med. 2025;27(3):101349. doi:10.1016/j.gim.2024.101349
33. Lee EH, Kim-Mcmanus O, Yang JH, Haas R, Zaki MS, Abdel-Salam GMH, et al. HPDL variant type correlates with clinical disease onset and severity. Ann Clin Transl Neurol. 2025;12(7):1360-1367. doi:10.1002/acn3.70047. Authorship correction: Ann Clin Transl Neurol. 2026;13(1):212. doi:10.1002/acn3.70084
34. Traschütz A, Schirinzi T, Laugwitz L, Murray NH, Bingman CA, Reich S, et al. Clinico-Genetic, Imaging and Molecular Delineation of COQ8A-Ataxia: A Multicenter Study of 59 Patients. Ann Neurol. 2020;88(2):251-263. doi:10.1002/ana.25751
35. Rahman S, Clarke CF, Hirano M. 176th ENMC International Workshop: diagnosis and treatment of coenzyme Q10 deficiency. Neuromuscul Disord. 2012;22(1):76-86. doi:10.1016/j.nmd.2011.05.001
36. Trautmann A, Vivarelli M, Samuel S, Gipson D, Sinha A, Schaefer F, et al. IPNA clinical practice recommendations for the diagnosis and management of children with steroid-resistant nephrotic syndrome. Pediatr Nephrol. 2020;35(8):1529-1561. doi:10.1007/s00467-020-04519-1
37. Drovandi S, Lipska-Ziętkiewicz BS, Ozaltin F, Emma F, Gulhan B, Boyer O, et al. Oral Coenzyme Q10 supplementation leads to better preservation of kidney function in steroid-resistant nephrotic syndrome due to primary Coenzyme Q10 deficiency. Kidney Int. 2022;102(3):604-612. doi:10.1016/j.kint.2022.04.029
38. Distelmaier F, Corral-Sarasa J, Jiménez-Sánchez L, Díaz-Casado ME, Rohmann M, Seibt A, et al. Preclinical and first-in-human evidence of 4-hydroxybenzoic acid for mitochondrial COQ2 deficiency. Brain. 2026;149(5):1784-1797. doi:10.1093/brain/awaf334
39. European Reference Networks and patient associations: MetabERN (https://metab.ern-net.eu/), ERKNet (https://www.erknet.org/), ERN-RND (https://www.ern-rnd.eu/), ERN EURO-NMD (https://ern-euro-nmd.eu/), UMDF (https://www.umdf.org/), The Lily Foundation (https://www.thelilyfoundation.org.uk/), Children's Mitochondrial Disease Network (https://www.cmdn.org.uk/), International Mito Patients (https://www.mitopatients.org/) [Internet]. [cited 2026 Oct 31].
40. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 490–495 (Mitochondrial DNA diseases).
41. Nyhan WL, Hoffmann GF, Al-Aqeel AI, Barshop BA. Introduction to lactic acidemias. In: Atlas of Inherited Metabolic Diseases. 4th ed. Boca Raton: CRC Press; 2020. p. 353–363.
42. Wang Y, Hekimi S. The efficacy of coenzyme Q10 treatment in alleviating the symptoms of primary coenzyme Q10 deficiency: A systematic review. J Cell Mol Med. 2022;26(17):4635-4644. doi:10.1111/jcmm.17488
43. Ogasahara S, Engel AG, Frens D, Mack D. Muscle coenzyme Q deficiency in familial mitochondrial encephalomyopathy. Proc Natl Acad Sci U S A. 1989;86(7):2379-2382. doi:10.1073/pnas.86.7.2379
44. Rötig A, Appelkvist EL, Geromel V, Chretien D, Kadhom N, Edery P, et al. Quinone-responsive multiple respiratory-chain dysfunction due to widespread coenzyme Q10 deficiency. Lancet. 2000;356(9227):391-395. doi:10.1016/s0140-6736(00)02531-9
45. Quinzii C, Naini A, Salviati L, Trevisson E, Navas P, Dimauro S, et al. A mutation in para-hydroxybenzoate-polyprenyl transferase (COQ2) causes primary coenzyme Q10 deficiency. Am J Hum Genet. 2006;78(2):345-349. doi:10.1086/500092
46. López LC, Schuelke M, Quinzii CM, Kanki T, Rodenburg RJ, Naini A, et al. Leigh syndrome with nephropathy and CoQ10 deficiency due to decaprenyl diphosphate synthase subunit 2 (PDSS2) mutations. Am J Hum Genet. 2006;79(6):1125-1129. doi:10.1086/510023
47. Mollet J, Giurgea I, Schlemmer D, Dallner G, Chretien D, Delahodde A, et al. Prenyldiphosphate synthase, subunit 1 (PDSS1) and OH-benzoate polyprenyltransferase (COQ2) mutations in ubiquinone deficiency and oxidative phosphorylation disorders. J Clin Invest. 2007;117(3):765-772. doi:10.1172/jci29089
48. Lagier-Tourenne C, Tazir M, López LC, Quinzii CM, Assoum M, Drouot N, et al. ADCK3, an ancestral kinase, is mutated in a form of recessive ataxia associated with coenzyme Q10 deficiency. Am J Hum Genet. 2008;82(3):661-672. doi:10.1016/j.ajhg.2007.12.024
49. Duncan AJ, Bitner-Glindzicz M, Meunier B, Costello H, Hargreaves IP, López LC, et al. A nonsense mutation in COQ9 causes autosomal-recessive neonatal-onset primary coenzyme Q10 deficiency: a potentially treatable form of mitochondrial disease. Am J Hum Genet. 2009;84(5):558-566. doi:10.1016/j.ajhg.2009.03.018
50. Heeringa SF, Chernin G, Chaki M, Zhou W, Sloan AJ, Ji Z, et al. COQ6 mutations in human patients produce nephrotic syndrome with sensorineural deafness. J Clin Invest. 2011;121(5):2013-2024. doi:10.1172/jci45693
51. Ashraf S, Gee HY, Woerner S, Xie LX, Vega-Warner V, Lovric S, et al. ADCK4 mutations promote steroid-resistant nephrotic syndrome through CoQ10 biosynthesis disruption. J Clin Invest. 2013;123(12):5179-5189. doi:10.1172/jci69000
52. Brea-Calvo G, Haack TB, Karall D, Ohtake A, Invernizzi F, Carrozzo R, et al. COQ4 mutations cause a broad spectrum of mitochondrial disorders associated with CoQ10 deficiency. Am J Hum Genet. 2015;96(2):309-317. doi:10.1016/j.ajhg.2014.12.023
53. Malicdan MCV, Vilboux T, Ben-Zeev B, Guo J, Eliyahu A, Pode-Shakked B, et al. A novel inborn error of the coenzyme Q10 biosynthesis pathway: cerebellar ataxia and static encephalomyopathy due to COQ5 C-methyltransferase deficiency. Hum Mutat. 2018;39(1):69-79. doi:10.1002/humu.23345
54. Jacquier A, Theuriet J, Fontaine F, Mosbach V, Lacoste N, Ribault S, et al. Homozygous COQ7 mutation: a new cause of potentially treatable distal hereditary motor neuropathy. Brain. 2023;146(8):3470-3483. doi:10.1093/brain/awac453
55. Saudubray JM, García-Cazorla Á. Clinical approach to inborn errors of metabolism in paediatrics. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 1.
56. Morgan NV, Yngvadottir B, O'Driscoll M, Clark GR, Walsh D, Martin E, et al. Evidence that autosomal recessive spastic cerebral palsy-1 (CPSQ1) is caused by a missense variant in HPDL. Brain Commun. 2021;3(1):fcab002. doi:10.1093/braincomms/fcab002
57. Husain RA, Grimmel M, Wagner M, Hennings JC, Marx C, Feichtinger RG, et al. Bi-allelic HPDL variants cause a neurodegenerative disease ranging from neonatal encephalopathy to adolescent-onset spastic paraplegia. Am J Hum Genet. 2020;107(2):364-373. doi:10.1016/j.ajhg.2020.06.015
58. Wiessner M, Maroofian R, Ni MY, Pedroni A, Müller JS, Stucka R, et al. Biallelic variants in HPDL cause pure and complicated hereditary spastic paraplegia. Brain. 2021;144(5):1422-1434. doi:10.1093/brain/awab041
59. Ghosh SG, Lee S, Fabunan R, Chai G, Zaki MS, Abdel-Salam G, et al. Biallelic variants in HPDL, encoding 4-hydroxyphenylpyruvate dioxygenase-like protein, lead to an infantile neurodegenerative condition. Genet Med. 2021;23(3):524-533. doi:10.1038/s41436-020-01010-y
60. Nyhan WL, Hoffmann GF, Al-Aqeel AI, Barshop BA. Mevalonic aciduria. In: Atlas of Inherited Metabolic Diseases. 4th ed. Boca Raton: CRC Press; 2020. p. 636–642.
61. Freeman PJ, Hart RK, Gretton LJ, Brookes AJ, Dalgleish R. VariantValidator: accurate validation, mapping, and formatting of sequence variation descriptions. Hum Mutat. 2018;39(1):61-68. doi:10.1002/humu.23348. Tool: https://rest.variantvalidator.org/
62. Desbats MA, Morbidoni V, Silic-Benussi M, Doimo M, Ciminale V, Cassina M, et al. The COQ2 genotype predicts the severity of coenzyme Q10 deficiency. Hum Mol Genet. 2016;25(19):4256-4265. doi:10.1093/hmg/ddw257
63. National Center for Biotechnology Information. ClinVar: COQ2 variants VCV000214217, VCV000060538, VCV000060537, VCV000060536 [Internet]. Bethesda (MD): NCBI; 2026 [cited 2026 Oct 31]. Available from: https://www.ncbi.nlm.nih.gov/clinvar/variation/214217/
64. Multiple-System Atrophy Research Collaboration. Mutations in COQ2 in familial and sporadic multiple-system atrophy. N Engl J Med. 2013;369(3):233-244. doi:10.1056/nejmoa1212115
65. Yu MH, Tsang MH, Lai S, Ho MS, Tse DML, Willis B, et al. Primary coenzyme Q10 deficiency-7: expanded phenotypic spectrum and a founder mutation in southern Chinese. NPJ Genom Med. 2019;4:18. doi:10.1038/s41525-019-0091-x
66. Montouro L, Dias A, Morita M, Coelho É, Kouyoumdjian J, Graca C, et al. The weight of genotype on the clinical presentation of COQ7-related hereditary motor axonal neuropathy: a case series and literature review. Arq Neuropsiquiatr. 2026;84(4):1-12. doi:10.1055/s-0046-1818609
67. Theuriet J, Ribault S, Fontaine F, Gueguen N, Jacquier A, Quadrio I, et al. COQ7-Related Neuropathy: Two New Cases and Review of the Literature. J Peripher Nerv Syst. 2026;31(2):e70132. doi:10.1111/jns.70132
68. Feng Z, Guo L, Gao Y, Zhang M, Liu X, Kang B, et al. Prenatal Diagnosis of COQ2 Variants in Suspected Coenzyme Q10 Deficiency. Kidney Int Rep. 2025;10(11):3945-3957. doi:10.1016/j.ekir.2025.08.023
69. OMIM® Online Mendelian Inheritance in Man. *618994 4-HYDROXYPHENYLPYRUVATE DIOXYGENASE-LIKE; HPDL [Internet]. Baltimore (MD): Johns Hopkins University; [cited 2026 Oct 31]. Available from: https://omim.org/entry/618994
70. Sun Y, Wei X, Fang F, Shen Y, Wei H, Li J, et al. HPDL deficiency causes a neuromuscular disease by impairing the mitochondrial respiration. J Genet Genomics. 2021;48(8):727-736. doi:10.1016/j.jgg.2021.01.009
71. Micule I, Lace B, Wright NT, Chrestian N, Strautmanis J, Diriks M, et al. Case report: two families with HPDL related neurodegeneration. Front Genet. 2022;13:780764. doi:10.3389/fgene.2022.780764
72. Baggiani M, Desbats MA, Naef V, Giacich M, Galatolo D, Mero S, et al. Loss of function variants in HPDL impair human cortical development via alterations of mitochondrial function. Cell Death Dis. 2026;17(1):237. doi:10.1038/s41419-026-08476-9
73. Vaghefi F, Khosravi T, Motallebi F, Zarghami S, Al Sudani ZM, Rahimzadeh A, et al. New insights into HPDL protein: identification of a novel bi-allelic variant, docking simulation study, and literature review. BMC Med Genomics. 2026;19(1):109. doi:10.1186/s12920-026-02389-4
74. Numata-Uematsu Y, Uematsu M, Yamamoto T, Saitsu H, Katata Y, Oikawa Y, et al. Leigh syndrome-like MRI changes in a patient with biallelic HPDL variants treated with ketogenic diet. Mol Genet Metab Rep. 2021;29:100800. doi:10.1016/j.ymgmr.2021.100800
75. Kojima F, Okamoto Y, Ando M, Higuchi Y, Hobara T, Yuan J, et al. A novel homozygous HPDL variant in Japanese siblings with autosomal recessive hereditary spastic paraplegia: case report and literature review. Neurogenetics. 2024;25(2):149-156. doi:10.1007/s10048-024-00746-y
76. Mochel F, Sedel F. Inborn errors of metabolism in adults: a diagnostic approach to neurological and psychiatric presentations. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 2.
77. Iványi B, Rácz GZ, Gál P, Brinyiczki K, Bódi I, Kalmár T, et al. Diffuse mesangial sclerosis in a PDSS2 mutation-induced coenzyme Q10 deficiency. Pediatr Nephrol. 2018;33(3):439-446. doi:10.1007/s00467-017-3814-1
78. Hou JY, Chou IJ, Kuo CY, Lin KL. Expanding the epilepsy spectrum in primary coenzyme Q10 deficiency with COQ4 variants - Case series and literature review. Seizure. 2026;143:31-38. doi:10.1016/j.seizure.2026.09.020
79. Tsalouchidou PE, Juenemann C, Hahn W, Zahnert F, Möller L, Hakel L, et al. Adult-onset status epilepticus in patients with COQ8A coenzyme Q10 deficiency: A case series. Epilepsy Behav Rep. 2024;28:100716. doi:10.1016/j.ebr.2024.100716
80. Fiscal-Carvajal AB, López-Uriarte GA, Martínez-Garza LE, Femat-Roldán G. Hereditary spastic paraplegia type 83 with homozygous variant in HPDL, report of 3 cases in a Mexican family. Mov Disord Clin Pract. 2026;13(10):2629-2631. doi:10.1002/mdc3.70645
81. Sartorelli J, Longo D, Travaglini L, Orlando V, D'Amico A, Bertini E, et al. Acute ophthalmoplegia with Wernicke-like MRI pattern in a patient with HPDL-related disorder. Mov Disord Clin Pract. 2024;11(9):1160-1162. doi:10.1002/mdc3.14153
82. Kasuya Y, Mitani T, Noguchi K, Mizuno T, Makino S, Osaka H. Exotropia in a patient with a novel homozygous 4-hydroxyphenylpyruvate dioxygenase-like protein (HPDL) variant. Cureus. 2025;17(4):e83070. doi:10.7759/cureus.83070
83. Della Marina A, Bertolini A, Wegener-Panzer A, Flotats-Bastardas M, Reinhardt T, El Naggar I, et al. Mitochondrial diseases mimicking autoimmune diseases of the CNS and good response to steroids initially. Eur J Paediatr Neurol. 2022;41:27-35. doi:10.1016/j.ejpn.2022.09.003
84. Mero S, Satolli S, Galatolo D, Canto FD, Armando M, Astrea G, et al. HPDL biallelic variants in cerebral palsy and childhood-onset hereditary spastic paraplegia: human and zebrafish insights. Mov Disord. 2025;40(10):2085-2101. doi:10.1002/mds.30296
85. 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. p. 179–190. doi:10.1007/978-3-030-45457-9_11
86. 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
87. Ferreira CR, Darling A, Vockley J. Disorders of nucleic acid metabolism, tRNA metabolism and ribosomal biogenesis. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 39. doi:10.1007/978-3-662-63123-2_39
88. Spranger JW, Brill PW, Hall C, Nishimura G, Superti-Furga A, Unger S. Bone Dysplasias: An Atlas of Genetic Disorders of Skeletal Development. 4th ed. Oxford University Press; 2018. p. 218–219 (Schimke immunoosseous dysplasia).
89. Nyhan WL, Hoffmann GF, Al-Aqeel AI, Barshop BA. Multiple acyl CoA dehydrogenase deficiency/glutaric aciduria type II/ethylmalonic-adipic aciduria. In: Atlas of Inherited Metabolic Diseases. 4th ed. Boca Raton: CRC Press; 2020. p. 331–340.
90. Morris AAM, Spiekerkoetter U. Disorders of mitochondrial fatty acid oxidation and riboflavin metabolism. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 12.
91. Perlman S. Hereditary ataxia overview. 1998 Oct 28 [updated 2022 Jun 16; revised 2026 Jul 9]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1138/
92. Parikh S, Saneto R, Falk MJ, Anselm I, Cohen BH, Haas R, et al. A modern approach to the treatment of mitochondrial disease. Curr Treat Options Neurol. 2009;11(6):414-430. doi:10.1007/s11940-009-0046-0
93. Colona VL, Travaglini L, Sartorelli J, Vasco G, Piluso A, D'Amico A, et al. Genotype variability in early-onset hereditary spastic paraplegia: a single-center study. Eur J Paediatr Neurol. 2025;57:82-90. doi:10.1016/j.ejpn.2025.06.001
94. Stolwijk NN, Penninx BMF, Bosch AM, Langeveld M, Salomons GS, Hollak CEM. Essential Oral Single Nutritional Therapy Products for Inherited Metabolic Diseases: Evidence and Consensus Assessment Using a Modified Delphi Method. J Inherit Metab Dis. 2026;49(6):e70252. doi:10.1002/jimd.70252
95. Morris AAM, Jones S. Medications used in the treatment of inborn errors of metabolism. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 45.
96. INFARMED. Q10 10 mg cápsula (ubidecarenona). Resumo das Características do Medicamento [Internet]. Lisboa: INFARMED; renewal 2024. Available from: https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=7293&tipo_doc=rcm
97. Nyhan WL, Hoffmann GF, Al-Aqeel AI, Barshop BA. Deficiency of the pyruvate dehydrogenase complex. In: Atlas of Inherited Metabolic Diseases. 4th ed. Boca Raton: CRC Press; 2020. p. 376–384.
98. Yue Y, Zhao F, Chen Q. COQ2-Associated Primary Coenzyme Q10 Deficiency Presenting With Proteinuria: A Case Report and Literature Review. Mol Genet Genomic Med. 2026;14(7):e70221. doi:10.1002/mgg3.70221
99. Mantle D, Hargreaves IP. Efficacy and Safety of Coenzyme Q10 Supplementation in Neonates, Infants and Children: An Overview. Antioxidants (Basel). 2024;13(5):530. doi:10.3390/antiox13050530
100. ClinicalTrials.gov. Trial search: coenzyme Q10 deficiency, COQ genes, 4-hydroxybenzoic acid [Internet]. Bethesda (MD): National Library of Medicine; 2026 [cited 2026 Oct 31]. Available from: https://clinicaltrials.gov/
101. ESCAPE Trial Group; Wühl E, Trivelli A, Picca S, Litwin M, Peco-Antic A, Zurowska A, et al. Strict blood-pressure control and progression of renal failure in children. N Engl J Med. 2009;361(17):1639-1650. doi:10.1056/NEJMoa0902066
102. Hahn W, Erffmeier K, Schulze M, Zahnert F, Knake S, Tsalouchidou PE. Intermittent ketogenic fasting with medium-chain triglycerides improves ataxia in COQ8A-related coenzyme Q10 deficiency: A case report. Mol Genet Metab Rep. 2026;46:101296. doi:10.1016/j.ymgmr.2026.101296
103. U.S. Food and Drug Administration. Drugs@FDA and openFDA drug label API: search «ubidecarenone» [Internet]. Silver Spring (MD): FDA; 2026 [cited 2026 Oct 31]. Available from: https://api.fda.gov/drug/label.json?search=openfda.generic_name:%22ubidecarenone%22
104. U.S. Food and Drug Administration. Search Orphan Drug Designations and Approvals: «treatment of primary CoQ10 deficiency» (coenzyme Q10 [CoQ10], BPGbio, 2026-01-13; 4-hydroxybenzoate, Gleeson Lab - UCSD, 2026-01-16; 4-hydroxybenzoic acid, University of Granada, 2026-01-20; 4-hydroxybenzoic acid, New York University, 2026-02-15) [Internet]. Silver Spring (MD): FDA; 2026 [cited 2026 Oct 31]. Available from: https://www.accessdata.fda.gov/scripts/opdlisting/oopd/index.cfm
105. European Medicines Agency. EU/3/25/3052: orphan designation for the treatment of primary ubiquinone deficiency (parahydroxybenzoic acid) [Internet]. Amsterdam: EMA; 2025 [cited 2026 Oct 31]. Available from: https://www.ema.europa.eu/en/medicines/human/orphan-designations/eu-3-25-3052
106. INFARMED. Ubenzima 50 mg cápsula (ubidecarenona). Resumo das Características do Medicamento [Internet]. Lisboa: INFARMED; renewal 2024. Available from: https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=8799&tipo_doc=rcm
107. Agência Nacional de Vigilância Sanitária. Open data: medicines registered in Brazil [Internet]. Brasília: ANVISA; 2026 [cited 2026 Oct 31]. Available from: https://dados.anvisa.gov.br/dados/DADOS_ABERTOS_MEDICAMENTOS.csv
108. BPGbio. BPGbio receives FDA Rare Pediatric Disease designation for its potential treatment for primary coenzyme Q10 deficiency [press release]. 2024 Sep 30. Available from: https://www.biospace.com/press-releases/bpgbio-receives-fda-rare-pediatric-disease-designation-for-its-potential-treatment-for-primary-coenzyme-q10-deficiency
109. BPGbio. BPGbio receives FDA Orphan Drug Designation for BPM31510 for the treatment of primary CoQ10 deficiency [press release]. Boston; 2026 Jan 21. Available from: https://www.santelog.com/actualites-sante-nasdaq/update-bpgbio-receives-fda-orphan-drug-designation-bpm31510-treatment
110. Corral-Sarasa J, Martínez-Gálvez JM, González-García P, Wendling O, Jiménez-Sánchez L, López-Herrador S, et al. 4-Hydroxybenzoic acid rescues multisystemic disease and perinatal lethality in a mouse model of mitochondrial disease. Cell Rep. 2024;43(5):114148. doi:10.1016/j.celrep.2024.114148
111. Münch J, Strahleck T, Seibt A, Smits SHJ, Herbrich S, Hohnecker AK, et al. Successful treatment of neonatal COQ2 deficiency with 4-hydroxybenzoic acid. Brain. 2026;149(5):e41-e44. doi:10.1093/brain/awag023
112. González-García P, Jiménez-Sánchez L, Corral-Sarasa J, López-Herrador S, Torres-Rusillo S, Díaz-Casado ME, et al. The treatment of primary CoQ deficiency requires the targeting of multiple pathogenic mechanisms. Commun Med (Lond). 2025;5(1):286. doi:10.1038/s43856-025-01000-8
113. Barriocanal-Casado E, Stopka SA, Pesini A, Aristizabal-Henao JJ, Karmacharya S, Van Cura D, et al. Spatial mapping of CoQ10 repletion by BPM31510 in a genetic mouse model (Coq4 F147C) of coenzyme Q deficiency. J Lipid Res. 2026;67(3):100987. doi:10.1016/j.jlr.2026.100987
114. ClinicalTrials.gov. NCT05848271: Natural History Study of Patients with HPDL Mutations (University of California, San Diego; observational; recruiting) [Internet]. Bethesda (MD): National Library of Medicine; last updated 30 Mar 2025 [cited 2026 Nov 1]. Available from: https://clinicaltrials.gov/study/NCT05848271
115. Alcázar-Fabra M, Rodríguez-Sánchez F, Trevisson E, Brea-Calvo G. Primary Coenzyme Q deficiencies: A literature review and online platform of clinical features to uncover genotype-phenotype correlations. Free Radic Biol Med. 2021;167:141-180. doi:10.1016/j.freeradbiomed.2021.02.046