Gyrate Atrophy of the Choroid and Retina (Hyperornithinaemia)

Ornithine-δ-aminotransferase (OAT) deficiency: a classic form with hyperornithinaemia and progressive chorioretinal degeneration, and a neonatal form with hyperammonaemia and hypo-ornithinaemia.

Literature and regulatory information updated: 28 October 2026.

0. Initial Reference Table

1. Introduction

1.1. Historical overview

Gyrate atrophy of the choroid and retina was first described as an ophthalmological entity, defined by the appearance of the ocular fundus: circular patches of chorioretinal atrophy, with sharp, scalloped borders, that appear in the mid-periphery and coalesce over the years. The link to an inborn error of metabolism was established in 1973 by Simell and Takki, who documented raised plasma ornithine in patients with this form of atrophy, the first published association between hyperornithinaemia and gyrate atrophy [18, 19]. In that series of nine patients, whose parents all came from the same region of Finland, ornithine was 10 to 20 times above normal in plasma, urine, cerebrospinal fluid and aqueous humour, there was no other consistent clinical abnormality apart from the ocular one, and ammonia remained normal when fasting, after meals and in loading tests [3]. Finland, where the disease is most frequent, provided the first clinical series and, later, the largest cohorts with neurological and muscle studies [4, 20, 21].

Before the biochemical discovery, François and colleagues (1960) had proposed that the Alder leucocyte anomaly, found in a patient with gyrate atrophy, in her consanguineous parents and in relatives across four generations, was the heterozygous expression of the disease gene; in 1966 the same group did not find the anomaly in nine patients and abandoned the hypothesis. OMIM concludes from this history that there is no true gyrate atrophy independent of hyperornithinaemia and that presumed cases without it are phenocopies [3].

The enzyme deficiency was demonstrated in 1977 by Valle, Kaiser-Kupfer and Del Valle, who found no detectable ornithine-δ-aminotransferase activity in transformed lymphocytes from a patient, whereas her heterozygous daughter had 44% of normal activity [8]. The following year, O'Donnell, Sandman and Martin showed that fibroblasts from a patient did not convert carbon-14-labelled ornithine into proline, owing to L-ornithine:2-oxoacid aminotransferase deficiency, with intermediate activity in the heterozygous father [22]. In 1980 the group at the United States National Eye Institute (NEI) described improved visual function after lowering plasma ornithine with diet, the starting point of a natural history and dietary intervention protocol that ran from 1978 to 2004 and included 65 patients [23, 24].

1.2. Milestones of molecular discovery

In 1986 Inana and colleagues cloned the human OAT complementary DNA (cDNA) from a retinoblastoma line, with an open reading frame of 1371 nucleotides encoding a precursor of 439 amino acids (48,534 Da), of which the first 32 form the mitochondrial targeting peptide; the mature protein has 407 amino acids [25]. In 1988 Mitchell and colleagues defined the exon–intron structure of the gene and described the first pathogenic variant, a change in the initiation codon (ATG to ATA) in four Lebanese patients [26]. In the same year Ramesh and colleagues compared a pyridoxine-responsive and a non-responsive patient, with different missense variants, and showed that the responder's enzyme increased its activity at higher concentrations of pyridoxal 5'-phosphate (PLP), the molecular basis of the vitamin B6-responsive form [27]. Genetic linkage to the locus was established in parallel: somatic cell hybrid analysis assigned the OAT locus to human chromosome 10 and to mouse chromosome 7 in 1985, subsequent studies narrowed the region to 10q23–qter, and family linkage data from 1988 confirmed the position on the long arm of chromosome 10 [7]. Barrett and colleagues (1987) located DNA sequences homologous to OAT at 10q26 and at Xp11.2; those on the X chromosome later proved to be pseudogenes, not a second functional locus (section 2.1) [3, 7]. A 1988 restriction fragment length polymorphism (RFLP) study showed that the chromosome 10 OAT locus segregated with the disease in one family and was in linkage disequilibrium with it, genetic proof that the defect lies in the structural gene at 10q26 [3]. In 1992 Brody and colleagues described 21 new alleles and found that 20 of 24 alleles had normal messenger RNA but only 2 of 24 had a normal amount of immunoreactive protein, which pointed to protein instability as the predominant mechanism [28]. In the same year, single-strand conformation polymorphism (SSCP) screening of 24 still uncharacterised alleles identified 13 different variants in 19 of them (79%) [7]. In 1995 the targeted Oat knockout mouse reproduced neonatal hypo-ornithinaemia and late retinal degeneration, and clarified the two faces of the disease [29].

1.3. Nomenclature and reclassifications

The disease is referred to in the literature as gyrate atrophy of the choroid and retina, as hyperornithinaemia with gyrate atrophy or as OAT deficiency; OMIM entry 258870 has the main title «gyrate atrophy of choroid and retina» (GACR) and, in the phenotype–gene relationship, names the phenotype with the phrase «with or without ornithinemia», which acknowledges the neonatal phase without hyperornithinaemia; the alternative titles record the historical names of the enzyme (ornithine-ketoacid aminotransferase, ornithine-δ-aminotransferase) and the acronym HOGA (hyperornithinemia with gyrate atrophy) [1–3]. The classic designation of the Finnish founder variant, L402P, uses the numbering of the 439-amino-acid precursor and coincides with Human Genome Variation Society (HGVS) nomenclature on the reference transcript: NM_000274.4:c.1205T>C, p.(Leu402Pro) [4, 6, 30]. Nyhan's Atlas also records a third historical hyperornithinaemia, described by Bickel in 1968 in two siblings with intellectual disability and renal tubular dysfunction, with hepatic OAT activity reduced by 60–80% but kinetically normal, which the authors accept may correspond to a partial deficiency of the same enzyme [31].

1.4. Definition of the entities in scope

The classic form is defined by the presence of biallelic pathogenic variants in OAT and/or by the combination of persistent hyperornithinaemia with the typical chorioretinal lesions described by an ophthalmologist; these are the criteria used in the international registry coordinated in Amsterdam [12]. The neonatal form is the same disease in the phase in which OAT flux favours ornithine synthesis: it presents in the first weeks, mainly in preterm infants, with poor feeding progression, growth failure, symptomatic hyperammonaemia and orotic aciduria, with low plasma ornithine, citrulline and arginine, and hyperornithinaemia appears only months later [4, 13]. The subdivision into pyridoxine-responsive and non-responsive forms is functional and depends on the variant; the proportion of responders varies between sources (see section 7.2) [4, 12, 32]. HHH syndrome, choroideremia and the gyrate atrophy-like choroidal dystrophy associated with C1QTNF5 are distinct entities, covered only in the differential diagnosis; secondary creatine deficiency and vitamin B6-dependent epilepsies are placed in comparison with cerebral creatine deficiencies and pyridoxine-dependent seizures.

2. Molecular Genetics and Pathophysiology

2.1. The OAT gene and protein

The OAT gene lies at 10q26.13 and is the only gene associated with gyrate atrophy; according to the ClinGen actionability report, all causal defects identified are located in this gene [2, 10]. The MANE Select reference transcript is NM_000274.4, which encodes the 439-amino-acid protein NP_000265.1, including the mitochondrial targeting peptide; shorter transcripts of unknown function have been described, as has a splice variant with retention of exon 2 observed only in a retinoblastoma line [5, 6]. The gene spans about 21 kb (GRCh38, 10:124 397 303–124 418 923) and has 11 exons; the 5' flanking region combines features of housekeeping genes (GC richness and three Sp1 transcription factor binding sites) and elements found in regulated and tissue-specific genes (a TATA box-like element and two CCAAT boxes) [7]. The human and rat proteins both have 439 amino acids, with 90.4% identity [7]. There are related non-functional sequences: a cluster of pseudogenes at Xp21.1–p11.1, which includes the OATL2 sequence at Xp11.2, and a partial non-processed pseudogene, OATL3, homologous to exon 3 and its flanking introns, located at 10q26 next to the structural gene [7]. The sources consulted do not describe interference from these pseudogenes in clinical sequencing of OAT. The human enzyme is a PLP-dependent mitochondrial matrix aminotransferase, organised as a homotetramer (dimer of dimers), with a «ping-pong» catalytic mechanism in which the cofactor alternates between PLP and pyridoxamine 5'-phosphate (PMP) [33]. Expression is highest in the liver, intestine and kidney; in the liver it is concentrated in perivenous hepatocytes, not in the periportal hepatocytes where the urea cycle operates; in the retina, OAT is expressed in the retinal pigment epithelium (RPE) and in photoreceptors [5, 33, 34]. Overexpression of OAT in some cancers has made the enzyme a target for inhibitors in oncology, a line of research unrelated to the treatment of gyrate atrophy [33].

2.2. Ornithine biochemistry and reversibility of the reaction

Ornithine is an intermediate of the urea cycle, of proline and glutamate synthesis, of creatine biosynthesis and of polyamine synthesis [4]. OAT transfers the δ-amino group of L-ornithine to α-ketoglutarate (2-oxoglutarate) and produces glutamate and glutamate-γ-semialdehyde, which cyclises spontaneously to Δ1-pyrroline-5-carboxylate (P5C), the precursor of proline and glutamate; it thus links arginine and ornithine metabolism to the pools of glutamate, proline and γ-aminobutyric acid (GABA) and is the main route of ornithine disposal outside the urea cycle [2, 33, 34]. Ornithine is formed from arginine by arginase 1 (cytoplasm) and arginase 2 (mitochondrion); cytosolic ornithine enters the mitochondrion through the ORNT1 antiporter (encoded by SLC25A15, the HHH syndrome gene), in exchange for mitochondrial citrulline [4, 34].

The reaction catalysed by OAT is freely reversible, and the direction of net flux changes with age. In the neonatal period, OAT has an anaplerotic role for the urea cycle: flux is directed towards the synthesis of ornithine and, through the cycle, of citrulline and arginine; initial expression of arginase 1 is low and ornithine behaves as an essential amino acid [4, 5]. A few months after birth, the flux reverses and comes to favour the disposal of arginine and ornithine by conversion to P5C [4]. This reversal explains the two faces of the disease. In the newborn, lack of OAT reduces the production of ornithine, citrulline and arginine, impairs ureagenesis and causes hyperammonaemia, especially when dietary arginine is below the needs of growth; the role of intramitochondrial arginase 2, expressed in enterocytes after weaning, has not been clarified [4]. After the flux reverses, lack of OAT prevents the breakdown of ornithine, which accumulates in plasma 5 to 20 times above normal [4]. Hyperornithinaemia does not increase the activity of the urea cycle, whose rate-limiting step is carbamoyl phosphate synthetase 1 [5].

When plasma ornithine reaches or exceeds 400 µmol/L, urinary excretion of ornithine, lysine, arginine and cystine increases, through competitive inhibition of the common renal tubular transporter of dibasic amino acids [4]. Excess ornithine also inhibits glycine amidinotransferase (AGAT), the first enzyme of creatine synthesis, which gives rise to secondary creatine deficiency in blood, urine, muscle and brain; subnormal serum creatinine reflects the reduction in total body creatine [4, 35]. Saudubray includes OAT deficiency among the causes of secondary disturbances of creatine metabolism, together with urea cycle defects, HHH syndrome and P5C synthetase (P5CS) deficiency [36].

2.3. Why the retina is the most affected organ

The mechanism of chorioretinal degeneration remains unclear, according to Saudubray and the ClinGen summary [4, 10]. The review by Buijs and colleagues (2025) groups the hypotheses into four categories, none demonstrated in humans [5]. The first attributes the damage to a deficit of downstream products: the RPE depends on proline catabolism as an energy source, through an RPE-specific proline transporter (SLC6A20), and interruption of the OAT→P5C→proline pathway would reduce the availability of proline and glutathione, with oxidative stress and ferroptosis. The second involves activation of the polyamine pathway by ornithine decarboxylase, with production of reactive oxygen species and acrolein; urinary excretion of polyamines, especially putrescine, is increased in patients. The third rests on the absence of ornithine transcarbamylase (OTC) in the RPE, which prevents local disposal of ornithine through the urea cycle, and on the cationic amino acid transporter CAT-1 (SLC7A1), whose silencing limited ornithine toxicity in vitro in OAT-deficient RPE cells. The fourth describes cumulative damage to the photoreceptor–RPE–choroid complex with irreversible loss of the RPE barrier, a mechanism also proposed for cystoid maculopathy [5].

The debate on the origin of the damage has therapeutic consequences. The Amsterdam group argues against toxicity due exclusively to circulating ornithine with two observations: most patients with HHH syndrome have very high plasma ornithine but no retinal disease, and studies by the same group on sub-RPE proteins indicate that about 70% come from the circulation and 30% from local retinal sources, from which they extrapolate that lowering systemic ornithine alone is not enough to prevent all the degeneration [34]. The Naples group responds that hepatic OAT deficiency is responsible for systemic hyperornithinaemia and that reducing ornithine prevents degeneration in the mouse and delays it in humans, which makes the liver the logical target [37]. Human histopathology is scarce: in the retina of a pyridoxine-responsive patient, studied post mortem, there were focal areas of photoreceptor atrophy with adjacent RPE hyperplasia, and electron microscopy showed abnormal mitochondria in the corneal endothelium, in the non-pigmented ciliary epithelium and, to a lesser extent, in the photoreceptors and iris smooth muscle [4, 5].

2.4. Animal models

The targeted Oat knockout mouse (Wang, 1995) has neonatal hypo-ornithinaemia that is lethal within about 48 hours, prevented by administration of arginine; survivors later develop hyperornithinaemia and progressive retinal degeneration, whose first changes appear in the RPE [4, 5, 29]. In the same 1995 study, a human infant with the deficiency also had transient hypo-ornithinaemia, which led the authors to conclude that the OAT reaction plays opposite roles in the neonatal period and in adulthood [3]. In this model, dietary arginine restriction completely prevented retinopathy at the age at which untreated animals developed it [4, 5]. In the study by Wang (2000), with assessment over 12 months by plasma amino acids, ERG, histology and ultrastructure, the diet markedly reduced plasma ornithine and prevented degeneration; the authors concluded that ornithine accumulation is a necessary condition for retinal damage and that, in the mouse, restoring OAT activity in the retina itself is not necessary for effective treatment, a finding relevant to the debate between ocular and hepatic therapy [3]. The Oat^rhg^ mouse derives from a spontaneous recessive mutation recognised by delayed hair growth (retarded hair growth); Bisaillon and colleagues (2014) showed that it is a transversion in exon 9 of Oat that replaces glycine 353 with alanine, a conserved residue and the same one affected by the human variant p.Gly353Asp [7]. rhg/rhg animals and rhg/null-allele compound heterozygotes appear normal at birth, grow more slowly, have raised plasma ornithine, low plasma lysine and reduced hepatic OAT activity, and develop chorioretinal degeneration similar to the human disease; intraperitoneal arginine until weaning improved weight gain in these two genotypes, but not in animals without any functional allele, which die within the first 48 hours [7]. The Oat^rhg^ mouse, homozygous for p.Gly353Ala, has at 12 months a well-defined superonasal patch of atrophy with RPE clumping, migration and hypertrophy, with the rest of the retina and the full-field ERG normal, whereas the null model loses central photoreceptors and has a reduced ERG from 4 months; the regional variation resembles that of patients, and multimodal imaging has been proposed as an outcome in therapeutic studies [38]. A 2026 proteomic and metabolomic study in the Oat^rhg^ mouse, before visual loss, found changes in methylation metabolites in all tissues, diversion of hepatic ornithine into the urea cycle, few proteomic changes in the retina and the most extensive changes in the RPE/choroid complex (mitochondrial metabolism, cytoskeleton, extracellular matrix, lysine metabolism and antioxidant capacity), with lysine degradation increased by ornithine in an OAT-dependent manner [39]. Cellular models include ARPE-19 cells with OAT silencing, without ornithine cytotoxicity over the duration of the assays, OAT-null HEK293 cells expressing human variants, and induced pluripotent stem cells [5].

2.5. Variant spectrum

Saudubray cites more than 70 pathogenic variants and a 2023 review more than 80, mainly missense, as well as nonsense, frameshift and splicing variants; at the time of the Buijs review, ClinVar listed 82 pathogenic and 44 likely pathogenic variants [4, 5, 40]. The historical nomenclature of OAT proteins uses the numbering of the 439-amino-acid precursor, which coincides with HGVS on NP_000265.1 for the validated variants (for example, V332M corresponds to p.Val332Met) [6]. The following table brings together variants cited in this article, normalised in VariantValidator without warnings and with the ClinVar classification consulted on 28 October 2026.

Other variants described in specific populations, in the authors' notation on NM_000274.4, include c.1058G>A p.(Gly353Asp), present in all patients of Dutch ancestry studied (14 of 16 Dutch and Belgian patients in the registry, in homozygous or compound heterozygous state), c.899C>G p.(Pro300Arg) homozygous in four Iraqi siblings, c.1173G>A p.(Trp391Ter) homozygous in patients of Italian-Romanian origin, c.772-1G>A in a Syrian patient and c.1A>G, with loss of the initiation codon, in a patient from Sri Lanka [12, 43]. Table 2 of Balfoort 2025 lists «c.897C>G p.(Tyr229Ter)»; in ClinVar, c.897C>G corresponds to p.(Tyr299Ter) (ID 176), which identifies a typographical error in the paper [12, 30]. In the Yunnan cohort, c.213G>A p.(Trp71Ter), c.799A>C p.(Thr267Pro), c.897C>A p.(Tyr299Ter) and c.967T>C p.(Tyr323His) were described in homozygous state in two Han patients, and the non-coding variant c.-30+22_-30+43del in all five patients from ethnic minorities (four homozygous), from three unrelated families; the haplotype was not studied [42]. In the Mexican cohort, c.796C>T p.(Gln266Ter) and c.721_722dupCC p.(Asp242ArgfsTer6) were described [44].

The OMIM catalogue of allelic variants for OAT brings together 40 selected entries, almost all published between 1988 and 1995 in patients of diverse origins, and adds data on ancestry, zygosity and mechanism that recent cohorts do not repeat [7]. The nucleotide positions in those descriptions are counted from the initiation codon and mostly coincide with current HGVS; the conversions in the following table were confirmed in VariantValidator on NM_000274.4 and the classification is that of ClinVar consulted on 28 October 2026 [6, 7, 30]. Four OMIM descriptions do not match the reference sequence: the variant designated R184T results from c.550G>A and alters alanine 184, being p.(Ala184Thr); Y245C is c.734A>G, not a change at nucleotide 743; the W391X from SSCP screening is c.1172G>A, different from the c.1173G>A described in patients of Italian-Romanian origin, although both produce p.(Trp391Ter); and the designation P199N corresponds to p.(Pro199Gln), as the title of the entry itself indicates [6, 7, 12, 30]. Variants already included in the previous table (p.(Leu402Pro), p.(Val332Met), p.(Arg180Thr), p.(Ala226Val), p.(Pro241Leu), p.(Gly353Asp), p.(Arg426Ter) and p.(Tyr299Ter)) are not repeated.

Four variants in the table with a functional or clinical description are now classified in ClinVar as of uncertain significance or conflicting (p.(Gln90Glu), p.(Asn54Lys), p.(Tyr245Cys) and p.(Arg154Leu)), which shows that description in a historical catalogue does not replace classification under current criteria [7, 30].

2.6. Molecular mechanisms of the variants and pyridoxine response

Study of purified recombinant enzymes made it possible to classify missense variants into three groups [33]. Catalytic variants affect the active site directly: p.Arg180Thr has a catalytic efficiency below one-thousandth of normal, a Michaelis constant for ornithine 40 times higher, and catalyses an abnormal α-transamination, and p.Gly237Asp has no measurable activity. Interface or oligomerisation variants (p.Gly51Asp, p.Gly121Asp, p.Arg154Leu, the substitution at residue 158 designated Y/S158F by the authors, p.Thr181Met with an efficiency of about 1/70, p.Pro199Gln with an efficiency of 1/500 or less) reduce thermal stability and shift the equilibrium from tetramer to dimer. Folding variants (p.Gln90Glu, p.Arg271Lys with about half the efficiency, p.Glu318Lys with near-normal efficiency, p.Cys394Tyr with about half and substrate inhibition, p.Gly353Asp with low solubility and aggregation) act mainly through the amount of functional protein [33]. The p.Val332Met variant has lower affinity for PMP, releases the cofactor during catalysis and is converted to apoenzyme; in a cellular model it is an inactive apodimer that partially becomes an active holotetramer when exogenous PLP is supplied, which offers an explanation for the pyridoxine response of carrier patients; the same defect was observed in p.Gly51Asp and p.Gly121Asp, but the clinical response of homozygotes for these variants is poorly known [33]. Saudubray classifies gyrate atrophy among the inborn errors of PLP-dependent enzymes in which pyridoxine acts as a pharmacological chaperone, with a response «in some patients», less frequent than in classic homocystinuria (about 50%) or in X-linked sideroblastic anaemia (about 90%) [45].

The p.Gly353Asp variant, frequent in the Netherlands, causes a folding and aggregation defect that PLP does not correct: no carrier responded to pyridoxine in vivo or in vitro [43]. In the assay of enzyme activity and labelled ornithine flux in fibroblasts from 14 pathogenic variants, activity and flux were below the limit of quantification in most, and only one patient line, apart from the positive control, responded to pyridoxine in vitro, in agreement with the in vivo response [43]. Pyridoxine response thus appears to be variant-specific [10, 43].

Cellular studies from the 1980s already pointed in this direction. Fusion of fibroblasts from four responsive and three non-responsive patients produced no complementation, which indicated that the two forms are allelic; fibroblasts from responders had more OAT activity in homogenates and greater incorporation of carbon-14-labelled ornithine into proteins (Wirtz, 1985) [3]. Kennaway and colleagues (1989), in nine patients, measured in fibroblast mitochondria the apparent affinity of the mutant enzyme for PLP and its thermal stability at 45 °C: the apparent Michaelis constant for PLP was lower in non-responders than in responders, the constant for ornithine was normal in the seven patients studied, and the patient with the mildest disease, responsive to pyridoxine, had the most stable enzyme but also the highest constant for PLP; OAT protein was detectable by Western blot in all responders and in two of five non-responders, and low or undetectable in the other three [3]. These data are compatible with the current interpretation that pyridoxine compensates for reduced affinity for the cofactor and stabilises the protein, and does not act on variants without residual protein.

The OMIM catalogue documents other mechanisms [7]. In variants that introduce a premature stop codon, the amount of messenger RNA depends on position: stop codons at positions 79, 209 and 299 reduced mRNA in fibroblasts, whereas p.(Arg426Ter), in the last exon, barely altered it, a pattern compatible with nonsense-mediated mRNA decay (NMD), which as a rule spares stop codons in the last exon [7]. Two nonsense variants, p.(Trp178Ter) in exon 6 and p.(Trp275Ter) in exon 8, caused skipping of the exon that contains them, an example of splicing alteration induced by a stop codon [7]. Other variants act on mitochondrial transport: the precursor with p.(His319Tyr) enters the mitochondrion but is barely processed and is degraded, and the precursor with p.(Gln90Glu), a substitution located in the mature sequence and not in the targeting peptide, remains in the cytosol and is degraded, which indicates that regions of the mature protein take part in import [7]. In splicing, besides the 9-bp deletion at the intron 4 acceptor site with loss of exon 5, a substitution in an intronic Alu element, oriented opposite to the direction of transcription, created a donor site and activated a cryptic acceptor, with insertion of 142 nucleotides into the mature mRNA, presented by the authors as the first example of Alu exonisation associated with a hereditary disease [7]. As for the cofactor, the p.(Val332Met) enzyme expressed in yeast increased its specific activity fourfold when assay PLP went from 50 to 600 µmol/L, and p.(Ala226Val) went from undetectable activity to about 10% of normal over the same range, in fibroblasts from a homozygote and in Chinese hamster ovary cells; a compound heterozygous p.(Ala226Val)/p.(Arg398Ter) patient lowered mean plasma ornithine with pyridoxine, with constant protein intake [7]. In contrast, p.(Asn54Lys) and p.(His319Tyr) were described in patients without a vitamin B6 response [7].

2.7. Heterozygotes, penetrance, expressivity and genotype–phenotype correlations

Obligate heterozygotes have intermediate OAT activity in fibroblasts and lymphocytes, for both responsive and non-responsive variants, and have no disease [4, 8, 22]. Heterozygotes for a responsive variant and for a non-responsive variant can be distinguished in vitro by stimulation of OAT activity with PLP: the ratio of activity without to activity with added PLP was lowest in controls, intermediate in heterozygotes for the non-responsive form and highest in those for the responsive form (Shih, 1988) [3]. A review of 80 published cases, cited by OMIM, found that transmission was always compatible with autosomal recessive inheritance, with 44 women (55%), and no quantitative differences in OAT activity were found between patients of the two sexes in fibroblasts and lymphocytes [3]. In homozygotes and compound heterozygotes, neural retinal dysfunction was present in all patients in the historical series, and hyperornithinaemia outside the neonatal period is the rule, with one published exception: a Mexican patient with a genetic diagnosis and plasma ornithine of 48 µmol/L, with no description in the abstract of the diet or of sampling conditions [10, 44]. Expressivity varies. The historical series collated by ClinGen (7 to 40 patients each) recorded posterior sutural cataract in 96–100% of lenses, RPE defects in 30–81%, optic disc drusen in 14%, macular involvement in 43%, peripapillary atrophy in 57% of adults, neural retinal dysfunction in 100%, non-disabling symptomatic neuropathy in 10%, type II muscle fibre atrophy in 100% of those biopsied, fatty changes of the lower limb musculature in 100% and focal epileptiform activity in 18% [10].

Sources differ on within-family variability. Kaiser-Kupfer (1991), in six sibling pairs from the NEI, and Saudubray describe much less intrafamilial than interfamilial variation, which suggests that allelic heterogeneity explains part of the variability [4, 46]. ClinGen, the Buijs review and the Finnish series by Peltola, with 35 patients carrying the same founder allele, describe wide phenotypic variation in visual acuity, fundus and fields, even with the same genotype, although the two eyes of each patient were always similar [5, 10, 20]. The genotype–phenotype correlation is, moreover, confounded by treatment and adherence: in the international registry, genotype did not modify the ornithine response to diet [5, 12]. The only consistent correlation is pyridoxine response in some variants, such as p.Val332Met, and lack of response in p.Gly353Asp [33, 43]. Comparison of Japanese and Finnish patients suggested worse visual function in the former, and women may retain larger visual fields than men, which indicates that natural history may vary between populations and between sexes [10]. The sources consulted, including the OMIM gene entry, do not describe allelic disorders of OAT or identified genetic modifiers [7].

3. Clinical Spectrum and Manifestations by System

3.1. Cohorts underpinning the clinical description

The natural history of gyrate atrophy rests on small series, almost all retrospective, that differ in the origin of the patients, the period and the treatment received. The following table summarises the cohorts used in this article; the figures cited in the following subsections should be read with the denominator and context of each one.

3.2. Ocular manifestations: sequence and chronology

The first visual symptoms are myopia, followed by night blindness (nyctalopia), usually in early or mid-childhood [4, 14]. In the European registry, ophthalmological symptoms began at a mean of 7.9 years (standard deviation 3.8; 3–16) in the ophthalmology subcohort and at 10 years (1–21) in the full registry, and diagnosis came, on average, only 5 to 7 years later [11, 12]. In the Manchester cohort, all patients over 6 years of age had ophthalmological symptoms, the presenting symptoms were progressive myopia in 10 of 18 and night blindness in 5 of 18, and the first record of chorioretinal changes occurred at a median age of 13 years (4–33) [40].

The patches of atrophy begin in the mid-periphery as early as the first decade; they are circular, well demarcated and with scalloped borders, enlarge, coalesce and advance towards the posterior pole, ending in loss of central vision and blindness [4, 10]. The fundus appearance is highly specific, and the retinopathy can be detected before the patient notices any visual disturbance [4]. On fundus autofluorescence, atrophic areas are hypoautofluorescent; the posterior pole is spared for a long time, and preserved macular islands maintain central vision for decades [5, 41, 42]. ClinGen dates irreversible blindness, with acuity worse than 20/200, to between 40 and 55 years, with wide variability [10]; Saudubray states that most patients become practically blind between 45 and 65 years [4]; the NEI sibling series placed blindness between the fourth and seventh decades [46]; the Amsterdam group describes severe visual impairment and functional blindness around the fourth or fifth decade [34]. These estimates come from old, small series with heterogeneous treatment, and there are no prospective data that allow a choice between them.

3.3. Myopia, cataract and maculopathy

The myopia of gyrate atrophy is high and progressive. In the European registry, all patients were myopic before cataract surgery, with a mean spherical equivalent of −7.9 D (−20.69 to −2.31) [12]; in the ophthalmology subcohort, the mean spherical equivalent was −8.96 D, with high myopia in 11 of 19 patients and worsening with age (r = −0.641), unlike RPE dystrophies, in which the mean is −0.10 D and high myopia affects fewer than 10% [11]. In Yunnan, high myopia was universal (mean −8.28 D; mean axial length 26.65 mm) [42]. ClinGen and Saudubray also report astigmatism, raised dark adaptation thresholds and, rarely, vitreous haemorrhage with sudden loss of vision [4, 10].

Posterior subcapsular cataract appears in late adolescence and is present in most patients by the end of the second decade [4, 10]. In the Amsterdam subcohort, 58% (11 of 19) had bilateral pseudophakia, with surgery at a mean age of 28.8 years (17–38), compared with about 73 years in the general population and 56.1 years in a series of 226 patients with retinitis pigmentosa, without intraoperative or postoperative complications [11]. In the full registry, 44% (12 of 27) had undergone cataract surgery, at a mean age of 28 years (18–37.5) [12]. In Manchester, 12 of 18 (66.6%) had presenile cataract in early adulthood, with 7 operated on at a median age of 27.5 years (22–36) [40]; in Yunnan, 4 of 7 had cataract, and they were older than the others (34.0 versus 8.0 years) [42].

Cystoid maculopathy is the main cause of loss of central vision before macular atrophy. In the Amsterdam subcohort, optical coherence tomography (OCT) showed cystoid maculopathy in 68% of patients (13 of 19; 22 of 38 eyes), with foveal involvement in 16 of those 22 eyes, bilateral in 48% and independent of age [11]. In Manchester, one third of patients (6 of 18) had macular oedema without leakage on angiography, and 78% had macular oedema and/or presenile cataract [40]. In Ankara, cystoid macular oedema (CMO) was present in 5 of 7 patients, including children aged 10 and 15 years, and 3 of 7 had hyperreflective foci in the ganglion cell layer [19]. In the Mexican cohort, spectral-domain OCT (SD-OCT) showed outer retinal tubulations in 58% of eyes, CMO in 50%, epiretinal membrane in 42%, foveoschisis and staphyloma in 17% and hyperreflective deposits in all [44]. In Yunnan, 12 of 13 eyes had parafoveal intraretinal cystoid spaces in the inner and outer nuclear layers, with disruption of the external limiting membrane in 91.67% and of the ellipsoid zone in 75%; the authors classified them as foveoschisis, because angiography showed no leakage, and state that they did not observe CMO [42]. The difference between «CMO» and «foveoschisis» seems to depend more on the angiographic criterion used than on the OCT appearance. Integrity of the foveal ellipsoid zone was continuous in 14 of 38 eyes, discontinuous in 9 and indiscernible in 15, with progressive loss between 25 and 40 years in three patients, while two patients under 18 years kept it continuous over 3 to 5 years of follow-up [11].

Peripapillary atrophy was present in 10 of 13 eyes in Yunnan and correlated with worse acuity (r = −0.593) and greater axial length (r = 0.500), but not with age; the authors propose it as a marker of severity [42]. Other changes described include epiretinal membrane (3 of 18 in Manchester; 38.46% of eyes in Yunnan), floaters, optic nerve drusen, chronic anterior uveitis and amblyopia [40, 42].

3.4. Visual function: acuity, fields and electroretinogram

Central visual acuity is maintained for decades and is therefore an insensitive outcome. In the European registry, best-corrected visual acuity (BCVA) worsened by 0.013 logMAR per year of age, with no association with plasma ornithine [12]; in the ophthalmology subcohort, median BCVA was 0.26 logMAR, with a correlation between age and logMAR of ρ = 0.685 [11]. In Manchester, median acuity at the last assessment was 0.36 logMAR (0.0–0.7) [40]. The visual field is more informative: on Humphrey 10-2 perimetry, the central 10 degrees were intact in 5 patients, all under 30 years, and in the rest the mean residual field was 6.08 degrees; mean foveal sensitivity was 33.23 dB and correlated with logMAR (r = −0.606) [11]. In the full registry, a residual central field greater than 10 degrees was present in 4 patients, all with protein restriction started before the age of 10 years, and fields of 20 degrees or more were recorded only in patients aged 25 years or younger [12]. On the Michigan Retinal Degeneration Questionnaire (MRDQ), administered to 8 adults, the greatest disability corresponded to scotopic vision, and the mean deviation on perimetry correlated with central vision, contrast sensitivity, scotopic vision and photopic peripheral vision [12].

The full-field ERG is reduced or undetectable in most patients at diagnosis, with markedly diminished rod and cone responses, and becomes abnormal with progression when still preserved [4, 14, 19]. In the Mexican cohort, the photopic and scotopic ERG were each preserved in only 17% of eyes [44]. The Ankara authors consider it useful for diagnosis and baseline recording, but insensitive as a marker of short-term therapeutic efficacy [19].

3.5. Vitreoretinal complications

Rhegmatogenous retinal detachment, recurrent macular neovascularisation and vitreous haemorrhage have been described, mostly in isolated reports [4, 19, 48, 49]. In Ankara, one patient had a rhegmatogenous detachment and was left without light perception in that eye at 27 years [19]. In a 25-year-old Polish man, a macula-off detachment, associated with atypical circumferential vitreous bands, was treated with vitrectomy, cryotherapy and gas tamponade, and final acuity was 6/9 [48]. Subfoveal macular neovascularisation, documented by OCT angiography in two women with a genetic diagnosis, was attributed to RPE dysfunction associated with high myopia [49].

3.6. Neonatal form with hyperammonaemia

Some patients, typically preterm infants, present in the first weeks with poor feeding progression, growth failure, symptomatic hyperammonaemia and orotic aciduria, a picture that mimics OTC deficiency [4, 13]. In this phase, plasma ornithine, citrulline and arginine are low, and hyperornithinaemia appears later [4]. Cleary and colleagues (2005) described two unrelated cases with enzymatic and molecular confirmation, in which the absence of hyperornithinaemia delayed the diagnosis [13]. An infant with a prenatal diagnosis had normal ornithine and arginine in cord blood but developed low values at 2–4 months on a normal diet, with a concomitant rise in ammonia and glutamine; arginine corrected the hypoargininaemia and hyperammonaemia and gave rise to hyperornithinaemia [4]. In the European registry, two Turkish siblings homozygous for p.(Arg426Ter) had neonatal hyperammonaemia with low-normal ornithine (56 µmol/L at birth in one of them), followed by a rapid rise, without permanent neurological sequelae and with regular schooling [12]. The Buijs review states that, without treatment, neonatal hyperammonaemia may go unnoticed or resolve spontaneously, and that intravenous arginine usually suffices to restore the urea cycle [5].

Sources differ on this point. The GeneReviews chapter on HHH syndrome states that OAT deficiency never presents with raised ammonia or glutamine or with excretion of homocitrulline or orotic acid, and Nyhan's Atlas writes that none of the hyperornithinaemias, except HHH, is hyperammonaemic [9, 31]. Saudubray, the European guideline on urea cycle disorders, the Cleary series and the European registry document neonatal hyperammonaemia in OAT deficiency [4, 12, 13, 50]. The French national protocol for urea cycle disorders (HAS, 2021) includes neonatal OAT deficiency among the «associated deficiencies» of the urea cycle, alongside carbonic anhydrase VA deficiency, lysinuric protein intolerance and citrullinaemia type II, and describes neonatal cases with symptomatic hyperammonaemia, orotic aciduria, faltering growth and encephalopathy; the proposed mechanism is substrate limitation, in which low ornithine and arginine reduce arginine synthesis and slow the cycle [51]. In the original Finnish series, ammonia was normal when fasting, after meals and under loading in nine patients in the hyperornithinaemic phase [3]. The statements do not contradict each other regarding the chronic phase: the hyperammonaemia of OAT deficiency is neonatal and accompanied by low ornithine, whereas the hyperornithinaemic phase is not hyperammonaemic. What GeneReviews and the Atlas do not consider is the neonatal phase.

3.7. Central nervous system and neurodevelopment

Most patients have normal intelligence, similar to that of their relatives [4]. The series differ, however, in the frequency of abnormalities. In the French series of 7 paediatric patients by Valayannopoulos (2009), 5 had behavioural changes and/or developmental delay; the authors considered a role for cerebral creatine deficiency, and Saudubray suggests also considering sequelae of unrecognised neonatal hyperammonaemia [4, 5, 35]. In Manchester, 6 of 18 (33%) had learning difficulties or developmental delay, in a mostly consanguineous cohort in which other genetic factors may contribute [40]. In the European registry, 4 of 27 (15%) had a diagnosed neurodevelopmental disorder (attention deficit hyperactivity disorder in 2, autism spectrum disorder in 1, dyslexia in 1) and 2 had developmental delay since childhood, one mild and the other non-verbal with regression; the neurological examination was normal in 12 of 13 assessed [12]. In the systematic review by Balfoort (2021), 9 of 97 patients had neurological symptoms, but only 13 of the 97 had a description beyond the ocular phenotype [5].

In the Finnish series by Valtonen (1999), with 37 patients according to the Buijs review, 3 had seizures, and changes on magnetic resonance imaging (MRI) and electroencephalography (EEG) did not correlate with the clinical picture [5, 52]. The OMIM summary describes the central comparison of that study with another denominator: 23 never-treated patients, compared with age-matched controls, and 9 patients who had been taking creatine or creatine precursors for several years. MRI showed degenerative white matter lesions in 50% of patients and premature atrophic changes in 70%, with a marked increase in the number of Virchow–Robin perivascular spaces; the EEG was abnormal in 58% of those recorded (slow background activity, focal changes or high-amplitude beta rhythm), with no association with MRI changes, age or sex; the MRI and EEG of supplemented patients did not differ from those of untreated patients [3, 52]. The two secondary sources attribute different denominators to the same study (37 patients; 23 untreated and 9 supplemented). An Italian report from 1978 described a child diagnosed at 3 years and 9 months with mild intellectual disability, language delay and speech disorder (Stoppoloni, 1978) [3]. Saudubray summarises these studies with diffuse slowing on EEG in about half of patients and early degenerative and atrophic changes on MRI, unrelated to age, in about 70% [4]. ClinGen collates focal epileptiform activity in 6 of 33 patients, including 5 of the 8 youngest (4–15 years), and describes intellectual disability, language difficulty, behavioural problems or epilepsy in a minority [10]. In the European registry, in 42 MRI reports from 22 patients, MR spectroscopy (MRS) showed cerebral creatine deficiency in 15 of 20 patients studied, 10 of whom were already supplemented, with cerebral creatine at about 45–70% of normal; there were non-specific white matter changes in 10, thinning of the corpus callosum in 2 and hypoplasia of the adenohypophysis with optic chiasm atrophy in two siblings [12]. The authors hypothesise that cerebral creatine of about 50% of normal is enough for normal neurodevelopment, which would contrast with primary creatine deficiencies (AGAT, SLC6A8 transporter) [12]. In Ankara and Yunnan, neurological comorbidities were isolated (epilepsy, depression, mild intellectual disability) [19, 42].

3.8. Peripheral nerve and muscle

In the Finnish series of 40 patients studied by Peltola (2002), with a mean age of 31.6 years (5–74), nerve conduction studies were abnormal in 21 (53%), in association with ophthalmological severity and age; quantitative sensory thresholds showed large-fibre impairment in 7 (18%) and small-fibre impairment in 4 (10%), and somatosensory and brainstem auditory evoked potentials were abolished in 5; involvement was largely subclinical, and ClinGen records non-disabling symptomatic neuropathy, with paraesthesiae, in 4 of 40 [5, 10, 21]. Muscle biopsy shows type II fibre atrophy with tubular aggregates and fatty change, present in all 7 patients biopsied in one series, and muscle creatine is depleted; some patients have muscle weakness [4, 10, 21]. In the study by Sipilä (1979), with 21 patients, type II fibre atrophy with tubular aggregates was almost universal; patients generally had no muscle symptoms, although some had reduced performance in tasks requiring speed or sudden strength, and progression to almost complete loss of type II fibres was slower than that of the ocular disease [3]. In the series by Valtonen (1996), the 7 patients biopsied had type II fibre atrophy and 6 had tubular aggregates, and computed tomography and MRI showed changes in the thigh muscles in all [3]. The hypothesis linking the muscle and ocular lesions to phosphocreatine deficiency induced by hyperornithinaemia prompted the creatine supplementation studies [3, 53]. The Finnish authors concluded that gyrate atrophy is a systemic disease, and Saudubray includes OAT deficiency in the lists of treatable inborn errors with peripheral neuropathy [21, 54].

3.9. Body composition and other findings

In Manchester, 9 of 11 patients assessed by bioimpedance had reduced fat-free mass, with a median body mass index of 23.4 kg/m² in adults [40]. In the European registry, the fat-free mass index was below the 25th centile in 8 of 11 (below the 10th centile in 5) and grip strength below the 25th centile in 5 of 11; four patients aged between 28 and 57 years had sarcopenia, only one on a restricted diet and none on creatine at that time, and growth and weight of patients on diet were normal [12]. The Manchester authors note that protein restriction may in itself reduce lean mass [40]. Hair may be fine, sparse and straight, with abnormalities on microscopy, and there are ultrastructural abnormalities of hepatic mitochondria [4, 54]. Other isolated comorbidities were recorded in Manchester (depression, psychosis, anorexia nervosa, spastic diplegia, congenital heart disease, precocious puberty, diabetes insipidus, among others), with no established causal relationship with the disease [40].

3.10. Natural history and treated disease

Most ophthalmological data come from patients with partial, late or interrupted treatment, and the separation between natural history and therapeutic effect is imperfect. The Yunnan cohort, without any treatment, showed progressive visual loss in all patients and night blindness in 75% [42]. In the treated cohorts, visual loss progressed in all Manchester patients despite treatment [40]. The clearest contrast is seen between siblings: in the European registry, a patient treated from the first year of life had no atrophic lesions at 11–12 years, whereas his sister, with poor adherence in the first decade, had peripheral atrophy at 25 years [12]; in Manchester, a patient diagnosed in childhood after hyperammonaemia and on a strict diet had an unusually mild phenotype at 23 years, whereas another, diagnosed at 9 years and non-adherent, had cataract and CMO at 22 years [40]. Patients with a significant reduction in ornithine on pyridoxine usually have a milder course and retain central vision to older ages [4]. The children of women with OAT deficiency on an unrestricted diet do not appear to suffer adverse effects from exposure to hyperornithinaemia in pregnancy [4].

4. Clinical Recognition and Warning Signs

4.1. When to suspect it, by age

In the newborn and young infant, especially if preterm, suspicion should arise in the presence of hyperammonaemia with orotic aciduria and low plasma ornithine, citrulline and arginine, a profile that mimics OTC deficiency [4, 13]. The European guideline on urea cycle disorders includes OAT deficiency in the diagnostic algorithm for neonatal hyperammonaemia and among the other causes of hyperammonaemia, alongside oxidative phosphorylation defects, citrin deficiency and lysinuric protein intolerance [50]. A newborn with neonatal hyperammonaemia and low ornithine who develops hyperornithinaemia in the following months has the characteristic pattern [4, 12].

In the school-age child, the warning signs are high, progressive myopia, night blindness and the detection of peripheral atrophic patches on fundus examination; early high myopia not explained by familial factors justifies examination of the retinal mid-periphery [4, 11, 42]. In the adolescent and young adult, posterior subcapsular cataract before the age of 30 years, cystoid maculopathy associated with high myopia and concentric constriction of the visual field should prompt measurement of plasma amino acids [11, 40]. Diagnosis is often late: the mean interval between first symptoms and diagnosis was about 5 years in the European registry and 6.8 years in the ophthalmology subcohort, and most patients are diagnosed in the second decade, already with chorioretinal atrophy [5, 11, 12].

In the family setting, all siblings of a patient should be studied, even if asymptomatic: 12 of 27 patients in the European registry and 5 of 19 in the ophthalmology subcohort were identified by targeted testing after diagnosis in a sibling, most already with myopia [11, 12]. In neurology and metabolic medicine, OAT deficiency enters the differential diagnosis of peripheral neuropathy and treatable retinopathies, of cataract associated with inborn errors and of secondary creatine deficiency [4, 54].

4.2. Emergency signs

Emergency situations are symptomatic neonatal hyperammonaemia, which requires treatment as an undiagnosed urea cycle disorder (section 7.7), sudden loss of vision due to retinal detachment, vitreous haemorrhage or macular neovascularisation, and hyperammonaemia or poor growth in a patient on excessive arginine restriction [4, 48–50].

4.3. Newborn screening

Newborn screening based on ornithine is not reliable, because hyperornithinaemia is not present in all, and perhaps in none, of affected newborns [4, 5]. The proline/citrulline ratio has been proposed as a neonatal marker; in the European registry, a ratio above 32 in the neonatal period, followed by persistent hyperornithinaemia, was accepted as a diagnostic criterion [5, 12]. A 2026 letter from Colorado describes the diagnosis of OAT deficiency from low citrulline on dried blood spot newborn screening; the full text was not read, and the observation is consistent with the neonatal physiology described in section 2.2 [55]. No newborn screening programme systematically includes this disease in the sources consulted.

5. Diagnosis: Criteria, Tests and Algorithms

5.1. Diagnostic criteria

There are no formal international criteria. The Amsterdam international registry uses three alternative criteria: biallelic pathogenic variants in OAT; or hyperornithinaemia with typical fundus lesions described by an ophthalmologist; or a proline/citrulline ratio above 32 in the neonatal period followed by persistent hyperornithinaemia [12]. With these criteria, a molecular diagnosis was obtained in 26 of 27 patients (96%) [12]. ClinGen considers plasma amino acids showing hyperornithinaemia the confirmatory biochemical test, the characteristic fundus highly suggestive but not confirmatory, and sequencing of OAT the molecular confirmation that enables cascade testing [10].

5.2. Biochemistry

Outside the neonatal period and on an unrestricted diet, plasma ornithine is 5 to 20 times above normal; the fasting mean given in the textbook The Online Metabolic and Molecular Bases of Inherited Disease (OMMBID) and reproduced in Saudubray is 915 µmol/L (400–1200) [4, 56]. Recent cohorts, with many non-fasting samples and partially treated patients, record values at diagnosis of 735.7 µmol/L on average (341–1207) in the European registry, 792.0 µmol/L on average (399–1086) in the ophthalmology subcohort, a median of 800 µmol/L (458–1244) in Manchester and 326.8–1462 µmol/L in Ankara [11, 12, 19, 40]. Ornithine rises with age, by about 8.1 µmol/L per year, independently of protein intake and adherence [12]. Reference ranges depend on the laboratory (27–98 µmol/L in Amsterdam, 40–150 µmol/L in Manchester, 30–110 µmol/L in the GeneReviews chapter on HHH syndrome), and the unit nmol/mL used in some laboratories is equivalent to µmol/L [9, 12, 40, 41].

The pyridoxine-responsive form tends to have lower ornithine than the non-responsive form, but the baseline level does not reliably distinguish between the two [4]. In the Yunnan cohort, ornithine measured in dried blood spots by liquid chromatography–tandem mass spectrometry (LC-MS/MS) was 97–319 µmol/L, with a reference of 15–80 µmol/L, modest values that the authors attribute to the method and diet; the dried blood spot is not equivalent to plasma [42]. Normal plasma ornithine in a genetically confirmed patient, described in the Mexican cohort, shows that a normal value does not exclude the diagnosis when the diet and sampling conditions are not known [44].

5.3. Ophthalmological assessment

The reference assessment, according to ClinGen (which follows the recommendations of the American Academy of Ophthalmology, AAO) and the 2022 AAO guideline, includes ocular and medical history with retinotoxic medicines and family history, best-corrected visual acuity, slit-lamp biomicroscopy, intraocular pressure, dilated indirect ophthalmoscopy, colour or wide-field fundus photography, SD-OCT, fundus and infrared autofluorescence, kinetic and static visual fields or microperimetry, and full-field ERG when appropriate [10, 15]. The Oxford Desk Reference adds full electrophysiology according to the International Society for Clinical Electrophysiology of Vision (ISCEV) [14]. High-density SD-OCT is the test for detecting and following cystoid maculopathy, epiretinal membranes and macular schisis [15]. In children, functional tests are often not possible; the AAO suggests examinations under sedation between 2 and 6 years, weighing the risks, and notes that sedation reduces ERG amplitude and that visual fields before the age of 7 years are unreliable [10, 15].

5.4. Molecular testing

Sequencing of OAT, alone or as part of an inherited retinal dystrophy panel, confirms the diagnosis and identifies the familial variants for cascade testing [10, 14]. The AAO considers genetic testing appropriate for most patients with suspected genetic retinal degeneration and states that the causal variant is identified in 56–76% of patients with inherited retinal dystrophies; this figure is the overall yield of panels in that group and not the sensitivity of the test for OAT deficiency [15]. The sensitivity of OAT sequencing in patients with a biochemical diagnosis is not quantified in the sources consulted; in the European registry, biallelic variants were identified in 96% of patients [12].

The limits of the test should be known. Exon-based panels may miss non-coding variants: in Yunnan, the c.-30+22_-30+43del variant, in the 5' untranslated region and intron 1, was detected only by whole-genome sequencing after the panel was negative, and was initially classified as a variant of uncertain significance (VUS) before segregation and a minigene assay demonstrated exon 1 skipping [42]. The sources consulted do not quantify the fraction of exon deletions or duplications (copy number variation) in OAT; their detection depends on coverage analysis or targeted methods that not all laboratories apply, and a negative result does not exclude the diagnosis in the presence of hyperornithinaemia and a typical fundus. Historical descriptions include structural and splicing variants that an analysis restricted to coding sequences and exon–intron junctions may not detect or interpret: a 1072-bp deletion removing the whole of exon 6, identified by Southern blot, the 9-bp deletion at the junction between intron 4 and exon 5, and the insertion of an Alu fragment triggered by the deep intronic variant c.199+303C>G [7, 30]. Before routine sequencing, SSCP screening detected variants in 19 of 24 uncharacterised alleles (79%), a value that measures the performance of that technique and not the sensitivity of current methods [7]. The report should state the gene, transcript and version (NM_000274.4), HGVS nomenclature at c. and p. level, zygosity and classification according to the criteria of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology (ACMG/AMP); the phase of the two variants in a compound heterozygote is confirmed by parental testing [41, 42]. Faced with a VUS, hyperornithinaemia, a history of neonatal hyperammonaemia and typical retinal lesions support pathogenicity, and family segregation and functional studies allow it to be reclassified [5, 42]. The same variant may be classified differently between laboratories: p.(Gly91Glu) is likely pathogenic for the authors of a 2026 case (criteria PS1, PM2 and PP3) and has conflicting classifications in ClinVar [30, 41].

5.5. Enzyme studies and in vitro assessment of pyridoxine response

OAT activity can be measured in cultured skin fibroblasts or lymphoblasts and has also been shown to be deficient in liver, muscle and hair roots [4]. In fewer than 10% of patients, fibroblast activity increases significantly when the assay is performed with high concentrations of PLP [4]. The Amsterdam group developed an activity and labelled ornithine flux assay in fibroblasts and an in vitro pyridoxine response test, concordant with the in vivo response in the only responsive line, and proposes integrating these tests into a diagnostic pathway [43]. Residual activity in fibroblasts may help to predict outcome and to select patients for the therapeutic trial [5]. Studies of PLP kinetics and in vitro stimulation in fibroblasts from patients and heterozygotes (sections 2.6 and 2.7) are research tools, not validated as clinical tests [3].

5.6. Therapeutic trial of pyridoxine

Saudubray recommends a 2- to 4-week trial of pyridoxine hydrochloride (300–500 mg/day) in all newly diagnosed patients, without changing protein intake, comparing fasting plasma ornithine before and after [4]. A response is defined as a reduction of more than 30% in baseline fasting ornithine with constant protein intake [4]. In Manchester, all new patients, regardless of genotype, received at least 200–400 mg daily for a minimum of 2 weeks [40]. The historical NEI protocol used 500 mg/day for 3 to 6 months before starting the diet [24]. The trial should precede the diet, because simultaneous protein reduction makes it impossible to attribute the fall in ornithine to the vitamin [4, 19].

5.7. Diagnosis in the neonatal phase

In the newborn with hyperammonaemia, OAT deficiency is difficult to distinguish from OTC deficiency, because both have low ornithine, citrulline and arginine and raised orotic acid [4]. Diagnosis rests on sequencing of OAT (usually in a hyperammonaemia panel or exome) and on reassessment of ornithine in the following months, which shows the characteristic rise [5, 12, 13]. The guideline on urea cycle disorders notes that hyperornithinaemia with hyperammonaemia can occur in OAT deficiency and in HHH syndrome, and that homocitrullinuria characterises the latter [50]. The 2021 French protocol places neonatal OAT deficiency among the primary causes of hyperammonaemia, once secondary causes have been excluded (hepatocellular failure, neonatal sepsis, urease-producing bacteria, medicines such as valproate and asparaginase, portocaval shunt, patent ductus venosus and transient hyperammonaemia of the newborn), and requests, for any hyperammonaemia, blood gases, lactate, electrolytes, blood glucose, liver function, coagulation, creatine kinase, urinary ketone bodies and markers of infection, together with plasma amino acids, acylcarnitine profile, urinary organic acids and, if possible, urinary orotic acid, with results intended within 48 hours and without delaying treatment [51]. In fulminant progression, the same protocol recommends storing EDTA blood for DNA extraction, frozen plasma, serum and urine and a skin biopsy for fibroblast culture, and confirming the diagnosis by next-generation sequencing in a panel of urea cycle genes, with family segregation analysis before offering prenatal or preimplantation diagnosis [51]. The table of biochemical profiles in that protocol assigns to neonatal OAT deficiency normal or raised orotic acid, raised glutamine, normal citrulline and raised ornithine; raised ornithine contradicts the annex of the same document, which describes hypo-ornithinaemia and hypoargininaemia in this phase, and the published series, in which neonatal ornithine is low or low-normal and rises only in the following months (table in section 5.2) [4, 12, 13, 51].

5.8. Practical algorithm

5.9. Prenatal and preimplantation diagnosis

Once the familial variants have been identified, prenatal diagnosis and carrier detection are carried out by molecular analysis [4]. A prenatal diagnosis has been described in an infant who had normal ornithine and arginine in cord blood and developed hypo-ornithinaemia at 2–4 months, which shows the importance of biochemical follow-up from birth in a newborn with a prenatal diagnosis [4]. Reproductive options are covered in section 8.

6. Differential Diagnosis

6.1. Comparative table

In the table, AR denotes autosomal recessive inheritance and AD autosomal dominant inheritance.

6.2. Hyperornithinaemias and hyperammonaemias

HHH syndrome is the main biochemical differential diagnosis. It results from biallelic variants in SLC25A15, which encodes the mitochondrial ornithine transporter, and is defined by the triad of hyperornithinaemia, episodic or chronic hyperammonaemia and homocitrullinuria [4, 9]. Plasma ornithine at diagnosis ranges between 200 and 1915 µmol/L and is, on average, slightly lower than in OAT deficiency; presentation is neonatal in about 8% of cases, and manifestations include protein intolerance, vomiting, seizures, developmental delay, hepatic dysfunction (more than one third in a French-Canadian series) and progressive spastic paraparesis [4, 9]. The fundus is usually normal: Saudubray reports retinal abnormalities in only 2 of more than 70 patients, and Nyhan's Atlas describes papilloedema during acute hyperammonaemia in one patient and retinal depigmentation with choroidal thinning, without functional loss, in another [4, 31]. Urinary homocitrulline is the discriminating test, with two pitfalls: on some analysers it may be confused with methionine, and it may appear in the urine of infants fed formulas or heated milk [4]. The French-Canadian founder variant c.562_564delTTC p.(Phe188del) reaches an incidence of 1:1,550 live births in an isolated population in northern Saskatchewan [9].

In the neonatal phase, the differential diagnosis is with OTC deficiency and the other proximal urea cycle disorders, because OAT deficiency shares low ornithine, citrulline and arginine and raised orotic acid [4]. The distinction is made by genetics and by the course: in OAT deficiency, ornithine rises in the following months [4, 12]. Lysinuric protein intolerance has the opposite pattern, with dibasic amino acids low in plasma and high in urine [9, 57]. P5CS deficiencies present with hypo-ornithinaemia, hypocitrullinaemia, hypoargininaemia and hypoprolinaemia, and hyperprolinaemias types I and II (deficiencies of proline oxidase, gene PRODH, and of P5C dehydrogenase) with raised proline; Saudubray's comparative table groups these disorders of ornithine and proline metabolism [4].

6.3. Chorioretinal dystrophies and treatable retinopathies

Gyrate-appearing chorioretinal atrophy with normal plasma ornithine outside the neonatal period should prompt a search for another cause: OMIM considers that there is no true gyrate atrophy without hyperornithinaemia and that such cases are phenocopies, although one exception has been published with a genetic diagnosis and ornithine of 48 µmol/L (section 2.7) [3, 44]. Ophthalmologically, the AAO groups gyrate atrophy and choroideremia in the class of chorioretinal degenerations, and plasma ornithine discriminates between the two [15]. Choroideremia has X-linked inheritance [15]. The gyrate atrophy-like choroidal dystrophy associated with C1QTNF5 was described in 2025 in a 70-year-old woman with the heterozygous variant c.538C>G p.(Gln180Glu), concentric visual field defects and photophobia of about 8 years' duration, acuity of 20/40 and 20/32 and symmetrical peripheral and peripapillary chorioretinal atrophy with foveal sparing, without the typical features of late-onset retinal degeneration associated with the same gene; dominant inheritance and age at onset distinguish it [59]. A 2026 multicentre series described the heterozygous variant EFEMP1 c.418C>T p.(Arg140Trp) in three families with late-onset peripheral retinal degeneration, with predominant rod dysfunction; two carriers, with symptom onset at 40 and 55 years, had lobular peripheral chorioretinal atrophy initially labelled as gyrate atrophy-like or choroideremia-like [60]. Pathological myopia is the most difficult differential when high myopia dominates the picture: circular, well-demarcated peripheral lesions with centripetal progression distinguish gyrate atrophy from the diffuse tessellated fundus, posterior staphyloma and lacquer cracks of pathological myopia, and absence of leakage on angiography rules out inflammatory or vascular retinopathies [42].

The Oxford Desk Reference includes plasma amino acids, alongside phytanic acid, very long-chain fatty acids, apolipoprotein B and vitamins A and E, in the metabolic investigation of retinal dystrophies, because these diseases have specific treatment: Refsum disease, abetalipoproteinaemia (vitamins A and E started before the age of 2 years attenuate the retinopathy), vitamin E deficiency and gyrate atrophy [14]. Syndromic differentials include the neuronal ceroid lipofuscinoses, Cohen, Senior–Løken, Alström, Cockayne and Joubert syndromes and the Zellweger spectrum disorders [14].

7. Treatments and Risk Management Strategies

7.1. Aims and framework of the recommendations

There are no international guidelines or formal treatment recommendations for gyrate atrophy; the review by Buijs and colleagues (2025) states this expressly, and the Dutch group wrote a national multidisciplinary care pathway based on the 2021 systematic review, which was not located in a published version [5, 32]. The level of evidence for the interventions is generally 4 to 5 on the Oxford Centre for Evidence-Based Medicine (OCEBM) scale [5]. The systematic review by Balfoort (2021), with a search up to December 2020, brought together 33 studies and 107 patients, mainly case reports (27) and series (4), without randomised trials or large cohorts, and concluded that the absence of predefined clinical outcomes precludes conclusions on clinical efficacy [32]. The sources guiding practice are the ClinGen actionability summary (AC1014), the AAO guideline for the assessment of inherited retinal dystrophies (2022), the European guideline on urea cycle disorders (2012, for neonatal hyperammonaemia), the European Delphi consensus on therapeutic nutritional products (2026) and the reference chapters of Saudubray [4, 10, 15, 50, 64].

Treatment pursues three targets: lowering plasma ornithine, correcting secondary creatine deficiency and treating ophthalmological complications [5]. The historical biochemical target is ornithine below 200 µmol/L [4]; in practice, normalisation is rarely achieved, even with strict adherence, and the NEI study associated a mean ornithine maintained between 400 and 500 µmol/L with slower functional loss [5, 47]. Treatment slows the degeneration but does not reliably halt it: some patients continue to lose retinal function with low ornithine [10, 34]. No medicine is authorised for this disease; all the uses described below are off-label or rely on food supplements.

7.2. Pyridoxine (vitamin B6)

Pyridoxine is the precursor of PLP, the OAT cofactor, and acts in some variants as a pharmacological chaperone that stabilises the enzyme or restores the holoenzyme form [33, 45]. The 2026 European Delphi consensus, with 36 experts from 18 countries, classified pyridoxine as an essential therapeutic nutritional product for OAT deficiency, with level 4 evidence [64]. The therapeutic trial should be performed in all newly diagnosed patients, regardless of genotype, before any change in diet (section 5.6) [4, 40].

The proportion of responders varies widely between sources. Saudubray states that fewer than 10% of patients have an in vitro enzymatic response and partial reduction of ornithine in vivo; the Buijs review cites about 5% according to OMMBID; the 2021 systematic review found a response in 30% of the variants included, without a clear dose–response relationship; in the European registry, only 3 of 27 patients responded in vivo, with quantified reductions of 24% and 36% [4, 5, 12, 19, 32]. The difference probably reflects publication bias towards responsive cases in the systematic review and the allelic composition of the cohorts (for example, the lack of response in p.Gly353Asp, frequent in the Netherlands) [43]. The response is, according to ClinGen, inconsistent: it occasionally slows the degeneration, and most patients do not respond [10]. Pyridoxine responders usually have a milder course [4].

As for dose, the reference is that of the Saudubray chapter, the only consensus source with a dose for this condition: a trial of 300–500 mg/day of oral pyridoxine hydrochloride for 2 to 4 weeks; doses of 200–500 mg/day reduced ornithine by 25–60% in a small number of patients; that chapter does not define a maximum dose or distinguish dose by age or weight [4]. The medicines table in the same book gives, for the responsive forms of cystathionine β-synthase, γ-cystathionase and OAT deficiencies, among others, 50–500 mg/day orally, without a maximum dose defined in the source for OAT deficiency; for cystathionine β-synthase deficiency, an analogous condition treated with the same medicine, it gives up to 10 mg/kg/day with a maximum of 500 mg/day, the reference by extrapolation for the child [65]. ClinGen cites 120–600 mg/day [10]. In cohort practice, Manchester uses at least 200–400 mg/day for a minimum of 2 weeks, the European registry recorded a mean dose of 260 mg/day (50–600) with the quantified responder taking 500 mg/day, and the NEI protocol used 500 mg/day for 3 to 6 months [12, 24, 40]. The only United States Food and Drug Administration (FDA) label consulted, for injectable pyridoxine hydrochloride, does not mention gyrate atrophy; for vitamin B6 dependency syndrome it allows a therapeutic dose of up to 600 mg/day, followed by 30 mg/day for life, values that coincide with the upper limit of the clinical sources but refer to another indication and another route [66]. Long-term oral pyridoxine for gyrate atrophy is thus outside the conditions of any label consulted. Adverse effects to monitor are sensory peripheral neuropathy at high doses (described above 900 mg/day in adults), paraesthesiae, somnolence and a fall in serum folate; pyridoxine antagonises levodopa when it is not combined with carbidopa [65, 66]. High-dose oral pyridoxine is marketed in the United States as a supplement, without a medicine label [5, 64].

7.3. Diet restricted in arginine or natural protein

Diet is the main treatment and the only one with evidence, albeit observational, of slowing visual loss [5, 10]. Dietary arginine is the main precursor of ornithine; dietary proteins contain on average 4–6% arginine, and nuts and seeds contain more [4]. Restricting arginine sufficiently usually requires marked restriction of natural protein and completing requirements with an arginine-free essential amino acid mixture, with adequate calories, vitamins and minerals to avoid malnutrition and catabolism, especially in children [4, 10]. ClinGen lists the arginine-rich foods to avoid: nuts, seeds, cereals, dairy products, shellfish, meat, chicken, watermelon and chocolate [10].

Published regimens vary between centres. In Manchester, the natural protein target in adults is 0.6–0.8 g/kg/day; when it falls below 0.8 g/kg/day, an essential amino acid mixture is added up to a total protein of 1 g/kg/day [40]. In the European registry, natural protein intake ranged between 0.18 and 1.1 g/kg/day (most between 0.43 and 0.85), with arginine-free essential amino acids when intake fell below dietary reference values, and the median duration of diet was 8 years [12]. The NEI protocol initially used an almost protein-free diet, with more than 2000 kcal of carbohydrate and fat, low-protein foods and amino acid supplements, with admission for 1 to 3 weeks, later moving to a less restricted low-protein diet [24]. An Italian woman, intolerant of the semi-synthetic diet, followed a low-protein diet with natural foods (0.8 g/kg/day) for 26 years according to the title of the paper and Saudubray (28 years in the published abstract), with a moderate reduction in ornithine and greatly delayed chorioretinal progression, which makes this approach an alternative for those who refuse amino acid supplements [4, 67].

The evidence of benefit comes from observational studies. In the NEI study, 17 patients adherent to the arginine-restricted diet, compared with 10 non-adherent patients, with mean follow-up of 13.9 and 14.1 years, had significantly slower rates of decline in the combined ERG, the flicker ERG and kinetic perimetry after adjustment for age (p < 0.05), and a trend in static perimetry (p = 0.06); the authors concluded that keeping mean ornithine below 400–500 µmol/L slows functional loss [47]. ClinGen grades this study at level 5, because the comparison group was self-selected by adherence [10]. In six sibling pairs (1991) and in two pairs followed for 16–17 years (2002), the sibling who started the diet earlier had slower progression of the lesions and, to a lesser extent, of retinal function [46, 68]. In the European registry, mean ornithine was 552.8 µmol/L in adherent patients, 718.8 µmol/L in those with adherence difficulties and 804.9 µmol/L without diet (p < 0.0001 between diet and no diet); natural protein intake and adherence were associated with ornithine, genotype did not modify the response, and values below 100 µmol/L occurred only in the neonatal period and in one patient with extreme restriction (0.18 g/kg/day) [12]. Fifteen of 27 patients retained more than 10 degrees of central field, and 7 of them had started the diet before the age of 10 years; of the 6 patients in the subcohort who started the diet by the age of 10, one had no lesions and three had lesions with the central 10 degrees intact and decimal acuity of 0.7–1.2 between 11 and 24 years [11, 12]. In Manchester, ornithine fell by 22% with restriction (n = 8) and by 7% without restriction (n = 6), with no significant difference in acuity at 5 years; in 7 patients with serial acuity, acuity fell significantly over 11 years in the no-diet group (n = 4; p = 0.025), but not in the diet group (n = 3); adherence was below 70%, and visual loss progressed in all [40]. The Amsterdam group concludes that diet may slow progression if started in the first decade, but does not prevent severe visual impairment [12]. In the Oat-null mouse, arginine restriction completely prevented retinopathy, which supports early treatment [4].

Adherence is the main obstacle. The registry of the GYROS natural history study (Gyrate Atrophy Ocular and Systemic Study) estimates that only about 20% of patients manage to follow the diet, and identifies adolescence, intercurrent illness and pregnancy as critical times [69]. In Ankara, adherence to diet and medication was poor and two patients were lost to follow-up [19]. Ornithine rises with age independently of protein intake and adherence, and the authors of the European registry warn that using ornithine alone as an indicator of adherence may lead to adults being unfairly accused of non-adherence [12].

Excessive arginine restriction has its own risks. Arginine becomes an essential amino acid in OAT deficiency, and hypoargininaemia causes poor growth, skin rash and even hyperammonaemia, especially with high total nitrogen intake [4]. Follow-up includes growth, physical examination, nutritional status and plasma amino acids [4]. In infants under 3–4 months of age, arginine should not be restricted until plasma ornithine begins to rise [4].

7.4. L-lysine

L-lysine reduces ornithine by competition in the renal tubular reabsorption of dibasic amino acids, which increases urinary excretion of ornithine [40, 65]. Two small studies with 10–15 g/day described reductions of 21–31% in plasma ornithine, and one of them included 5 pyridoxine-non-responsive patients [5, 19]. The systematic review found reductions of 21–34% [32]. The reference adult dose, in Saudubray's medicines table, is 10–15 g/day of L-lysine hydrochloride divided into 5 doses, orally, without a maximum dose defined in the source and without a paediatric dose for this indication [65]. In Manchester, 4 g/day is used in patients with plasma lysine below 100 µmol/L (some with 5–10 g/day), and in the European registry 2–15 g/day; in the latter, three patients stopped because of gastrointestinal effects, another had poor adherence for the same reason, and the clinical notes showed no marked effect on ornithine [12, 40]. The authors warn of the theoretical risk of anorexia, with catabolism and a rise in ornithine [12]. There is no medicine label for lysine in this indication, and its combination with the other modalities has not been studied [5].

7.5. Creatine

Creatine monohydrate corrects the secondary creatine deficiency. In the Finnish series by Heinänen (1999), it corrected the phosphorus-31 MRS spectrum of muscle, and creatine or its precursors partially restored the creatine peak in brain and muscle, without evident symptomatic effect and without halting retinal degeneration [4, 5, 53]. ClinGen states that creatine reduced or eliminated the type II fibre changes, which reappeared in two patients who stopped it, and that chorioretinal progression continued in a series of 13 patients followed for 5 years [10]. Saudubray's medicines table gives no dose for OAT deficiency (only for GAMT and AGAT deficiencies) [65]. The available dosing reference is that of the European registry: 2–22.5 g/day, most between 2.5 and 10 g/day, and 5–8 g/day in children; plasma creatine normalised in those supplemented, but cerebral creatine remained low in many, and one patient normalised it with 14 g/day [12]. The Manchester paper publishes «1.5 mg four times a day», a unit that is probably a typographical error for grams and should not be used as a dose [40]. Creatine is a supplement without a medicine label, and the 2026 Delphi consensus did not include it among the essential products for OAT deficiency [64].

7.6. Proline and other approaches

Oral proline was used in a single series, with improved acuity in one patient, and in a historical trial of 4 patients with proline or lysine the ophthalmological results were mixed [5, 10]. Proline was not used in the European registry, because plasma proline was normal in most (median 204 µmol/L) [12]. There is no consensus on proline or on creatine, whose modalities vary between centres [19].

7.7. Neonatal hyperammonaemia

The neonatal hyperammonaemia of OAT deficiency responds to standard treatment, in particular to arginine supplementation [4]. Infants with symptomatic hyperammonaemia or with signs of impaired nitrogen metabolism (hyperglutaminaemia, orotic aciduria) should receive arginine, and intravenous arginine usually suffices to restore the urea cycle [4, 5]. Saudubray gives no dose for this situation. The reference is, by extrapolation, the dose in the European guideline on urea cycle disorders for the patient with hyperammonaemia still without a diagnosis, whose algorithm includes OAT deficiency (UCD guideline, from urea cycle disorders): L-arginine hydrochloride 250 mg/kg (up to 400 mg/kg; 1–2 mmol/kg) intravenously in 10% glucose, as a bolus over 90–120 minutes, followed by maintenance of 250 mg/kg/day (1.2 mmol/kg/day), not exceeding 12 g/day, with sodium benzoate 250 mg/kg as a bolus over 90–120 minutes and maintenance of 250–500 mg/kg/day (5.5 g/m²/day above 20 kg) while the diagnosis is not established; the grade of recommendation is D [50]. The 2021 French national protocol, which draws on the European consensus revised in 2019, keeps the same doses for the undiagnosed patient and makes the maxima explicit: sodium benzoate at most 6 g in the 2-hour bolus and 12 g/24 h in maintenance, L-arginine hydrochloride and sodium phenylbutyrate at most 12 g/24 h each; L-citrulline is not indicated while the deficiency has not been identified [51]. For hyperammonaemia above 100 µmol/L without a diagnosis, the same protocol adds carglumic acid (Carbaglu®) orally or enterally, with a loading dose of 100 mg/kg and maintenance of 100–250 mg/kg/day in 4 doses until the final diagnosis (100 mg/kg/day in the general initial treatment table; the upper limit of 250 mg/kg/day is the only maximum given), and, until an organic aciduria is excluded, L-carnitine 50–100 mg/kg/day (continuous infusion or 3 doses), hydroxocobalamin 1 mg/day intramuscularly and biotin 10–20 mg/day orally; it stops protein for a maximum of 24–48 hours, withdraws lipids until a beta-oxidation defect is excluded and uses a hyperosmolar poly-ionic glucose infusion with age-based rates (6 mL/kg/h, about 10 mg/kg/min of glucose, between 0 and 24 months), with clinical, ammonia and acidosis reassessment every 3–6 hours [51]. As for preparation, the protocol describes intravenous arginine hydrochloride in a 400 mL bag at 6.25% (25 g) or in a 20 mL ampoule at 21%, administered via a Y-connector at a maximum rate of 1 g/kg/h, with a risk of necrosis on extravasation; sodium benzoate in a 1 g/10 mL ampoule diluted in an equal volume of 10% glucose, through a dedicated venous line, taking into account the sodium load (about 7 mmol per gram) [51]. Extracorporeal clearance (continuous veno-venous haemodiafiltration as the method of choice, haemodialysis as an alternative, peritoneal dialysis not advised) is discussed case by case; persistent hyperammonaemia above 500 µmol/L in the child or 200 µmol/L in the adult, acidosis with pH below 7.10 and deep coma are unfavourable criteria [51]. The only FDA label consulted for intravenous arginine hydrochloride (R-Gene 10®) is approved only as a diagnostic test of growth hormone reserve, with 0.5 g/kg over 30 minutes in children (maximum 30 g) and 30 g in adults, and warns of fatal paediatric overdoses with hyperchloraemic metabolic acidosis and cerebral oedema; the guideline dose is outside the conditions of this label, and the solution contains 47.5 mEq of chloride per 100 mL [70]. After the acute phase, the child continues arginine supplementation while ornithine has not risen, and restriction begins only after 3–4 months, once hyperornithinaemia is established [4].

7.8. Ophthalmological complications

Cystoid maculopathy is treated with topical or oral carbonic anhydrase inhibitors, topical non-steroidal anti-inflammatory drugs (NSAIDs), intravitreal or periocular corticosteroids and, in selected cases, intravitreal anti-VEGF (vascular endothelial growth factor) agents; the structural benefit is short-lived and the effect on visual function is uncertain, without large studies [5, 40]. In Manchester, 5 of 6 patients with macular oedema received oral acetazolamide [40]. In the Amsterdam subcohort, 5 patients received carbonic anhydrase inhibitors, usually topical; mild residual maculopathy was the rule, and only one patient became cyst-free, with recurrence on stopping; the authors consider the role of systemic inhibitors limited because of adverse effects and uncertain long-term benefit [11]. In Ankara, topical nepafenac four times a day in 5 of 7 patients produced no sustained response; combination with oral acetazolamide twice a day produced mild regression in one patient; brinzolamide, experimental sub-Tenon platelet-rich plasma (without improvement), coenzyme Q10 with topical vitamin E (transient complete regression in one patient) and intravitreal aflibercept (stopped for lack of response) were also used, without sustained improvement in acuity or CMO [19]. Octreotide was used in two patients and stopped because of inefficacy or adverse effects [5]. In a 4-year-old boy, CMO and foveoschisis improved rapidly after starting the arginine-restricted diet and supplementation, and in a 20-year-old pyridoxine-responsive man the low-protein diet with pyridoxine stabilised vision; these are isolated cases, known only from the abstract [71, 72].

None of these medicines has a studied dose for the CMO of gyrate atrophy; doses follow the labels for other indications, and the extrapolation should be declared to the patient. The FDA label for acetazolamide tablets gives, in chronic glaucoma, 250 mg to 1 g per 24 hours, in divided doses above 250 mg, with no additional benefit above 1 g; it is contraindicated with low sodium or potassium, marked renal or hepatic disease, adrenal insufficiency, hyperchloraemic acidosis and cirrhosis, may cause severe sulfonamide-type reactions (Stevens–Johnson syndrome, toxic epidermal necrolysis, hepatic necrosis, agranulocytosis), requires electrolyte monitoring and has been associated with growth retardation in children on prolonged therapy [73]. The label for dorzolamide 2% eye drops gives one drop three times a day in the affected eye (indication: raised intraocular pressure), with burning or discomfort in about one third of patients, bitter taste and risk of sulfonamide hypersensitivity through systemic absorption [74]. The label for nepafenac 0.1% (Nevanac®) gives one drop three times a day, from the day before until 2 weeks after cataract surgery, with increased bleeding time, delayed healing and keratitis as precautions; use four times a day in Ankara exceeds the label [19, 75].

Macular neovascularisation responds to intravitreal anti-VEGF agents: in one patient, acuity went from 20/60 to 20/30 after three monthly injections followed by a treat-and-extend regimen; in another, five recurrences over 5 years were treated and acuity stabilised at 20/40, with persistent subretinal fibrosis in both, which justifies prolonged surveillance [49]. Cataract is often operated on at the end of the third decade, with good results and without complications in the Dutch series [5, 11]. In rhegmatogenous retinal detachment, vitrectomy with cryotherapy and gas tamponade achieved reattachment; the authors suggest considering scleral buckling for inferior breaks, because of the presence of atypical circumferential vitreous bands [48].

7.9. Visual rehabilitation and support

Low vision rehabilitation should start early, with benefit to the quality of life of adults with visual impairment [5]. The AAO recommends low vision assessment at baseline and every 1–2 years, and ClinGen includes educational support, orientation and mobility training and information on registries and trials in supportive care [10, 15]. Follow-up is multidisciplinary, with paediatrics or internal medicine, metabolic medicine, ophthalmology, neurology, genetics and dietetics [10, 40].

7.10. Dosing table

7.11. Regulatory status

The Portuguese summaries of product characteristics (SmPC) from INFARMED, the Brazilian package inserts from the National Health Surveillance Agency (ANVISA) and the European public assessment reports for these medicines were not consulted; national brand names and label differences between jurisdictions are not included in this article for that reason.

7.12. Intervention and natural history studies

None of these studies is randomised. The NEI comparison and the recent cohorts compare groups that are self-selected by adherence, and adherent patients may differ from the others in factors that influence progression; reductions in ornithine do not translate linearly into visual outcomes, and the European registry found no association between ornithine and acuity [12, 47]. Percentage reductions in ornithine should not be compared indirectly between studies, because they differ in baseline values, fasting, laboratory and duration. The systematic review mainly includes case reports, subject to publication bias, and series with negative results (visual loss in all Manchester patients; no sustained improvement of CMO in Ankara) should be read alongside the positive ones [19, 32, 40]. The slowness of functional decline measured by Caruso (2001) led the authors to conclude that an efficacy trial on preservation of vision would have to be long, which explains the choice of structural imaging outcomes in GYROS [3, 69]. The 1998 gene therapy trial was the only registered intervention trial; no intervention trial is active on ClinicalTrials.gov in October 2026, according to the search performed, which does not prove that none exists in other databases [69, 77].

7.13. Developments in the last twelve months and research

Publications between November 2025 and October 2026 are mainly observational. In 2026 the following were published: the Turkish Ankara cohort (7 patients, with CMO as the dominant finding), a Japanese case of diagnosis in an adult woman with loss to follow-up, the first description of recurrent macular neovascularisation, a case of rapid improvement of CMO with diet in a 4-year-old child and a case of a pyridoxine-responsive form with CMO (level of evidence: observational studies and case reports) [19, 41, 49, 71, 72]. The Chinese Yunnan cohort, published in 2025, described the first non-coding variant in the 5' region and a possible founder effect in ethnic minorities [42]. A 2026 letter described diagnosis from low citrulline on newborn screening [55]. The 2026 European Delphi consensus classified pyridoxine as an essential product (expert consensus) [64]. The 2026 biochemical review consolidated the molecular mechanisms of the variants and the basis of the response of p.Val332Met (in vitro study) [33]. The 2025–2026 mouse studies characterised the regional variation of the disease and early proteomic and metabolic signatures (preclinical) [38, 39].

Therapeutically, research is at the preclinical stage. Liver-directed gene therapy with an adeno-associated virus serotype 8 (AAV8) vector expressing OAT under a hepatocyte promoter reduced ornithine in blood and in the eye, improved the ERG and partially restored retinal structure up to one year after injection in two mouse models (preclinical, 2023) [78]. In 2025 the same group showed that intravenous administration reduced ornithinaemia, but that only the combined intravenous and subretinal route significantly lowered intraocular ornithine and improved retinal morphology and function (preclinical) [79, 80]. The Amsterdam group advocates local ocular gene therapy, possibly combined with a mild diet or supplements, and estimates horizons of 2–3 years for repurposed medicines, 5–10 years for mutation-specific RNA therapies and more than 10 years for AAV vectors (expert opinion) [34]. The Naples group counters that there is no in vivo proof of concept for retina-only gene therapy, that subretinal injection carries risks (iatrogenic macular detachment, inflammation, ocular hypertension, cataract, tears) and corrects only the area around the injection, and that the hepatic route has its own problems (ectopic periportal expression, theoretical oncogenic risk through polyamine synthesis, neutralising antibodies, vector dilution with liver growth) [37]. In the United States, the GYROS natural history study, funded by the Foundation Fighting Blindness, the FDA and Conquering Gyrate Atrophy, is intended to prepare a trial of single-dose systemic gene therapy developed at Johns Hopkins, still without a registered intervention trial (press release and registry) [69, 76]. The 2023 international symposium on gyrate atrophy listed the open research questions, including the definition of outcomes and of pathogenic metabolites beyond ornithine [32, 81].

7.14. Agents and circumstances to avoid

Arginine-rich foods in patients on diet, excessive arginine restriction and catabolic states due to insufficient caloric intake or anorexia, which raise ornithine, should be avoided [4, 10, 12]. Pyridoxine doses above 900 mg/day in adults are associated with peripheral neuropathy, and patients treated with levodopa without carbidopa should not take supplements containing more than 5 mg of pyridoxine per day [65, 66]. The initial assessment should include retinotoxic medicines in use [10]. High-dose lysine may cause anorexia and, through it, catabolism [12].

7.15. Pregnancy, lactation, perioperative care and transition

The arginine-restricted diet has been maintained during pregnancy in several cases; as in other aminoacidopathies, the requirement for the restricted amino acid increases in the third trimester, and the pregnant woman should be followed closely with weight, nutritional assessment and plasma amino acids [4]. The children of mothers on an unrestricted diet do not appear to have adverse effects from maternal hyperornithinaemia [4]. Pyridoxine is in category A of the FDA label, and requirements increase in pregnancy and lactation [66]. Acetazolamide was teratogenic in animals and should be used in pregnancy only if the benefit justifies it, and dorzolamide caused fetal vertebral malformations in rabbits at high oral doses [73, 74]. A Turkish series on pregnancy in inherited metabolic diseases includes gyrate atrophy, but the text was not accessible [82].

Cataract surgery was uneventful in 11 patients in the Dutch subcohort [11]. Retinal detachment after vitrectomy is described as a rare event [48]. During perioperative fasting or intercurrent illness, caloric intake should be maintained to avoid catabolism, and intercurrent illness is one of the critical times for dietary control [10, 69].

Transition from paediatric to adult care is a period of risk: adherence falls in adolescence, there is loss to follow-up and ornithine rises with age [12, 40, 69]. The Manchester and Amsterdam centres recommend formal transition programmes and multidisciplinary teams [12, 40]. The case in section 10 shows that diagnosis in adulthood, without family support and with language and cultural barriers, can lead to abandonment of treatment [41].

7.16. Recommendations of the sources and divergences

ClinGen considers dietary modification the actionable intervention (strong in adults, moderate in children) and cites pyridoxine as an intervention with inconsistent evidence, creatine without an established dose and proline and lysine with mixed results; no supplement dose is firmly established [10]. Saudubray recommends the pyridoxine trial in all patients, the arginine-restricted diet with an ornithine target below 200 µmol/L and arginine supplementation in the neonatal phase [4]. The 2026 Delphi consensus considers pyridoxine essential, but does not include creatine or lysine among the essential products for this disease [64]. The main divergences concern the proportion of pyridoxine responders (fewer than 10%, about 5%, 30% or 3 of 27, depending on the source), the pyridoxine dose (120–600 mg/day in ClinGen, 200–500 mg/day in Saudubray, 50–500 mg/day in the medicines table of the same book), the ornithine target (below 200 µmol/L in Saudubray, 400–500 µmol/L as the mean associated with benefit at the NEI) and the role of systemic carbonic anhydrase inhibitors in CMO (used in Manchester, considered of limited role in Amsterdam) [4, 5, 10–12, 32, 40, 47, 65]. The French national protocol for urea cycle disorders (HAS, 2021) covers OAT deficiency only in its neonatal form and makes no chronic recommendations for it; it diverges from the other sources on three points: the table of biochemical profiles assigns raised ornithine and normal citrulline to neonatal OAT deficiency, whereas the protocol's own annex and the published series describe low ornithine; the annex links blindness to creatine deficiency, a causal relationship that the other sources do not establish; and the preparation sheet describes the risk of arginine hydrochloride as hyperchloraemic alkalosis, whereas the FDA label and chloride physiology point to hyperchloraemic metabolic acidosis [4, 13, 51, 70]. The doses for undiagnosed hyperammonaemia coincide with those of the European guideline, and the protocol's emergency sheet does not cover OAT deficiency, only OTC and CPS1 deficiencies, citrullinaemia type 1, argininosuccinic aciduria, arginase deficiency and HHH syndrome [50, 51].

8. Reproductive Options and Family Counselling

8.1. Recurrence risks

Inheritance is autosomal recessive [1, 4]. The parents of a patient are, as a rule, heterozygous, and each pregnancy of the couple carries a 25% risk of an affected child, 50% of a heterozygous child and 25% of a child without the variant. Heterozygotes have intermediate enzyme activity and no disease [4, 8, 22]. The children of a patient necessarily receive one variant and are affected only if the other parent is heterozygous; the heterozygote frequency is about 1:110 in Finland, and there are no population estimates outside that country in the sources consulted [10]. Consanguinity increases the risk, and most families in the Manchester, Ankara and Yunnan cohorts were consanguineous or from populations with founder variants [19, 40, 42]. Confirming the phase of the two variants in the parents avoids interpreting two variants on the same allele as compound heterozygosity, and makes it possible to recognise a de novo variant or uniparental disomy, scenarios that the sources consulted do not describe in this disease [41, 42]. Germline mosaicism is also not described in the sources consulted.

8.2. Cascade testing and presymptomatic diagnosis

Testing of siblings is the measure with the greatest impact on prognosis, because it allows the diet to be started before the age of 10 years, an age associated with preservation of the central visual field [11, 12]. In the European registry, 12 of 27 patients were identified by targeted testing after diagnosis in a sibling, and the patient treated from the first year of life had no lesions at 11–12 years [12]. ClinGen bases actionability precisely on the possibility of starting treatment before symptoms, once the familial variants have been identified [10]. In younger siblings, testing is justified by the direct clinical benefit in childhood, and should combine molecular testing with plasma amino acids and ophthalmological assessment; in a newborn sibling, ornithine may be low in the first months and require ammonia monitoring (section 3.6) [4, 12]. Carrier testing in unaffected adult siblings and in partners of patients or carriers can be offered in the context of reproductive planning.

8.3. Prenatal and preimplantation diagnosis

Once the familial variants have been identified, prenatal diagnosis by molecular analysis of a fetal sample, and carrier detection, are possible [4]. Preimplantation genetic testing applies on the same principle, provided the variants are classified as pathogenic or likely pathogenic. A newborn with a prenatal diagnosis should have biochemical surveillance from birth, because hypo-ornithinaemia and hyperammonaemia may develop in the first months even with normal values in cord blood [4].

8.4. Content of counselling

Counselling should explain the inheritance and the recurrence risk, the importance of early diagnosis and diet in siblings, the variability of progression, the chronic nature of treatment and its demands, the possibility of a pyridoxine response and the absence, to date, of gene therapies in clinical trials [10, 12, 69]. The authors of the case in section 10 stress the need to involve the family, to use professional interpreters and cultural mediators when the patient's language is not that of the team, and to adapt dietary education to cultural and religious dietary practices [41]. The AAO recommends genetic counselling after genetic test results in inherited retinal dystrophies [15].

9. Special Considerations: Population Diversity and Epidemiology

9.1. Prevalence and what each figure measures

The available figures are old and based on diagnosed cases, with a mean diagnostic delay of 5 to 7 years and many patients classified only as «retinal dystrophy» [11, 12]. There are no recent population prevalence studies, and the frequency of pathogenic alleles in population databases was not analysed in the sources consulted.

9.2. Geography, ancestry and founder effect

Finland has the highest frequency described, owing to a founder effect: the p.(Leu402Pro) variant accounts for more than 85% of Finnish alleles and has been described only in people of Finnish origin [4]. In the 1989 Finnish analysis, p.(Leu402Pro) was homozygous in 21 of 27 probands, all with the same RFLP haplotype, which indicates a single origin, and explained the disease in 14 of 16 families; p.(Arg180Thr), homozygous in the other two, had a different haplotype from the same allele in a North American proband described as non-Finnish, a sign of independent origin [7]. That North American proband was compound heterozygous p.(Arg180Thr)/p.(Leu402Pro), which diverges from the statement that p.(Leu402Pro) occurs only in Finns; the sources do not clarify his ancestry [4, 7]. In the series of 35 Finnish L402P carriers, examined between 1993 and 1995, severe visual impairment appeared as early as early adulthood and diagnosis was often very late [20]. In the Netherlands, the p.(Gly353Asp) variant was found in all patients of Dutch ancestry studied, and does not respond to pyridoxine [12, 43]. In the Chinese region of Yunnan, a non-coding variant in the 5' region was found in all patients from three ethnic minorities, suggesting a founder effect [42]. The first variant described, at the initiation codon, was found in four Lebanese patients, and cohorts from South Asia, Turkey and the Middle East show frequent homozygosity in consanguineous families (p.(Arg398Ter) and p.(Pro241Leu) in South Asian patients; p.(Arg331Ter), c.424+1G>A and p.(Arg426Ter) in Turkish families; p.(Pro300Arg) in Iraqi siblings) [12, 19, 26, 40]. The initial molecular descriptions also cover Lebanese Maronite, Iraqi Jewish, Indian-origin, Algerian, Nicaraguan, Japanese, Scandinavian-ancestry and English patients [7]. Comparison of Japanese and Finnish patients suggested worse visual function in the former [10].

In the United States, the NEI followed 65 patients between 1978 and 2004, and the GYROS study recruited 46 participants in North American, European and Brazilian centres [24, 69]. In Europe, besides Finland and the Netherlands, patients have been described in the United Kingdom, Belgium, Italy, France and Poland [12, 35, 40, 48, 67]. In Brazil, INRET (Clinic and Research Centre, Belo Horizonte) is one of the GYROS study centres; the sources consulted do not include Brazilian epidemiological data [69]. For Portugal, the OMIM variant catalogue records two Portuguese patients studied at the end of the 1980s: one homozygous for p.(Ala270Pro) and one compound heterozygous for the in-frame deletion p.(Ala184del) and an unidentified second allele [7]. The sources consulted contain no Portuguese epidemiological data, which does not demonstrate the absence of other patients or of Portuguese publications.

9.3. Special populations

In young children, functional tests require adaptation or sedation and visual fields before the age of 7 years are unreliable [10, 15]. In preterm infants, the risk of neonatal hyperammonaemia is greater [4]. In pregnancy, the arginine requirement increases in the third trimester [4]. In patients from cultures and languages different from those of the team, adherence depends on interpreters and on cultural adaptation of the diet [41]. In adults diagnosed late, vision is already compromised and the expected benefit of diet is smaller, which explains the abandonment of treatment in patients over 45 years in the European registry [12].

9.4. Reference networks and patient associations

In Europe, the disease is relevant to two European Reference Networks: ERN-EYE, for rare eye diseases, and MetabERN, for inherited metabolic disorders [16, 17]. Amsterdam UMC is a centre of expertise for gyrate atrophy recognised by the ERN and by the Dutch Ministry of Health and coordinates the international registry with Ghent and Toronto [12]. The Dutch Bird's Eye View consortium, which brings together scientists, clinicians and patient representatives, maintains a portal on the disease [11, 83]. In the United States, Conquering Gyrate Atrophy, a non-profit research and advocacy organisation founded by Karabi Acharya, and the Foundation Fighting Blindness, which also maintains a registry of inherited retinal dystrophies, collaborate in the GYROS study [69, 76, 84]. In France, OAT deficiency with neonatal expression falls within the network for inherited metabolic diseases coordinated by the G2M health filière, with reference centres (the Lille centre coordinated the national protocol for urea cycle disorders, with participation from Necker Hospital) and competence centres; long-term treatment is 100% covered under the long-term condition scheme ALD 17 (inherited metabolic diseases requiring prolonged treatment), and the patient at risk of decompensation carries an emergency certificate written by the reference centre [51].

10. Illustrative Clinical Case

Taniai and colleagues (2026) described a 21-year-old woman, born in South Asia, referred to Jikei University Hospital, in Japan, with reduced visual acuity suggestive of an inherited retinal dystrophy [41]. She had no relevant past history, and parental consanguinity and sibship composition were unknown. At the first examination, best-corrected acuity was 0.4 in the right eye and 0.5 in the left (decimal scale), with refraction of −2.75 and −3.25 dioptres, and slit-lamp biomicroscopy showed posterior subcapsular cataracts without other anterior segment changes [41].

Ultra-wide-field fundus photography showed confluent chorioretinal atrophy in the periphery and around the vascular arcades, with greater extent in the right eye; fundus autofluorescence was hypoautofluorescent in the atrophic areas; OCT revealed multiple intraretinal cystic spaces with preservation of the outer layers, including the ellipsoid zone; on Goldmann perimetry, the I-2 isopter was reduced to less than 5 degrees and the I-4 to 10–20 degrees; and the full-field ERG, according to ISCEV standards, recorded no responses under any dark- or light-adapted condition [41]. Ornithine was markedly raised in serum, 775.8 nmol/mL (reference 31.3–104.7), and in urine, 3353 nmol/mL (reference 0–45.1), and the diagnosis of gyrate atrophy was established on these data [41]. The figures of the paper, whose legends were consulted in the full text, show the fundus photography, autofluorescence and OCT (figure 1), the Goldmann fields (figure 2) and the sequencing visualisation (figure 3), and are available in open access (https://pmc.ncbi.nlm.nih.gov/articles/PMC13125854/).

Exome sequencing identified the homozygous variant NM_000274.4:c.272G>A p.(Gly91Glu) in exon 3 of OAT (NC_000010.11:g.124408893C>T; rs386833603), absent from the population databases gnomAD v2.1 and ToMMo 61KJPN and classified by the authors as likely pathogenic (PS1, PM2, PP3); Sanger sequencing confirmed homozygosity, and the father was heterozygous [41]. In ClinVar, this variant has conflicting classifications [30]. The authors knew of a single previous case with the same genotype, a 10-year-old girl with high myopia, astigmatism and hearing loss [41].

The patient started pyridoxine 100 mg/day for 2 months, and serum ornithine remained at 930.5 nmol/mL. The team planned to increase the dose to 200 mg/day and to start the arginine-restricted diet, but the patient stopped attending appointments and was lost to follow-up [41]. The authors attributed the abandonment to insufficient explanation in a language that was not her mother tongue and without a professional interpreter, to the lack of therapeutic education adapted to cultural and religious dietary practices and to the absence of communication with her parents, in a patient who received the information about the disease on her own [41].

The case illustrates late diagnosis, already with an extinguished ERG and a markedly reduced field, cystoid maculopathy with a preserved ellipsoid zone and the difficulty of maintaining demanding treatment in an adult without family support. Management did not fully follow the reference recommendations: the pyridoxine trial used 100 mg/day, below the 300–500 mg/day recommended in Saudubray and the at least 200–400 mg/day used in Manchester, so the lack of response at this dose does not allow the patient to be classified as a non-responder [4, 40]. The contrast with the siblings in the European registry, in which the younger, treated from the first year of life, had a normal retina at 11 years and the older sister, with poor adherence in the first decade, had peripheral atrophy at 25 years (figures 1A and 1B of Balfoort 2025, known from the legends), shows what is lost when diagnosis and treatment come late [12].

11. Prognosis and Long-Term Follow-up

11.1. Prognosis

Without treatment, degeneration progresses throughout life, with blindness in middle adulthood: between 40 and 55 years according to ClinGen, between 45 and 65 years according to Saudubray, in the fourth to seventh decade in the NEI sibling series [4, 10, 46]. The 2021 French protocol reproduces the 45–65-year range and attributes blindness to creatine deficiency; in the series in which creatine was supplemented, retinal degeneration continued (section 7.5) [10, 51, 53]. Central vision is preserved for decades, but is compromised by early cataract, cystoid maculopathy, macular neovascularisation and epiretinal membrane [5]. In the NEI series of five patients followed for 4 to 6 years, median half-lives were 11 to 17 years for the visual field and 11 to 16 years for the ERG, reflecting slow functional decline [3]. Acuity worsens by about 0.013 logMAR per year, and the central 10 degrees of the field were intact only in patients under 30 years in the Amsterdam subcohort [11, 12]. Pyridoxine responders have a milder course [4]. Diet started in the first decade is associated with better preservation of the central field, but no cohort has shown prevention of severe visual impairment, and visual loss progressed in all Manchester patients [12, 40]. The effect of prolonged early treatment on vision at older ages has not been evaluated [4]. Patients and relatives consider visual loss the manifestation with the greatest impact [34].

The extraocular prognosis is mostly favourable: cognition is normal in most, neuromuscular involvement is mainly subclinical, and treated neonatal hyperammonaemia left no sequelae in the two siblings in the European registry [4, 12, 21]. Sarcopenia and reduced lean mass are frequent in adults, and cerebral creatine remains low in many supplemented patients [12]. The sources consulted do not describe an effect of the disease on life expectancy.

11.2. Surveillance

There are no standardised imaging surveillance guidelines for this disease, and the intervals derive from ClinGen, the AAO and centre practice [10, 15, 40].

The historical NEI protocol, with an almost protein-free diet, used monthly tests and complete eye examination every 6–12 months [24]; the GYROS study uses fasting dried blood spot ornithine every 4 months and annual imaging, plasma ornithine and perimetry, in a research context [69]. In Manchester, retinal imaging was irregular (autofluorescence and wide-field with an annual median; OCT with a median of four examinations in 5 years), and the authors recommend a formal imaging programme [40]. The Amsterdam group considers acuity a poor follow-up outcome and the visual field preferable, but the methods (Goldmann or Humphrey) are not harmonised [12].

12. Summary and Final Considerations

Gyrate atrophy of the choroid and retina is a rare autosomal recessive disease, caused by biallelic variants in OAT, which encodes a pyridoxal phosphate-dependent mitochondrial aminotransferase. Owing to a physiological reversal of enzymatic flux, the disease has two phases: in the newborn, lack of ornithine synthesis can cause hyperammonaemia with low ornithine, citrulline and arginine, which mimics OTC deficiency and responds to arginine; after the first months, ornithine accumulates 5 to 20 times above normal, without hyperammonaemia. The typical presentation is a child with high, progressive myopia and night blindness, with circular patches of chorioretinal atrophy in the mid-periphery that coalesce and advance towards the posterior pole, posterior subcapsular cataract before the age of 30 years and cystoid maculopathy; blindness occurs in middle adulthood, with estimates that vary between sources. Extraocular involvement includes secondary creatine deficiency, type II muscle fibre atrophy, subclinical peripheral neuropathy and, in a minority, neurodevelopmental disorders.

Diagnosis rests on plasma amino acids and sequencing of OAT; homocitrullinuria distinguishes HHH syndrome, and normal ornithine distinguishes chorioretinal dystrophies such as choroideremia. Newborn screening by ornithine is not reliable, and low citrulline or a raised proline/citrulline ratio are neonatal markers under study. All patients should undergo a pyridoxine trial before any dietary change, although a response is rare and variant-dependent. The diet restricted in arginine or natural protein, with an essential amino acid mixture, slows visual loss in observational studies, especially when started in the first decade, but does not prevent it and requires an adherence that few patients maintain. Creatine corrects part of the muscle and cerebral deficit without a demonstrated retinal effect, and lysine has a modest effect and poor tolerance. Ocular complications are treated with the therapies used in other indications, with doses extrapolated from labels. Testing of siblings and early start of treatment are the measures with the greatest impact. Hepatic, ocular or combined gene therapy is at the preclinical stage, and the GYROS natural history study is preparing the outcomes for future trials.

13. References

1. Orphanet. Gyrate atrophy of choroid and retina. ORPHA:414. Orphadata rd-epidemiology/rd-natural_history/rd-cross-referencing (version 2026-06-23). Available from: https://www.orpha.net/en/disease/detail/414
2. GeneGuidance. OAT — ornithine aminotransferase; gyrate atrophy of choroid and retina (accessed 2026-09-26).
3. OMIM® Online Mendelian Inheritance in Man. #258870 GYRATE ATROPHY OF CHOROID AND RETINA; GACR. Johns Hopkins University; last edited 23/06/2022. https://omim.org/entry/258870. Accessed 28 October 2026.
4. Baumgartner MR, Valle D, Dionisi-Vici C. Disorders of ornithine and proline metabolism. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 21, sections 21.1 and 21.2. doi:10.1007/978-3-662-63123-2_21
5. Buijs MJN, Balfoort BM, Brands MM, Ten Asbroek ALMA, Boon CJF, Diederen RMH, et al. Molecular and cellular mechanisms underlying gyrate atrophy: Why is the retina primarily affected?. Acta Ophthalmol. 2025;103(7):e436-e455. doi:10.1111/aos.17498.
6. 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. Available from: https://variantvalidator.org/
7. OMIM® Online Mendelian Inheritance in Man. *613349 ORNITHINE AMINOTRANSFERASE; OAT. Johns Hopkins University; last edited 24/09/2022. https://omim.org/entry/613349. Accessed 28 October 2026.
8. Valle D, Kaiser-Kupfer MI, Del Valle LA. Gyrate atrophy of the choroid and retina: deficiency of ornithine aminotransferase in transformed lymphocytes. Proc Natl Acad Sci U S A. 1977;74(11):5159-61. doi:10.1073/pnas.74.11.5159.
9. Camacho J, Rioseco-Camacho N. Hyperornithinemia-hyperammonemia-homocitrullinuria syndrome. 2012 May 31 [updated 2020 Feb 13; revised 2025 Apr 10]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews®. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK97260/
10. ClinGen Actionability Working Group. Gyrate atrophy (OAT) — Adult Actionability Summary Report AC1014, version 1.0.2 (literature search 2020-06-29); Pediatric Summary Report version 1.0.3. Available from: https://actionability.clinicalgenome.org/ac/Adult/ui/stg2SummaryRpt?doc=AC1014
11. Balfoort BM, Van Den Broeck F, Brands MM, van Karnebeek CD, Bergen AA, van den Born LI, et al. A cohort study of 19 patients with gyrate atrophy of the choroid and retina (GACR). Graefes Arch Clin Exp Ophthalmol. 2024;262(11):3589-3596. doi:10.1007/s00417-024-06540-8.
12. Balfoort BM, Van den Broeck F, Boon CJF, Brouwers MCGJ, Diederen RMH, Dhillon P, et al. Novel Insights Into Gyrate Atrophy of the Choroid and Retina (GACR): A Cohort Study. J Inherit Metab Dis. 2025;48(1):e12842. doi:10.1002/jimd.12842.
13. Cleary MA, Dorland L, de Koning TJ, Poll-The BT, Duran M, Mandell R, et al. Ornithine aminotransferase deficiency: diagnostic difficulties in neonatal presentation. J Inherit Metab Dis. 2005;28(5):673-9. doi:10.1007/s10545-005-0074-1.
14. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 302–303.
15. American Academy of Ophthalmology Quality of Care Secretariat, Hoskins Center for Quality Eye Care. Guidelines on Clinical Assessment of Patients with Inherited Retinal Degenerations – 2022. San Francisco: AAO; 2022 Oct. Available from: https://www.aao.org/clinical-statement/guidelines-on-clinical-assessment-of-patients-with
16. European Reference Network for Rare Eye Diseases (ERN-EYE). Available from: https://www.ern-eye.eu/ [accessed 2026 Oct 28].
17. European Reference Network for Hereditary Metabolic Disorders (MetabERN). Available from: https://metab.ern-net.eu/ [accessed 2026 Oct 28].
18. Simell O, Takki K. Raised plasma-ornithine and gyrate atrophy of the choroid and retina. Lancet. 1973;1(7811):1031-3. doi:10.1016/s0140-6736(73)90667-3.
19. Aktaşoğlu EÖ, Özdemir HB, Özmen MC, İnci A, Okur İ, Ezgü FS, et al. Gyrate atrophy of the choroid and retina: a tertiary center experience. Orphanet J Rare Dis. 2026;21(1):104. doi:10.1186/s13023-026-04265-x.
20. Peltola KE, Näntö-Salonen K, Heinonen OJ, Jääskeläinen S, Heinänen K, Simell O, et al. Ophthalmologic heterogeneity in subjects with gyrate atrophy of choroid and retina harboring the L402P mutation of ornithine aminotransferase. Ophthalmology. 2001;108(4):721-9. doi:10.1016/s0161-6420(00)00587-x.
21. Peltola KE, Jääskeläinen S, Heinonen OJ, Falck B, Näntö-Salonen K, Heinänen K, et al. Peripheral nervous system in gyrate atrophy of the choroid and retina with hyperornithinemia. Neurology. 2002;59(5):735-40. doi:10.1212/wnl.59.5.735.
22. O'Donnell JJ, Sandman RP, Martin SR. Gyrate atrophy of the retina: inborn error of L-ornithin:2-oxoacid aminotransferase. Science. 1978;200(4338):200-1. doi:10.1126/science.635581.
23. Kaiser-Kupfer MI, de Monasterio FM, Valle D, Walser M, Brusilow S. Gyrate atrophy of the choroid and retina: improved visual function following reduction of plasma ornithine by diet. Science. 1980;210(4474):1128-31. doi:10.1126/science.7444439.
24. National Eye Institute. The diagnosis, pathogenesis and treatment of gyrate atrophy of the choroid and retina. ClinicalTrials.gov identifier NCT00001166. Available from: https://clinicaltrials.gov/study/NCT00001166
25. Inana G, Totsuka S, Redmond M, Dougherty T, Nagle J, Shiono T, et al. Molecular cloning of human ornithine aminotransferase mRNA. Proc Natl Acad Sci U S A. 1986;83(5):1203-7. doi:10.1073/pnas.83.5.1203.
26. Mitchell GA, Brody LC, Looney J, Steel G, Suchanek M, Dowling C, et al. An initiator codon mutation in ornithine-delta-aminotransferase causing gyrate atrophy of the choroid and retina. J Clin Invest. 1988;81(2):630-3. doi:10.1172/JCI113365.
27. Ramesh V, McClatchey AI, Ramesh N, Benoit LA, Berson EL, Shih VE, et al. Molecular basis of ornithine aminotransferase deficiency in B-6-responsive and -nonresponsive forms of gyrate atrophy. Proc Natl Acad Sci U S A. 1988;85(11):3777-80. doi:10.1073/pnas.85.11.3777.
28. Brody LC, Mitchell GA, Obie C, Michaud J, Steel G, Fontaine G, et al. Ornithine delta-aminotransferase mutations in gyrate atrophy. Allelic heterogeneity and functional consequences. J Biol Chem. 1992;267(5):3302-7.
29. Wang T, Lawler AM, Steel G, Sipila I, Milam AH, Valle D. Mice lacking ornithine-delta-aminotransferase have paradoxical neonatal hypoornithinaemia and retinal degeneration. Nat Genet. 1995;11(2):185-90. doi:10.1038/ng1095-185.
30. National Center for Biotechnology Information. ClinVar [Internet]. Bethesda (MD): NCBI. Available from: https://www.ncbi.nlm.nih.gov/clinvar/ (OAT; IDs 153, 156, 162, 168, 177, 181, 56121, 56131, 56138, 56141, 456519)
31. Nyhan WL, Hoffmann GF, Al-Aqeel AI, Barshop BA. Atlas of Inherited Metabolic Diseases. 4th ed. Boca Raton: CRC Press; 2020. p. 247–252 (chap. 31, Hyperornithinemia, hyperammonemia, homocitrullinuria syndrome).
32. Balfoort BM, Buijs MJN, Ten Asbroek ALMA, Bergen AAB, Boon CJF, Ferreira EA, et al. A review of treatment modalities in gyrate atrophy of the choroid and retina (GACR). Mol Genet Metab. 2021;134(1-2):96-116. doi:10.1016/j.ymgme.2021.07.010.
33. Floriani F, Borri Voltattorni C, Montioli R. Current Insight into Human Ornithine Aminotransferase: A Review. Proteins. 2026;94(8):1431-1441. doi:10.1002/prot.70134.
34. Bergen AA, Buijs MJ, Ten Asbroek AL, Balfoort BM, Boon CJ, Brands MM, et al. Vision on gyrate atrophy: why treat the eye?. EMBO Mol Med. 2024;16(1):4-7. doi:10.1038/s44321-023-00001-1.
35. Valayannopoulos V, Boddaert N, Mention K, Touati G, Barbier V, Chabli A, et al. Secondary creatine deficiency in ornithine delta-aminotransferase deficiency. Mol Genet Metab. 2009;97(2):109-13. doi:10.1016/j.ymgme.2008.12.010.
36. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 9 (introduction).
37. Boffa I, Brunetti-Pierri N. Vision on gyrate atrophy: why treat the liver?. EMBO Mol Med. 2024;16(1):8-9. doi:10.1038/s44321-023-00002-0.
38. Wilder RJ, An AF, Bell BA, Fossett G, Wojciechowski AM, Shpylchak I, et al. Intraretinal variation in disease severity in the Oat(rhg) mouse model of gyrate atrophy. Exp Eye Res. 2025;255:110382. doi:10.1016/j.exer.2025.110382.
39. Puja A, Xu R, Mascari I, Ngo T, Zhang Y, Wang Q, et al. Early proteomic and metabolic signatures of liver and eye in OAT-deficient mice. Exp Eye Res. 2026;271:111140. doi:10.1016/j.exer.2026.111140.
40. Palmer E, Stepien KM, Campbell C, Barton S, Iosifidis C, Ghosh A, et al. Clinical, biochemical and molecular analysis in a cohort of individuals with gyrate atrophy. Orphanet J Rare Dis. 2023;18(1):265. doi:10.1186/s13023-023-02840-0.
41. Taniai H, Mizobuchi K, Hara T, Hayashi T, Nakano T. Gyrate Atrophy of the Choroid and Retina Diagnosed in Adulthood With a Homozygous OAT Variant: A Case Report Highlighting the Need for Long-Term Care. Case Rep Ophthalmol Med. 2026;2026:8380205. doi:10.1155/crop/8380205.
42. Gao FJ, Duan C, Chen KX, Ju YQ, Chang Q, Li JJ, et al. Expanding the genetic spectra of gyrate atrophy of the choroid and retina in a Chinese cohort in Yunnan province. Hum Genomics. 2025;19(1):145. doi:10.1186/s40246-025-00857-z.
43. Balfoort BM, Pampalone G, Ruiter JPN, Denis SW, Brands MM, Timmer C, et al. Extending diagnostic practices in gyrate atrophy: Enzymatic characterization and the development of an in vitro pyridoxine responsiveness assay. Mol Genet Metab. 2024;143(1-2):108542. doi:10.1016/j.ymgme.2024.108542.
44. Díazceballos-García AL, Matsui R, Chairez Miranda MG, Rosales Padrón JF, Graue-Wiechers F, Zenteno JC. Multimodal imaging and genetic screening in Mexican patients with Gyrate atrophy: identification of novel OAT pathogenic variants. Int Ophthalmol. 2024;45(1):1. doi:10.1007/s10792-024-03260-0.
45. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 29, sections 29 (table) and 29.2.7.
46. Kaiser-Kupfer MI, Caruso RC, Valle D. Gyrate atrophy of the choroid and retina. Long-term reduction of ornithine slows retinal degeneration. Arch Ophthalmol. 1991;109(11):1539-48. doi:10.1001/archopht.1991.01080110075039.
47. Kaiser-Kupfer MI, Caruso RC, Valle D, Reed GF. Use of an arginine-restricted diet to slow progression of visual loss in patients with gyrate atrophy. Arch Ophthalmol. 2004;122(7):982-4. doi:10.1001/archopht.122.7.982.
48. Fong N, Lee YM, Agrawal S, Durkin S. Rhegmatogenous retinal detachment in gyrate atrophy: a case report and review of literature. Doc Ophthalmol. 2026;152(1):79-84. doi:10.1007/s10633-025-10073-8.
49. Bierbrier J, Ahmed Y, Ventresca M, Arshinoff SA, Koushan K. Macular neovascularization in patients with Gyrate atrophy. Ophthalmic Genet. 2026:1-7 (early online publication). doi:10.1080/13816810.2026.2671407.
50. Häberle J, Boddaert N, Burlina A, Chakrapani A, Dixon M, Huemer M, et al. Suggested guidelines for the diagnosis and management of urea cycle disorders. Orphanet J Rare Dis. 2012;7:32. doi:10.1186/1750-1172-7-32.
51. Haute Autorité de Santé. Protocole National de Diagnostic et de Soins (PNDS): Déficits du cycle de l’urée. Saint-Denis La Plaine: HAS; 2021 jun 3. Available from: https://www.has-sante.fr/jcms/p_3269570.
52. Valtonen M, Näntö-Salonen K, Jääskeläinen S, Heinänen K, Alanen A, Heinonen OJ, et al. Central nervous system involvement in gyrate atrophy of the choroid and retina with hyperornithinaemia. J Inherit Metab Dis. 1999;22(8):855-66. doi:10.1023/a:1005602405349.
53. Heinänen K, Näntö-Salonen K, Komu M, Erkintalo M, Alanen A, Heinonen OJ, et al. Creatine corrects muscle 31P spectrum in gyrate atrophy with hyperornithinaemia. Eur J Clin Invest. 1999;29(12):1060-5. doi:10.1046/j.1365-2362.1999.00569.x.
54. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 1, sections 1.4.16, 1.5, 1.6.2, 1.6.11.
55. Summerlin M, Harding C, Starosta RT, Sliwoski L, Viall S, Kripps K. Letter to the editor: Newborn screening for low blood citrulline leads to diagnosis of ornithine aminotransferase deficiency. Mol Genet Metab. 2026;148(3):110161. doi:10.1016/j.ymgme.2026.110161.
56. Valle D, Simell O. The hyperornithinemias. In: Valle DL, Antonarakis S, Ballabio A, Beaudet AL, Mitchell GA, editors. The Online Metabolic and Molecular Bases of Inherited Disease. McGraw-Hill; 2019. https://ommbid.mhmedical.com/content.aspx?bookid=2709&sectionid=225083672
57. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 3, section 3.1.
58. Orphanet. Hyperornithinemia-hyperammonemia-homocitrullinuria syndrome. ORPHA:415. Available from: https://www.orpha.net/en/disease/detail/415
59. Yang-Seeger D, Pauleikhoff LJB, Atiskova Y, Thiele S, Spitzer MS, Birtel J. C1QTNF5 missense variant causing autosomal dominant gyrate atrophy-like choroidal dystrophy. Ophthalmic Genet. 2025;46(4):397-400. doi:10.1080/13816810.2025.2492037.
60. Stanton CM, Ansari G, Pfau K, Lipsky T, Halachev M, Gerasimavicius L, et al. Widening the spectrum of disease expression due to heterozygous variants in EFEMP1. JAMA Ophthalmol. 2026;144(10):915-924. doi:10.1001/jamaophthalmol.2026.3828.
61. Wanders RJA, Engelen M, Vaz FM. Inborn errors of non-mitochondrial fatty acid metabolism including peroxisomal disorders. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 42, section 42.3. doi:10.1007/978-3-662-63123-2_42.
62. Ramaswami U, Humphries SE. Inborn errors of lipoprotein metabolism presenting in childhood. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 36, section 36.1. doi:10.1007/978-3-662-63123-2_36.
63. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 74–77.
64. 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.
65. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 45, Medications used in the treatment of inborn errors of metabolism (rows Pyridoxine, L-Lysine-HCl, Creatine monohydrate).
66. Fresenius Kabi USA. Pyridoxine Hydrochloride Injection, USP [prescribing information]. ANDA080618. Revised 2024-10-15. DailyMed setid a56d11c0-b033-4201-85ff-fc710506481a. Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=a56d11c0-b033-4201-85ff-fc710506481a
67. Santinelli R, Costagliola C, Tolone C, D'Aloia A, D'Avanzo A, Prisco F, et al. Low-protein diet and progression of retinal degeneration in gyrate atrophy of the choroid and retina: a twenty-six-year follow-up. J Inherit Metab Dis. 2004;27(2):187-96. doi:10.1023/B:BOLI.0000028779.29966.05.
68. Kaiser-Kupfer MI, Caruso RC, Valle D. Gyrate atrophy of the choroid and retina: further experience with long-term reduction of ornithine levels in children. Arch Ophthalmol. 2002;120(2):146-53. doi:10.1001/archopht.120.2.146.
69. Jaeb Center for Health Research. Gyrate Atrophy Ocular and Systemic Study (GYROS). ClinicalTrials.gov identifier NCT05312736. Last update posted 2026-10-27. Available from: https://clinicaltrials.gov/study/NCT05312736
70. Pharmacia & Upjohn. R-Gene 10 (arginine hydrochloride injection, USP) [prescribing information]. NDA016931. Revised 2024-04-02. DailyMed setid 6accf33f-9841-4327-8923-93c42020ef12. Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=6accf33f-9841-4327-8923-93c42020ef12
71. Bates B, Bilberry S, Gibson J, Hoang L, Armitage Harper C 3rd. Rapid improvement of cystoid macular edema in a patient with gyrate atrophy. J AAPOS. 2026;30(4):104909. doi:10.1016/j.jaapos.2026.104909.
72. Jagadish R, Giri N, Laturkar GR, Grover S, Trapasiya JK. Case report on Vitamin B6-Responsive Gyrate atrophy with Cystoid Macular Edema in a Consanguineous Young Male: A Treatable Metabolic Retinopathy. Ann Afr Med. 2026;25(Suppl 3):S494-S497. doi:10.4103/aam.aam_218_26.
73. Avet Pharmaceuticals. Acetazolamide tablets, USP [prescribing information]. ANDA205530. Revised 2026-05-12. DailyMed setid 0700a910-79aa-4431-aecf-637fe14fcae8. Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=0700a910-79aa-4431-aecf-637fe14fcae8
74. Sagent Pharmaceuticals. Dorzolamide hydrochloride ophthalmic solution [prescribing information]. ANDA215660. Revised 2026-01-09. DailyMed setid 0246463c-6c24-4611-8e22-2d5254ace66a. Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=0246463c-6c24-4611-8e22-2d5254ace66a
75. Harrow Eye. NEVANAC (nepafenac ophthalmic suspension) 0.1% [prescribing information]. NDA021862. Revised 2024-11-05. DailyMed setid 11b9a0cf-2a70-17ef-e063-6394a90a5d9e. Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=11b9a0cf-2a70-17ef-e063-6394a90a5d9e
76. Foundation Fighting Blindness. Foundation Fighting Blindness launches GYROS, a natural history study for gyrate atrophy [press release]. Columbia (MD): FFB; 2024 Feb 6. Available from: https://www.fightingblindness.org/news/gyros-study-425
77. National Eye Institute. Phase I study in the safety and efficacy of transduced keratinocytes for possible treatment of gyrate atrophy. ClinicalTrials.gov identifier NCT00001735. Available from: https://clinicaltrials.gov/study/NCT00001735
78. Boffa I, Polishchuk E, De Stefano L, Dell'Aquila F, Nusco E, Marrocco E, et al. Liver-directed gene therapy for ornithine aminotransferase deficiency. EMBO Mol Med. 2023;15(4):e17033. doi:10.15252/emmm.202217033.
79. Dell'Aquila F, Di Cunto R, Marrocco E, Del Prete E, D'Alessio A, De Stefano L, et al. Combined intraocular and intravenous gene delivery for therapy of gyrate atrophy of the choroid and retina. Mol Ther. 2025;33(7):2997-3005. doi:10.1016/j.ymthe.2024.10.019.
80. Chekuri A. Combinatorial gene therapy for gyrate atrophy of the choroid and retina. Mol Ther. 2025;33(7):2955-2956. doi:10.1016/j.ymthe.2025.03.063.
81. Brands M, Balfoort B, Acharya K, Bergen A, Brunetti-Pierri N, Buijs M, et al. A mini-review on the international gyrate atrophy symposium 2023: More than meets the eye. Focus on outstanding research questions. Mol Genet Metab. 2024;143(4):108609. doi:10.1016/j.ymgme.2024.108609.
82. İşler-Soylu E, Zubarioglu T, Uygur E, Aras-Çöl T, Babazade-Aghakishili H, Çalışgan K, et al. Managing Pregnancy in Inherited Metabolic Disorders: Experience From a Single Tertiary Metabolic Center. J Obstet Gynaecol Res. 2026;52(8):e70419. doi:10.1111/jog.70419.
83. Bird's Eye View Consortium. Gyrate atrophy of the choroid and retina [portal of the Dutch consortium]. Amsterdam: Amsterdam UMC. Available from: https://gyrateatrophy.nl/ [accessed 2026 Oct 28].
84. Conquering Gyrate Atrophy [patient organisation]. Available from: https://www.conqueringgyrateatrophy.org/ [accessed 2026 Oct 28].