Severe Methylenetetrahydrofolate Reductase (MTHFR) Deficiency

Autosomal recessive homocystinuria caused by loss of function of 5,10-methylenetetrahydrofolate reductase, with severe hyperhomocysteinaemia and low or normal methionine, neonatal encephalopathy or late-onset neurological and psychiatric disease, treatable with betaine; an entity distinct from the common polymorphisms c.665C>T and c.1286A>C, which have no clinical utility as a thrombophilia test or in the prevention of neural tube defects.

Literature and regulatory information updated: 17 October 2026.

0. Reference Summary Table

1. Introduction

1.1. Initial description and historical milestones

The first biochemical description of homocystinuria associated with decreased methylenetetrahydrofolate reductase activity was published in 1972 by Mudd, Uhlendorf, Freeman, Finkelstein, and Shih, in the context of investigating homocystinurias that did not correspond to cystathionine β-synthase (CBS) deficiency [22]. In 1975, Freeman, Finkelstein, and Mudd described in the New England Journal of Medicine an adolescent girl with mild intellectual disability, homocystinuria, and homocystinaemia without hypermethioninaemia, lacking the body habitus of CBS deficiency, with recurrent episodes of schizophrenia-like behaviour that responded to folate; the pathway enzymes were normal except for MTHFR, which had 18% of control activity [23]. This case established two features that still guide clinical recognition: homocystinuria with low or normal methionine, and the possibility of an isolated psychiatric presentation [23, 24].

In 1994, Goyette and colleagues isolated the human MTHFR complementary DNA by homology with the porcine enzyme and the bacterial metF gene, mapped the gene to 1p36, and identified the first two variants in patients, an Arg→Gln missense variant at a conserved residue and an Arg→Ter nonsense variant [25]. A year later, Frosst and colleagues identified the common thermolabile variant, present on about 38% of unselected chromosomes in the sample studied, associated with reduced activity and thermolability in lymphocytes and with higher plasma homocysteine in homozygotes, and proposed it as a vascular risk factor [26]. This second discovery generated an extensive literature on risk associations that bears no relation to the severe Mendelian disease and that, as discussed below, does not support clinical genotyping [18].

The following period brought the demonstration that early treatment alters the natural history. In 2007, Strauss and colleagues described five Amish children homozygous for the same variant, in whom the four oldest had irreversible brain injury before diagnosis, whereas the youngest, treated with betaine from birth, was healthy at age 3 years [9]. The case meta-analysis by Diekman and colleagues (2014) associated early betaine with survival and normal psychomotor development [27]. The European network E-HOD (European network and registry for homocystinurias and methylation defects), launched in February 2013, published the first guideline for the diagnosis and treatment of remethylation disorders, including MTHFR deficiency, in 2017 [17], and its first revision in 2026, which is now the reference guideline [2]. The multicenter E-HOD cohort of Yverneau and colleagues (2022) showed that presymptomatic diagnosis is the independent factor most strongly associated with good neurodevelopmental outcome [10].

1.2. Nomenclature and definitions

The term "MTHFR deficiency" is used ambiguously: in the metabolic literature it denotes severe homocystinuria caused by biallelic loss-of-function variants, but it is also applied, mostly outside clinical genetics, to people with common genotypes that slightly reduce enzyme activity and who do not have homocystinuria [1]. For this reason, GeneReviews prefers the name "homocystinuria due to N(5,10)-methylenetetrahydrofolate reductase deficiency" [1], and the E-HOD guideline uses "severe MTHFR deficiency" [2]. This article follows the latter name, reserving the term "polymorphisms" for the common variants c.665C>T p.(Ala222Val) and c.1286A>C p.(Glu429Ala), historically designated "C677T" and "A1298C" [18, 28].

Severe MTHFR deficiency is defined by the combination of hyperhomocysteinaemia, low or normal plasma methionine, absence of methylmalonic aciduria, and biallelic pathogenic or likely pathogenic variants in MTHFR, or severely reduced enzyme activity in fibroblasts when genetic testing is inconclusive [1, 2]. The 2026 guideline classifies it among the isolated remethylation disorders (iRD), because the defect lies in the substrate for methionine synthase, 5-methyltetrahydrofolate, rather than in cobalamin [2].

There is no single age cutoff separating early-onset from late-onset forms. The study by Froese and colleagues uses the first year of life as the boundary (early <1 year; late >1 year) [7]; the Yverneau cohort defines early onset as diagnosis by 3 months of age [10]; and the French and E-HOD adult series of Marelli includes patients diagnosed after age 10 years [11]. The 2026 guideline describes the manifestations separately for early and late onset, without setting a numerical threshold in its clinical table [2]. The cohort data presented later always state the definition used by each study.

1.3. Scope and related entities

This article addresses severe MTHFR deficiency as a Mendelian disease. Classic homocystinuria due to CBS deficiency, with elevated methionine, ectopia lentis, and a skeletal phenotype, is part of the differential diagnosis. Common MTHFR polymorphisms are related to neural tube defects and to folic acid prevention, and have no value in thrombophilia evaluation. Sections 2.7, 6.3, and 8.6 summarise the position of the American College of Medical Genetics and Genomics (ACMG) on these polymorphisms, because confusion between the two entities is common in practice and has direct consequences for counselling [18, 19].

2. Molecular Genetics and Pathophysiology

2.1. Gene and protein

The MTHFR gene maps to 1p36.22, has 11 exons, and encodes a protein of 656 amino acids (NP_005948.3) [4, 5]. The MANE Select reference transcript is NM_005957.5, which corresponds to ENST00000376590.9 [4]. The enzyme functions as a homodimer; each monomer has an N-terminal catalytic domain, with binding sites for NADPH, 5,10-methylenetetrahydrofolate, and the cofactor FAD, linked to a C-terminal regulatory domain, where allosteric inhibition by S-adenosylmethionine (SAM, also called AdoMet) takes place [5, 29]. The reaction catalysed is the irreversible, NADPH-dependent reduction of 5,10-methylenetetrahydrofolate to 5-methyltetrahydrofolate (5-MTHF), the main circulating form of folate and the methyl donor for methionine synthase in the remethylation of homocysteine to methionine [1, 12].

2.2. Biochemical mechanism and pathophysiology

In severe MTHFR deficiency, the lack of 5-MTHF limits remethylation of homocysteine by methionine synthase, which raises plasma total homocysteine and lowers methionine; total folate in the CSF is severely reduced [12]. Unlike methionine synthase defects (cblE and cblG), the block lies upstream of methyl-THF formation and does not cause "trapping" of folates as methyl-THF (folate trap), so it does not interfere with the availability of reduced folates for purine and pyrimidine synthesis [12, 17]. This mechanism explains why megaloblastic anaemia is not part of the typical picture, in contrast to intracellular cobalamin defects and B12 or folate deficiency [12, 24].

Pathophysiological hypotheses include toxicity of homocysteine and related metabolites, lack of products (methionine and SAM), decreased SAM-dependent methylation capacity (DNA methylation, creatine and myelin synthesis), oxidative stress, and nonenzymatic functions of the protein [17]. Saudubray emphasises that it is not established whether the neuropathology results from elevated homocysteine, from decreased methionine interfering with methylation reactions, or from some other metabolic effect [12]. Low methionine and SAM are associated with demyelination, and accumulation of S-adenosylhomocysteine (SAH, or AdoHcy) may contribute to hypomyelination [10].

The particular vulnerability of the central nervous system has an anatomical basis: in the brain, homocysteine remethylation depends exclusively on 5-MTHF, because betaine-homocysteine S-methyltransferase (BHMT), the alternative pathway that uses betaine as a methyl donor, is not expressed in brain tissue [10, 12]. BHMT is active mainly in the liver; orally administered betaine lowers homocysteine and increases systemic methionine, and its effects on the central nervous system are mediated by circulating methionine and by metabolites that cross the blood–brain barrier [12]. This is the rationale for treatment, described in section 7, and also the explanation for persistently low CSF 5-MTHF on betaine: in the Amish series, betaine at doses of 534 ± 222 mg/kg/day increased plasma SAM and tripled the calculated cerebral methionine influx, but did not normalise tHcy or correct the cerebral 5-MTHF deficiency [9].

2.3. Inheritance pattern

Transmission is autosomal recessive; the disease requires loss-of-function variants on both alleles, in the homozygous or compound heterozygous state in trans [1]. Heterozygotes are asymptomatic and are not at risk of developing the disease [1]. Consanguinity is frequent in published series: 13 of 33 families in the 2016 Huemer cohort, and 7 of 10 patients with available information in the Yverneau early-onset cohort [6, 10].

2.4. Variant spectrum

The mutational review by Froese and colleagues (2016) compiled 109 pathogenic variants in 171 families: 70 missense, 17 with a predominant effect on splicing, 11 nonsense, 7 small deletions, 2 no-stop, 1 small duplication, and 1 large duplication; only 36% of variants recur in unrelated families, meaning that most are private [7]. In the study by Burda and colleagues (2015), with 76 patients and fibroblast enzymology in 72, a deleterious variant was identified on all 152 alleles, for a total of 65 different variants, 42 of them novel [29]. Missense variants with virtually absent activity cluster exclusively in the N-terminal catalytic domain, whereas those in the C-terminal regulatory domain decrease NADPH binding and alter SAM inhibition [29]. Among the 42 cell lines with residual activity (1.7–42% of control), 28 had reduced affinity for NADPH, 1 for methylenetetrahydrofolate, 5 were FAD-responsive, and 24 had abnormal SAM inhibition kinetics [29].

Large intragenic deletions and duplications are rare: GeneReviews, based on HGMD data, attributes about 99% of pathogenic variants to sequencing and about 1% to deletions or duplications, and the Froese study records one large duplication [1, 7]. For large deletions and duplications, GeneReviews cites a 2014 article by Ibn-Salem that deals with deletions of regulatory boundaries associated with congenital disease and not with MTHFR in particular [30].

Much of the MTHFR literature uses a historical complementary DNA numbering shifted by 12 nucleotides relative to transcript NM_005957.5; amino acid numbers coincide. On validation with VariantValidator, the historical descriptions "c.677C>T" and "c.1141C>T" are rejected on NM_005957.5 because of a mismatch with the reference base, and the correct equivalents are c.665C>T and c.1129C>T [28]. Table 2.1 lists the variants cited in the article, with current numbering, historical designation, and ClinVar classification.

Table 2.1. Normalisation of the MTHFR variants cited (GRCh38; NM_005957.5, NP_005948.3)

The remaining Huemer 2016 variants (for example c.188G>C p.Trp59Ser, c.482G>A p.Arg157Gln, c.596C>T p.Ala195Val, or c.1027T>G p.Trp339Gly) were published using the Burda 2015 numbering; the protein numbers are valid, but the c. position should be converted in VariantValidator before any use in a report [6, 28]. The same applies to variants c.891T>A p.(Tyr297Ter) and c.1916C>T p.(Thr639Ile) from the case of Li and colleagues (2026), which were not normalised in the survey on which this article is based [34].

2.5. Founder variants

Two populations have documented founder variants. In Bukharan Jews (originating mainly from Uzbekistan), the c.474A>T variant was found in the homozygous state in two unrelated children with severe deficiency; it causes abnormal splicing and premature termination, and the carrier frequency in healthy Bukharans was 1:39 (5 of 196), which led the authors to propose population carrier screening [8]. In the Old Order Amish of Somerset County, Pennsylvania, the c.1129C>T p.(Arg377Cys) variant is present at high frequency [9, 35]; in the Froese review it is the most frequent missense variant in unrelated families, with 10 families [7]. The splicing variant c.1530G>A p.(Lys510=) is the most frequent across the literature as a whole (13 families in the Froese review; 21 of 152 alleles, 13.8%, in Burda's series of 76 patients referred to the Zurich reference laboratory) [7, 29]. The seven Saudi patients of Aljassim and colleagues (2020) were all homozygous for c.680C>T p.(Thr227Met), written in the article and in GeneReviews as "680 T>C" [1, 32].

2.6. Residual activity and genotype–phenotype correlations

No genotype–phenotype correlations have been established for individual variants, largely because most variants are private; neither variant type nor location correlates consistently with phenotype [1, 17]. Residual enzyme activity is the best available correlate of severity: patients with early onset have less residual activity than those with late onset [7]. In the 2016 Huemer cohort, the 14 patients with activity below 1.5% of control had a mean onset at 1.5 months (median 1; 0.1–6), whereas the 19 with higher activity had a mean onset at 36 months (median 2.5; 0.1–216) [6]. Very low activity was associated with apnoea, hydrocephalus, microcephaly, and feeding problems, and higher activity with psychiatric symptoms, confusion, intellectual disability, myelopathy, ataxia, and spasticity; patients with higher activity did not die and responded better to treatment [6]. The authors caution that in vitro activity does not serve as an individual prognostic marker, because the relationship is not linear and siblings with the same genotype may follow different courses; the 1.5% threshold was chosen arbitrarily [6].

In Marelli's series of late-onset adults, missense variants in the C-terminal regulatory domain were overrepresented relative to early-onset cases, and residual activity in fibroblasts (n = 4) was 17–58% [11]. In the two Israeli families of Lossos, with adult-onset spastic paraparesis, fibroblast activity was 18–52% [36]. Some variants from the Huemer series illustrate these mechanisms: homozygous c.482G>A p.Arg157Gln (FAD-responsive, 10.4% activity) and homozygous c.596C>T p.Ala195Val (17%, onset at 216 months), in Burda numbering; none of the patients treated with riboflavin had FAD-responsive variants [6]. The variant effect map of Weile and colleagues (2021), which tested every possible amino acid substitution, allows prediction of the functional impact of missense variants and is cited by the 2026 guideline as a resource for interpretation [2, 37].

2.7. Common polymorphisms: c.665C>T and c.1286A>C

Both common variants are missense variants that decrease enzyme activity: c.665C>T p.(Ala222Val), the thermolabile variant, which is less active at higher temperatures, and c.1286A>C p.(Glu429Ala), which has a milder effect [1, 18]. More than 25% of Hispanic people and 10–15% of North American White people are homozygous for the thermolabile variant [18]. The two are in linkage disequilibrium, the combination usually occurs as compound heterozygosity in trans, and homozygosity for one with heterozygosity for the other is rare [18]. c.665C>T homozygotes tend to have slightly higher homocysteine and lower serum folate than controls, especially when folate intake is low [1, 18, 38].

These variants cause neither homocystinuria nor severe MTHFR deficiency [1, 2]. The 2026 guideline states that testing for polymorphisms such as c.665C>T is not indicated in the differential diagnosis of remethylation disorders [2]. In ClinVar, c.665C>T is classified as a drug response variant by an expert panel, and c.1286A>C has conflicting classifications [31]. An interaction with severe variants has been documented: in Selzer's patient, the rare variant c.1743G>A p.(Met581Ile) was in cis with c.665C>T, and the authors attributed the severe enzyme deficiency to this combination together with c.1286A>C on the other allele [33]. The c.665C>T polymorphism has also been studied as a modifier in cblC, with no effect on age at onset, phenotype, or outcome [39].

2.8. Allelic disorders and unsupported claims

No other Mendelian disorders caused by germline MTHFR variants are known [1]. OMIM and Orphanet entries linking MTHFR to susceptibility to vascular disease, thromboembolism, schizophrenia, or neural tube defects (for example ORPHA:563612 and ORPHA:563609, isolated anencephaly and exencephaly, with MTHFR as a susceptibility factor) refer to associations with common variants and not to monogenic disease [3, 40]. GeneReviews states that MTHFR has pseudogenes and that testing must distinguish them from the functional gene [1]; HGNC lists no MTHFR pseudogene, and none of the primary molecular genetics sources consulted (Burda 2015, Froese 2016, the 2017 and 2026 guidelines) mentions pseudogenes, so the claim is not reproduced here as a feature of the gene [2, 4, 29].

3. Clinical Spectrum and Manifestations by System

3.1. Overview and frequency of manifestations

Severe MTHFR deficiency presents mainly in early infancy but can become manifest at any age, with a high prevalence of cognitive impairment [17]. The 2026 guideline summarises the frequency of manifestations in patients with early and late onset, most of them diagnosed clinically rather than by newborn screening, drawing on case series, pooled reports, and meta-analyses (Table 3.1) [2]. The frequencies reflect the published literature, which is biased towards the most severe and most extensively studied cases, and an empty cell means that the guideline assigns no frequency, not that the manifestation is absent [2].

Table 3.1. Frequency of manifestations in severe MTHFR deficiency according to the 2026 E-HOD guideline (++++ >50%; +++ 25–50%; ++ 10–25%; + <10%; (+) isolated cases; — no frequency assigned)

Source: Table 2 of the 2026 E-HOD guideline, MTHFR deficiency columns [2]. HUS: haemolytic uraemic syndrome.

The megaloblastic anaemia scored ++ for early onset contrasts with the text of the same guideline, which considers macrocytic anaemia typical of the cblG and cblE defects but not of severe MTHFR deficiency, and with the 2017 guideline, which describes only isolated reports of megaloblastic anaemia in older children and adults with long-standing neurological disease [2, 17]. In the 2016 Huemer cohort, 4 patients with very low activity had macrocytosis or transient macrocytic anaemia, 6 had isolated normocytic or microcytic anaemia, and 2 had anaemia with low platelets and leucocytes [6]. In practice, the absence of megaloblastic anaemia favours MTHFR deficiency over methionine synthase defects, but its presence does not exclude it [12, 17].

The GeneReviews frequency table for neonatal onset, which distinguishes patients with early diagnosis and treatment from those with late or no treatment (for example microcephaly in 10–20% versus 50–70%), lists as its sources a review of polymorphisms and congenital anomalies, a single case, a series of 7 patients, and a meta-analysis of polymorphisms in psychiatric disorders, with no cohort denominator for severe disease [1]. For this reason, this article uses data from the cohorts with denominators described below.

3.2. Cohorts with denominators

Table 3.2. Main cohorts and series of severe MTHFR deficiency

The numbers in the European cohorts overlap: 43 patients from the 50-patient series of the 2019 E-HOD registry were included in the Yverneau cohort [10]. The 2026 guideline cites Marelli's adult series as comprising 13 patients; the article abstract reports 14 [2, 11].

3.3. Early-onset form

Pregnancy and delivery are usually normal. In the Yverneau cohort, 28 of 29 newborns were born at term, 26 of 27 were appropriate for gestational age, transition to extrauterine life was normal in 10 of 10 with available information, and 5 of 19 (26%) had neonatal microcephaly [10]. In the 2016 Huemer cohort, mean gestational age was 39.5 weeks and mean birth weight 3285 g, with two newborns below the 3rd percentile [6].

Newborns and young infants present with feeding difficulties, decreased level of consciousness, hypotonia, and hydrocephalus; apnoea, of central origin, is a frequent complication [10, 17]. In the Yverneau cohort, the 28 clinically diagnosed patients had a median age at diagnosis of 32 days (6–90) and a median interval between first symptoms and diagnosis of 4 days (0–65); at diagnosis, 21/28 (75%) had feeding difficulties and 23/28 (82%) had neurological symptoms, including hypotonia (17/28, 61%), microcephaly (9/28, 32%), and seizures (6/28, 21%); 9/28 (30%) had respiratory failure or apnoea, and 10/26 (38%) had poor or absent eye contact [10]. Hypothermia (5/10), somnolence (3/10), and haemodynamic failure (3/10) were other signs; cardiac and hepatic disease and thromboembolic events were rare (2/28 each), and there was no renal disease [10].

In older infants and children, seizures and cognitive impairment predominate, often with acquired microcephaly [17]. The 2017 guideline describes a "triphasic" course in some patients: initially normal development, then acquired microcephaly and psychomotor delay, and finally abrupt deterioration with potentially fatal respiratory failure [17]. Seizures are myoclonic, clonic, or tonic, and infantile spasms have been described [17]. GeneReviews describes generalised tonic-clonic, myoclonic, multifocal, and febrile seizures, with focal or generalised discharges on the electroencephalogram (EEG) [1]. Acute neurological deterioration in newborns and infants (seizures, lethargy, apnoea, coma) can be fatal or leave severe deficits [17].

Family series show the severity of the untreated natural history. In the Massadeh family, three siblings of the proband died at 40 days, 2 months, and 45 days with lethargy, hypoactivity, weak suck, and prolonged somnolence, having been admitted as cases of neonatal sepsis [5]. In the seven Saudi patients of Aljassim, homozygous for p.(Thr227Met), the presentation included apnoeas in early infancy, seizures, and hydrocephalus, which required shunting in one patient [32]. In the Huemer cohort, the four deaths occurred at 5, 6, 11, and 20 months, from apnoea (2), apnoea with pulmonary artery thrombosis (1), and infection (1) [6].

3.4. Late-onset form

In adolescence and adulthood, the picture is dominated by gait disturbance with lower-limb spasticity, weakness, cognitive decline, epilepsy, and schizophrenia-like psychosis, with leukoencephalopathy, demyelinating polyneuropathy predominantly affecting the lower limbs, and, in some patients, subacute combined degeneration of the spinal cord similar to that of cobalamin deficiency [12, 17]. Saudubray summarises the adult phenotype as psychiatric disorder, spastic paraparesis, thromboembolic events, and polyneuropathy, with specific treatment available [43].

In Marelli's series, the 14 adults had a median age at diagnosis of 32 years (11–54) and a median onset at 20 years (9–38), with gait difficulties (8), cognitive decline (3), or seizures (3), sometimes with preexisting mild intellectual disability (6) [11]. Over the course of the disease, manifestations were almost exclusively neurological: cognitive dysfunction 93%, gait disturbance 86%, epilepsy 71%, psychiatric symptoms 57%, polyneuropathy 43%, and visual impairment 43%; vascular events occurred in 28% and obesity in 36% [11]. The mean diagnostic delay was 14 years, and one patient was asymptomatic at age 55 years [11]. Cognitive and gait difficulties were associated with posterior white matter abnormalities on magnetic resonance imaging (MRI) [2].

Lossos and colleagues described two unrelated Israeli families, each with two affected siblings, with complicated hereditary spastic paraparesis beginning between ages 29 and 50 years, associated with polyneuropathy, behavioural changes, cognitive impairment, psychosis, seizures, and leukoencephalopathy; diagnosis was made about a decade later, and betaine lowered homocysteine by 50–70% in all, with improvement of the three ambulatory patients over 9–15 years [36]. GeneReviews reports one of these patients alive at age 67 years [1]. A Chinese case published in 2026, a 45-year-old man with spastic gait since age 38 years and recurrent pulmonary thromboembolism, was accompanied by a review that counted 28 patients from 17 families with adult onset published up to October 2025 [34]. In the 2026 Turkish cohort, acute deterioration also occurred in patients with chronic disease [42].

3.5. Manifestations by system

Central nervous system and neurodevelopment. Neurological disease is the dominant manifestation at all ages [41]. In the 2016 Huemer cohort, encephalopathy, lethargy, and hypotonia, and to a lesser extent apnoea, feeding problems, confusion, microcephaly, and epilepsy, improved over time, whereas intellectual disability, gait disturbance, and spasticity became more frequent [6]. At last follow-up in the Yverneau cohort (median 8.1 years), the neurological examination was abnormal in 19/29 (65%), mainly because of pyramidal syndrome (14/28, 50%), gait abnormalities (9/27, 33%), and hypotonia (9/28, 32%); 7/23 (30%) had microcephaly [10]. Gross motor skills were impaired in 14/28 and fine motor skills in 14/26; there were behavioural problems in 13/27, language difficulties in 7/9, and special educational needs in 12/20 [10]. GeneReviews reports that most children treated early walk independently and learn to speak [1].

Epilepsy. Epilepsy occurs in 70% of late-onset patients, at onset, at diagnosis, or during the course of the disease, and betaine is considered stabilising and sometimes improves seizures [2]. In the Yverneau early-onset cohort, clinical epilepsy was recorded in 3/10 patients with available information and was controlled with one antiseizure medication, although the EEG was abnormal in 14/17 (82%) [10]. Presymptomatic diagnosis with betaine treatment was associated with prevention of epilepsy [2].

Psychiatric and behavioural manifestations. These may be present at diagnosis; their proportion increases with age and does not appear to be influenced by treatment [2]. In the adult series, 57% had psychiatric symptoms [11]. Psychosis was the presenting feature of the index case in 1975 [23], and there are reports of psychotic symptoms and mood disorders reversible with betaine and B-complex vitamins [44, 45]. In children, GeneReviews describes anxiety, attention-deficit/hyperactivity disorder (ADHD), aggressiveness, sleep disorder, and autism spectrum disorder (ASD) [1]. Saudubray includes MTHFR deficiency among the treatable metabolic disorders that can present as isolated schizophrenia [24].

Eye. In the Yverneau cohort, 14/29 (48%) early-onset patients had ocular abnormalities at last follow-up, mainly nystagmus and strabismus and, less often, uncoordinated eye movements, optic atrophy, and cataract; the frequency did not improve with treatment [2, 10]. Up to 43% of late-onset adults had visual impairment [11], and in the 2016 Huemer cohort the eye was affected in 18%, with nystagmus, visual impairment, and optic atrophy [6]. One late-onset adult with complete visual loss fully recovered vision on treatment, and two cousins recovered only partially [46]. Bilateral lens subluxation has been described in a single young adult with MTHFR deficiency, in a report whose authors state there was no previous association; GeneReviews includes lens dislocation among the manifestations in adolescents and adults, but the published data document only this isolated case, and ectopia lentis remains a characteristic feature of CBS deficiency, not of MTHFR deficiency [1, 2, 47].

Vascular system. Arterial and venous thromboses are rare in childhood and occur in 18–28% of adults [2, 11, 41]. Stroke can be the presenting manifestation of any severe homocystinuria (tHcy >100 µmol/L) from late childhood to adulthood [24]. In the Yverneau early-onset cohort, there were no thromboembolic events at last follow-up (0/30) [10]. Recurrence on treatment appears to be low [2].

Respiratory. Central apnoea is a presentation in newborns and infants; in reports of subacute neurological deterioration, respiratory failure resolved with treatment over weeks to months, and the proposed mechanisms are central myelopathy and peripheral neuropathy [2, 10].

Kidney, heart, and other organs. MTHFR deficiency is not associated with renal disease, unlike cblC [2]. Structural heart disease and cardiomyopathy are very rare [2]; one patient in the Yverneau cohort died at 39 months of cardiomyopathy not identified at diagnosis [10]. Skeletal deformity and skin hyperpigmentation have been described in early-onset patients [2].

3.6. Neuroimaging and neuropathology

Reported MRI abnormalities include communicating hydrocephalus, cerebral and cerebellar atrophy, microgyria, white matter lesions, and thinning of the corpus callosum [1]. In the Yverneau cohort, all 9 available initial MRI scans were abnormal, with ventricular enlargement (6/9), cortical atrophy (5/9), enlarged extra-axial CSF spaces (5/9), and delayed myelination (4/9); one patient required a ventriculoperitoneal shunt [10]. In the 2016 Huemer cohort (n = 30), MRI was normal in 3, with cerebral atrophy in 15, white matter disease or delayed myelination in 13, and enlarged ventricles or hydrocephalus in 9; in some patients MRI stabilised or improved on treatment, including one with atrophy and hypomyelination at 3 months and a normal MRI at 3 years [6]. Basal ganglia calcifications on computed tomography are described in disorders of folate metabolism, including MTHFR deficiency [24]. Autopsy findings have included dilated cerebral vessels, microgyria, hydrocephalus, perivascular changes, demyelination, gliosis, astrocytosis, and macrophage infiltration; in some patients, thrombosis of cerebral arteries and veins was the main cause of death [12].

3.7. Natural history and treated disease

The untreated natural history of the early-onset forms includes high mortality in the first months of life: 31% of patients in the Diekman meta-analysis died, all untreated or treated late, and in the subanalysis using deceased siblings as genotypically identical controls, betaine prevented mortality (p = 0.002) [27]. Mortality is nil or very low in patients presenting in adolescence or adulthood [2]. When disease is treated after symptom onset, survival improves but established neurological deficits tend to persist: 76% of treated early-onset patients in the Yverneau cohort had a severe neurodevelopmental outcome, despite a median interval of 4 days between symptoms and diagnosis [10]. The nine patients diagnosed before symptom onset because they had an affected sibling all had normal cognitive development and neurological examination at last follow-up (3 months to 52 years); eight had no symptoms, and one needed educational support because of severe visual problems [10]. These data support the newborn screening discussed in sections 4 and 5 and the surveillance described in section 11.

4. Clinical Recognition and Red Flags

4.1. When to suspect

The 2017 guideline strongly recommends considering a remethylation disorder, acquired or genetic, in the presence of unexplained neurological, visual, or haematological symptoms, and also in unexplained thrombosis, spinal cord degeneration, or atypical haemolytic uraemic syndrome (recommendations 1 and 2, moderate quality) [17]. For MTHFR deficiency, the 2026 guideline adds three expert statements: consider the diagnosis at any age in the presence of unexplained neurocognitive impairment or decline, especially with other neurological abnormalities or multiorgan disease (92% consensus); consider it in the differential diagnosis of acute psychoses and unexplained behavioural changes (strong consensus, 96%); and test for it, by measuring tHcy, in patients of any age with unexplained thromboembolic events (96%) [2].

Table 4.1. Settings of suspicion by age

4.2. Red flags and emergencies

Apnoea, decreased level of consciousness, and seizures in a newborn or infant with elevated tHcy are emergencies: the guideline recommends starting empirical parenteral cobalamin in suspected remethylation disorders, before the defect is identified, and betaine as soon as hyperhomocysteinaemia with normal or low methionine is confirmed [17]. Samples for the differential diagnosis should be collected before treatment, but treatment should not be delayed while awaiting confirmation of the exact defect [2, 17]. Other red flags are neurological deterioration after anaesthesia with nitrous oxide, which in Selzer's case began 17 days after discharge, and signs of intracranial hypertension on betaine, which require measuring methionine and stopping the drug if it is markedly elevated [16, 17, 33].

Some patterns also point away from the diagnosis. Elevated or high-normal methionine suggests CBS deficiency rather than a remethylation disorder [17]. Elevated methylmalonic acid points to functional B12 deficiency (acquired or genetic), and marked megaloblastic anaemia to methionine synthase defects or to B12 or folate deficiency [12, 17]. Mild to moderate hyperhomocysteinaemia, common in B12 or folate deficiency, renal failure, and hypothyroidism, does not fit the profile of severe deficiency [17].

4.3. Newborn screening

Newborn screening for MTHFR deficiency is technically feasible using low methionine and a low methionine/phenylalanine ratio measured by tandem mass spectrometry (MS/MS) in dried blood spots, followed by tHcy as a second-tier test on the same sample [2, 17]. The 2026 guideline recommends newborn screening for MTHFR deficiency to allow early treatment with betaine (recommendation 31, moderate quality) [2]. MTHFR deficiency is not part of the US Recommended Uniform Screening Panel (RUSP), in which homocystinuria due to CBS deficiency is a primary condition and methylmalonic acidaemia with homocystinuria and hypermethioninaemia are secondary conditions [13], nor of the Portuguese National Newborn Screening Programme [14]. Details of the algorithm and its limitations are given in section 5.8.

5. Diagnosis: Criteria, Testing, and Algorithms

5.1. Diagnostic criteria

No published consensus clinical criteria exist [1]. The diagnosis is established in a patient with suggestive clinical and biochemical findings (hyperhomocysteinaemia, low or normal methionine, normal MMA) and biallelic pathogenic or likely pathogenic variants in MTHFR, classified according to ACMG/AMP criteria [1]. When genetic testing is inconclusive, confirmation is by enzyme assay in fibroblasts in an experienced laboratory [2]. Biallelic variants of uncertain significance (VUS), or one pathogenic variant with one VUS, neither establish nor exclude the diagnosis [1].

5.2. First-line biochemistry and preanalytical considerations

Evaluation of a suspected remethylation disorder begins with measurement of total homocysteine in plasma, or in serum if plasma is not available (recommendation 1 of the 2026 guideline, high quality) [2]. The sample should be centrifuged within 1 hour and kept at +4 °C or frozen until analysis; if immediate centrifugation is not possible, it is kept on ice or in ice water and centrifuged within 6 hours; once separated, homocysteine is stable for weeks at 4 °C and for years frozen (recommendation 2, moderate) [2, 17]. Immunoassays, chromatography, and liquid chromatography–tandem mass spectrometry (LC-MS/MS) are suitable; chromatographic methods allow simultaneous measurement of methionine, cysteine, and cystathionine [2, 17]. The guideline recommends against measuring free homocystine instead of tHcy (recommendation 3), because free homocystine may be undetectable even when tHcy is very high, and a poorly stored sample yields falsely low values because of protein binding [2, 48].

When tHcy is above the age-adjusted range, plasma for methionine, folate, vitamin B12, holotranscobalamin (if available), and MMA should be collected before any treatment (recommendation 4, high quality); renal function (creatinine and/or cystatin C) should be assessed, especially in adults (strong consensus, 96%) [2]. A pretreatment sample can be frozen so as not to delay treatment, and dried blood spots can be used when shipping is difficult [2, 17]. Untreated patients with remethylation disorders usually have tHcy above 100 µmol/L, but values below 100 and even below 50 µmol/L have been observed [2].

Table 5.1. Biochemical markers in severe MTHFR deficiency

Table 20.1 in Saudubray shows that the profile of normal or low methionine, elevated tHcy, normal or low SAM, elevated SAH, and normal or elevated cystathionine is shared by MTHFR deficiency, the cblE, cblG, and cblD-variant 1 defects, the combined cobalamin defects, and B12 and folate deficiencies; what distinguishes MTHFR deficiency is normal MMA, normal B12, the usual absence of macrocytosis, and low CSF folate [17, 49]. Low CSF 5-MTHF is not specific: it is also found in folate receptor α deficiency (FOLR1), in which depletion is usually more marked, and in other secondary causes such as Rett syndrome, Kearns-Sayre syndrome, and white matter diseases [12].

5.3. Diagnostic algorithm

Figure 3 of the 2026 guideline organises the diagnosis around three entry points: a compatible clinical picture with elevated tHcy, an abnormal newborn screen, or suspicion raised by next-generation sequencing (NGS) with elevated tHcy [2]. In all of them, the pretreatment panel is collected: serum or plasma B12 (and holotranscobalamin, if available), folate, plasma tHcy and methionine, acylcarnitines, and plasma or urinary MMA [2].

Table 5.2. Branches of the 2026 E-HOD guideline algorithm with elevated tHcy and normal or low methionine

Source: Figure 3 of the 2026 E-HOD guideline [2]. With normal tHcy and normal or low methionine, and normal B12, folate, and MMA, the guideline concludes that there is no remethylation disorder and suggests considering MTHFD1 deficiency. Folate or B12 deficiency may coexist with a genetic remethylation disorder.

The coexistence of nutritional deficiency and genetic disease is relevant: a patient with MTHFR deficiency may have low serum folate, and a newborn of a mother with B12 deficiency may have homocystinuria and elevated MMA, mimicking a genetic defect [1, 2]. In cases detected by screening, plasma from both the infant and the mother should be collected to assess B12 status before treatment (recommendation 7, high quality) [2].

5.4. Enzyme studies

Direct assay of MTHFR activity in cultured fibroblasts, preferably in the physiological direction of the reaction, is available in reference laboratories and is, among the remethylation disorders, the only direct enzyme assay available besides that for MTHFD1 [2, 17]. Activity can also be measured in leucocytes, lymphocytes, liver, and cultured amniocytes [12, 17]. In indirect fibroblast assays, MTHFR deficiency shows normal incorporation of propionate and methyltetrahydrofolate, normal or increased incorporation of formate into serine and decreased incorporation of formate into methionine, with normal adenosylcobalamin and decreased methylcobalamin synthesis [17]. The enzyme assay is useful when only one pathogenic variant is identified, when there are VUS, and to characterise novel variants [2, 17]. In the Huemer cohort, in vitro activity did not correlate with initial tHcy or methionine [6].

5.5. Molecular genetics

Molecular genetic testing is the method of choice to confirm a remethylation disorder (recommendation 5, high quality): NGS, independent confirmation of variants, and phase determination in the parents; if the result is inconclusive, an enzyme assay [2]. GeneReviews describes two strategies: single-gene sequencing, followed by deletion/duplication analysis if only one or no variant is found, with targeted testing for the founder variants in people of Bukharan Jewish or Amish ancestry; or a multigene panel that includes the genes in the differential diagnosis; for an atypical phenotype, exome or genome sequencing [1]. For interpretation of missense variants, the Weile variant effect map is available as a functional reference [37].

MTHFR testing must be interpreted with caution because of the abundance of polymorphisms in the gene; the thermolabile variant c.665C>T does not cause severe disease and need not be tested in this differential diagnosis [2, 38]. An adequate report identifies the gene, the versioned transcript (NM_005957.5), the variant in HGVS nomenclature at the DNA and protein levels, zygosity, phase confirmed in the parents, and ACMG/AMP classification; when the laboratory or the literature uses historical numbering, the correspondence should be confirmed in VariantValidator or ClinVar (Table 2.1) [2, 28].

5.6. Test limitations

The proportion of variants detectable by each method (about 99% by sequencing and 1% by deletion/duplication analysis, according to the HGMD data cited by GeneReviews) describes the composition of the published spectrum and does not equal the clinical sensitivity of the test in a given patient [1]. A negative result does not exclude the diagnosis when the biochemistry is suggestive: deep intronic variants, rearrangements, and splicing variants not recognised as such can escape exome sequencing, and the most frequent variant in the disease, c.1530G>A, is synonymous and is recognised as pathogenic only through its effect on splicing [7, 31]. Apparent homozygosity may reflect an undetected exon deletion on the other allele or uniparental isodisomy, and a variant found in only one parent may be de novo in the proband or result from parental mosaicism [1]. In these scenarios, parental testing and fibroblast enzyme assay resolve most cases [2].

5.7. Variants of uncertain significance and segregation

When a VUS is found, confirmation of trans phase in the parents, the complete biochemical profile, and fibroblast enzyme activity are the elements that allow a decision; the enzyme assay is specifically indicated for VUS [2]. GeneReviews points out that two VUS, or one pathogenic variant and one VUS, neither confirm nor rule out the disease [1].

5.8. Newborn screening: algorithm and limitations

Screening for cblD-Hcy, cblE, cblG, and MTHFR is based on low methionine and/or a low methionine/phenylalanine ratio in dried blood spots, with second-tier tHcy, which distinguishes patients from controls [2, 17]. Saudubray considers that effective screening requires critical selection of the low-methionine cutoff, supported by second-tier tHcy [12]. tHcy has recently begun to be used as a first-tier marker, with multiplexing of homocysteine in MS/MS assays through selective thiol derivatisation developed by the US Centers for Disease Control and Prevention (CDC) [2, 50]. Genomic newborn screening (exome or genome sequencing for a predefined list of genes) can identify remethylation disorders, including MTHFR deficiency, missed by biochemical screening [2].

The absence of abnormal acylcarnitines distinguishes this scenario from that of the combined cobalamin defects, in which propionylcarnitine (C3) and the C3/C2 ratio are the primary markers and low methionine (for example <13.4 µmol/L in a proposed algorithm for cblC) is a secondary analyte [39]. In MTHFR deficiency, C3 is not elevated, so detection depends on the primary marker of low methionine [2, 39]. In the USA, each laboratory sets its own cutoffs, and an out-of-range result requires immediate confirmation, consultation with a metabolic specialist, and consideration of presumptive treatment [1]. An abnormal result is not specific: low methionine and elevated tHcy are also found in newborns of mothers with B12 deficiency [2, 40].

Evidence for the effectiveness of screening programmes is limited. Wong and colleagues described two patients with MTHFR deficiency detected by low methionine on screening, already symptomatic despite early diagnosis, who improved markedly with treatment, compared with a patient diagnosed clinically at 3 months with significant morbidity [51]. The authors of the Yverneau cohort note that the yield of methionine and of the methionine/phenylalanine ratio appears imperfect and that the cost of tHcy in dried blood spots is high, but consider that the availability of effective treatment justifies screening, and call for a prospective study [10]. At the time of that study, few screening programmes detected MTHFR deficiency in 2022 [10].

5.9. Prenatal and preimplantation diagnosis

Once the variants have been identified in the index case and the parents' carrier status confirmed, prenatal or preimplantation diagnosis is performed by molecular genetic testing of chorionic villi or amniotic fluid (recommendation 6, moderate quality) [2]. As a rare alternative, MTHFR activity can be measured in cultured amniocytes; indirect assays in chorionic villi should be avoided [2, 17]. Classic methods included measurement of tHcy and MMA in amniotic fluid [17]. Preconception or prenatal carrier screening panels that include remethylation disorder genes are already offered [2].

5.10. Other investigations at diagnosis

Brain MRI is part of the initial evaluation, and MR angiography and MR venography are indicated when there are signs of a neurovascular complication [1]. The 2026 guideline advises EEG at diagnosis in patients with neurological signs (88% consensus) and standardised neuropsychological assessment with validated instruments (100% consensus) [2]. The 2017 guideline recommends ophthalmology consultation in all newly diagnosed patients with a remethylation disorder, regardless of age and severity (recommendation 16, low quality); the 2026 guideline recommends ophthalmological examination as clinically indicated (100% consensus) [2, 17].

6. Differential Diagnosis

6.1. Comparison table

Table 6.1. Differential diagnosis of severe MTHFR deficiency

Sources: Tables 3 and 4 of the 2017 guideline, Table 3 and Figure 3 of the 2026 guideline, Table 3 of GeneReviews, and Table 20.1 of Saudubray [1, 2, 12, 17, 49]. AR: autosomal recessive; XL: X-linked; FIGLU: formiminoglutamic acid.

6.2. Main differentials

The most difficult differential is with the isolated methionine synthase defects, cblE (MTRR) and cblG (MTR), which share the profile of homocystinuria without methylmalonic aciduria and with low methionine [12]. What distinguishes them clinically is megaloblastic anaemia, frequent in cblE and cblG and usually absent in MTHFR deficiency [2, 12]. More than 30 patients have been described with each of these defects, most symptomatic in the first year, but one cblG patient was diagnosed at age 21 years after having been treated for multiple sclerosis [12]. The distinction is made by sequencing and, if necessary, by complementation studies and indirect fibroblast assays; parenteral hydroxocobalamin has a proven clinical effect in the cobalamin defects but not in MTHFR deficiency [12, 17].

The combined cobalamin defects, of which cblC is the most common, are distinguished by elevated MMA and elevated propionylcarnitine on newborn screening [2, 49]. Clinically, MTHFR deficiency may be indistinguishable from cblC, but it usually lacks renal involvement and has less ocular involvement than the cobalamin disorders; maculopathy, frequent in early-onset cblC, is exceptional in MTHFR deficiency [2].

Homocystinuria due to CBS deficiency is distinguished by elevated or high-normal methionine, low cystathionine, and a phenotype of ectopia lentis, severe myopia, skeletal abnormalities, and thromboembolism [1, 17]. The biochemical difference has therapeutic consequences: in CBS deficiency, betaine raises an already elevated methionine, and methionine above 1000 µmol/L is associated with cerebral oedema, whereas in MTHFR deficiency betaine corrects a low methionine [16, 49]. The treatment of CBS deficiency follows a pyridoxine-responsiveness algorithm.

Acquired B12 and folate deficiencies are the most common causes of hyperhomocysteinaemia [17]. In newborns and infants, maternal B12 deficiency must be excluded, especially in infants breastfed by strict vegan mothers or by mothers with undiagnosed pernicious anaemia; maternal B12 may not be markedly low, so B12 and MMA are assessed in both the infant and the mother [1, 17]. Table 3 of the 2017 guideline lists other mimics: B12 malabsorption (intrinsic factor deficiency, autoimmunity, parasites, short bowel), nitrous oxide, antifolates, severe iron deficiency, infections with cytomegalovirus, Epstein-Barr virus, human herpesvirus 6, parvovirus B19, or HIV-1, leukaemias and myeloproliferative disorders, myelodysplasia, respiratory chain disorders (such as Pearson syndrome), orotic aciduria, SLC19A2 deficiency, methylmalonic and propionic acidaemias, and lysinuric protein intolerance [17].

6.3. Common MTHFR polymorphisms: a frequent false diagnosis

A person with a report of heterozygous or homozygous c.665C>T or c.1286A>C, obtained on a thrombophilia panel or a direct-to-consumer test, does not have severe MTHFR deficiency [18, 40]. Saudubray lists "MTHFR polymorphism" among the causes of mildly increased homocystine with an otherwise normal profile, alongside B12 and folate deficiencies [48]. If there is clinical concern, the appropriate test is plasma tHcy, not the genotype; anticoagulation, aspirin, or high-dose folic acid are not indicated on the basis of genotype alone [18, 40]. The place of these variants in thrombophilia and neural tube defects is addressed in section 8.6.

6.4. Differential by presentation

In a newborn with encephalopathy, apnoea, and seizures, the differential includes the treatable epileptic encephalopathies: Saudubray recommends systematic trials of pyridoxine, pyridoxal 5'-phosphate, biotin, and folinic acid in refractory neonatal seizures, and points out that methyl transfer defects (MTHFR and cobalamin) may require hydroxocobalamin, folates, pyridoxine, betaine, or methionine depending on the defect [52]. In Saudubray's table of epilepsies, cblC and MTHFR deficiency share multifocal and partial spikes on the EEG, developmental delay, intellectual disability, psychiatric symptoms, and abnormal movements, and the recommended workup includes organic acids, tHcy, and folate [24]. In an adult with spastic paraparesis, severe hyperhomocysteinaemia with low methionine distinguishes MTHFR deficiency from genetic hereditary spastic paraplegias without a biochemical marker [36]. In an adult with new-onset psychosis, tHcy is part of the workup for treatable metabolic disorders [2, 24].

7. Treatment and Risk Management Strategies

7.1. Goals and principles

There is no cure; treatment aims to restore homocysteine remethylation through a pathway that does not depend on MTHFR, correct methionine and SAM, lower tHcy, and supply 5-MTHF to the central nervous system [1, 2]. The 2026 guideline recommends lowering plasma tHcy to the lowest concentration achievable in each patient and safely maintaining plasma methionine within age-adjusted reference ranges (recommendation 9, moderate quality) [2]. The heterogeneity of the remethylation disorders precludes a single tHcy target: in retrospective cohorts, treatment lowers tHcy from values above 100 µmol/L at diagnosis to 20–80 µmol/L, and in MTHFR deficiency the reduction is about 40%, to medians of 70–110 µmol/L (range 16–328) [2]. The panel considers tHcy values below 100 µmol/L safe in CBS deficiency and individual titration to the lowest achievable value useful in the remethylation disorders (92% consensus) [2].

The target in the European betaine label, tHcy below 15 µM or as low as possible, and that in the US label, undetectable tHcy or tHcy present only in small amounts, are rarely achieved in MTHFR deficiency [15, 16]. In the Yverneau cohort, tHcy fell on average 1.8-fold on treatment but remained above the reference range in 26/28 (93%) patients, and only 2/28 (7%) fell below 70 µmol/L; by contrast, methionine increased more than threefold and was within the reference range in all 24 patients with data [10]. tHcy did not differ between patients with severe and nonsevere outcomes, either at diagnosis or at last follow-up [10]. The authors suggest that the benefit of presymptomatic treatment is due to normalisation of methionine and SAM, despite persistently elevated tHcy [10].

7.2. Acute phase and suspected but unconfirmed disease

In a patient with a suspected remethylation disorder, the 2017 guideline recommends immediate treatment with parenteral cobalamin, before the defect is identified (recommendation 18, moderate quality), and considering betaine as soon as hyperhomocysteinaemia with normal or low methionine is confirmed (recommendation 19, moderate) [17]. In individual cases, it suggests enteral supplementation with folinic acid or L-methionine (recommendation 20, low quality) [17]. In cblC, the recommended initial hydroxocobalamin dose is 1 mg/day parenterally (about 0.33 mg/kg/day in a 3-kg newborn), with the intramuscular route preferred and no maximum dose defined in the source; frequency is reduced as the patient stabilises, and dose escalation improved metabolic control in some reports [17]. Once cobalamin defects have been excluded and MTHFR deficiency confirmed, parenteral hydroxocobalamin is no longer needed if enteral cobalamin absorption is intact [2]. For severe MTHFR deficiency, Figure 3 of the 2026 guideline indicates starting oral betaine and considering folinic acid or methylfolate and/or L-methionine [2]. Supportive treatment in the acute phase includes nasal continuous positive airway pressure (CPAP) and, in severe cases, intubation and ventilation for apnoeas, and treatment of seizures [1].

7.3. Anhydrous betaine

Mechanism. Betaine (trimethylglycine), derived from choline, is the methyl donor for hepatic BHMT, which remethylates homocysteine to methionine by a pathway independent of 5-MTHF and methylcobalamin; the effect is twofold, lowering homocysteine and raising plasma methionine and SAM [12, 16, 17]. Because BHMT is not expressed in the brain, the neurological benefit depends on circulating methionine [12]. Adherence can be monitored with plasma betaine, dimethylglycine, and sarcosine [2, 53]; plasma betaine does not define efficacy [16].

Evidence. Betaine is the treatment of choice in severe MTHFR deficiency (high quality, according to the 2026 guideline) [2]. The guideline recommends early betaine treatment because it reduces mortality (recommendation 25, moderate quality) and because, if started before symptom onset, it significantly improves clinical outcome and potentially prevents neurological deterioration (recommendation 26, moderate), and suggests treating patients with thromboembolism with betaine (recommendation 27, low) [2]. In 2017, the equivalent recommendations had a high quality of evidence, based on the Diekman meta-analysis: betaine at a dose ≥100 mg/kg/day was associated with better survival, with 2 deaths among 26 patients after treatment initiation [17, 27]. In Marelli's adult series, acute symptoms (confusion, psychiatric decompensation, worsening gait, seizures) improved in most treated patients (n = 9), and chronic symptoms such as mild cognitive impairment stabilised (n = 4) [2, 11].

Table 7.1. Anhydrous betaine dosing according to the sources consulted

The recommended dose is that of the 2026 E-HOD guideline: 100–250 mg/kg/day in children and 6–9 g/day in adolescents and adults, in 2–3 divided doses [2, 17]. The labels specify 100 mg/kg/day in 2 doses, titrated to tHcy and methionine, in children and adults (European SmPC for Cystadane® and Amversio®), and 6 g/day from age 3 years or 100 mg/kg/day with weekly 50 mg/kg increments below age 3 years (FDA label); both allow titration beyond the starting dose (more than 200 mg/kg/day in some patients in the EU; up to 20 g/day in the USA) [15, 16, 54]. The 2026 guideline defers the maximum dose to the manufacturer [2]; the European SmPC adds that, below the age of 10 years, more than two daily doses or more than 150 mg/kg/day did not increase the reduction in tHcy [16]. By comparison, the guideline's paediatric initiation above 100 mg/kg/day and division into 3 doses fall outside the label conditions, while the adult dose of 6–9 g/day falls within the range of the US label [2, 15]. Recorded practice varies. In the 2016 Huemer cohort, 29 patients received betaine at 200 to 24,000 mg/day [6]; in the Yverneau cohort, all early-onset patients were receiving betaine at last follow-up, at a median dose of 157 mg/kg/day (52–350), and the nine presymptomatic patients with data (5/5) were receiving a median of 174.5 mg/kg/day (44–284) [10]. In the RoCH post-authorisation safety registry (France and Spain, 2007–2014; 125 patients with homocystinuria), the median total daily dose in patients with MTHFR deficiency was 9 g/day [56]. In the Amish series, doses of 534 ± 222 mg/kg/day were used, which increased SAM but did not normalise tHcy [9]. In one patient with cblC and cor pulmonale, 850 mg/kg/day was used [17]. Frequent administration of moderate doses may be beneficial, and the pharmacokinetics of betaine are poorly characterised [17].

Pharmacokinetics, formulation, and administration. In volunteers aged 21–49 years given 50 mg/kg, time to peak was 0.9 ± 0.3 h, peak concentration 0.9 ± 0.2 mM, mean elimination half-life 14 h, and renal clearance 5% [16]. Cystadane® is a hygroscopic oral powder in a 180-g bottle; in the USA, one level measuring scoop (1.7 mL) corresponds to 1 g, and in the EU there are three measuring scoops, of 100 mg, 150 mg, and 1 g [15, 16]. The powder is dissolved in 120–180 mL of water, juice, milk, or formula, or mixed with food, and taken immediately [15]. The European SmPC recommends separating betaine intake by 30 minutes from amino acid mixtures and from vigabatrin or GABA analogues, because of an interaction described in vitro, and requires no adjustment in renal failure or nonalcoholic fatty liver disease [16]. The product has a shelf life of 3 years unopened and 3 months after opening, to be stored below 25 °C in the EU and between 15 and 30 °C, protected from moisture, in the USA [15, 16].

Adverse effects and management. Betaine is generally well tolerated even at high doses (3–6 g/day in infants; 9 g/day in adults) [17]. In the survey of 41 physicians covering 111 patients that supports the US label, the reported effects were nausea (2), gastrointestinal discomfort (2), diarrhoea (1), bad taste (1), odour (1), questionable psychological changes (1), and aspiration of the powder (1) [15]. The European SmPC lists increased blood methionine as very common and, as uncommon, anorexia, agitation, irritability, cerebral oedema, diarrhoea, glossitis, nausea, gastric discomfort, vomiting, alopecia, urticaria, abnormal skin odour, and urinary incontinence [16]. In the RoCH registry there were 3 drug-related adverse events out of 277: bad taste and headache, both nonserious, and one serious case of interstitial lung disease [56]. In a 15-month-old infant, seizures appeared shortly after starting 250 mg/kg/day [17].

The most serious risk, cerebral oedema due to hypermethioninaemia, has been described in patients with CBS deficiency, with methionine between 1000 and 3000 µM, between 2 weeks and 6 months after initiation, with complete recovery after discontinuation [15, 57]. The labels recommend keeping methionine below 1000 µM; the European SmPC advises measuring methionine at baseline and then about once a year or every six months, intensifying monitoring and reviewing the diet if it exceeds 700 µmol/L, interrupting betaine if symptoms of cerebral oedema appear (morning headaches with vomiting and/or visual changes), and discontinuing it permanently if symptoms recur on rechallenge [15, 16]. In MTHFR deficiency, in which betaine corrects a low methionine, the sources consulted describe no cases of betaine-associated cerebral oedema [2, 57]; methionine monitoring is still required, especially when L-methionine is added [1, 16].

7.4. Folates: methylfolate, folinic acid, and folic acid

Rationale. Betaine does not supply 5-MTHF to the brain, and CSF 5-MTHF remains low in patients treated with betaine alone [9]. Methylfolate (5-methyltetrahydrofolate, as calcium mefolinate) bypasses MTHFR; folinic acid (5-formyltetrahydrofolate, calcium folinate) and folic acid require functional MTHFR to be converted to 5-MTHF [1, 55]. Knowles and colleagues described three patients with severe deficiency (activity ≤1%) in whom CSF 5-MTHF was undetectable at diagnosis and on folic acid or calcium folinate, and became measurable only with oral calcium mefolinate at doses of 15–60 mg/day [58, 59]. The original abstract of this article stated "15–60 mg/kg/day"; the published erratum corrects the unit to mg/day, and the abstract indexed in PubMed retains the error [59]. The 2017 guideline notes that 5-MTHF up to 45 mg/day did not correct low CSF 5-MTHF in one of these patients, and that the instability of oral 5-MTHF was an unresolved problem [17].

Folic acid. Folic acid at doses of 5–80 mg/day has been associated with both good and poor responses, with no dose–response relationship, and may worsen cerebral methyl-THF deficiency [17]. The 2017 guideline listed folic acid, together with nitrous oxide and protein restriction, among the agents to be avoided in MTHFR deficiency, and the 2026 guideline maintains that folic acid should be avoided [2, 17]. GeneReviews mentions folic acid in treatment during pregnancy, in contradiction with both guidelines [1, 2].

Recommendations. In 2017, the evidence could neither demonstrate nor exclude benefit or harm from folic acid, folinic acid, or 5-MTHF as adjunctive therapy (recommendation 31, low quality) [17]. In 2026, expert statement 39 establishes that every patient with severe MTHFR deficiency should be treated with betaine and with folinate or methylfolate, regardless of the presence of psychiatric symptoms; the text does not give the consensus percentage for this statement [2]. Table 4 of the guideline recommends using methylfolate if available and, failing that, folinic acid, with commonly used doses of 45 mg/day in children and 75 mg/day in older patients (expert opinion) [2]. GeneReviews considers that folinic acid does not substitute for 5-MTHF, because it requires functional MTHFR to be converted [1], and Saudubray recommends 5-methyltetrahydrofolate in MTHFR deficiency, at an oral dose of 15–60 mg/day as calcium mefolinate [55]. GeneReviews gives L-5-MTHF 20 mg orally twice daily, adjusted according to clinical response, tHcy, and CSF 5-MTHF [1].

In recorded practice, use of 5-MTHF has been in the minority: in the Yverneau cohort, at last follow-up, 17 patients were receiving folinic acid (median 0.8 mg/kg/day), 7 folic acid (0.05 mg/kg/day), and only 2 5-MTHF (0.3 mg/kg/day) [10]; in the 2016 Huemer cohort, 19 patients were receiving folate or folinate at doses of 3 to 100 mg/day, and the impact of folinic acid was not convincing [6]. The availability of calcium mefolinate as a medicinal product varies between countries; there is no FDA label or centralised European marketing authorisation, and in INFARMED's Infomed database a search on 26 October 2026 for mefolinate, methylfolate, and levomefolate returned no authorised product; its use in MTHFR deficiency is off-label. In Portugal, oral calcium folinate (Lederfoline® 15 mg) has an SmPC only for methotrexate rescue and folic acid antagonist overdose, with no dosing applicable by analogy, so the folate doses in this article are those of the 2026 guideline and GeneReviews, all off-label [1, 2, 60].

7.5. Cobalamin

In confirmed MTHFR deficiency, parenteral hydroxocobalamin is unnecessary when enteral cobalamin absorption is intact; the 2026 guideline recommends oral cobalamin to prevent B12 deficiency and optimise tHcy (expert opinion), without specifying a dose [2]. The 2017 guideline considered the efficacy of cobalamin in MTHFR deficiency not assessable [17]. GeneReviews proposes vitamin B12 1 mg (up to 2 mg) subcutaneously every few days, reducing to once weekly and then once monthly according to tHcy (the subcutaneous route does not appear on hydroxocobalamin labels, which specify only the intramuscular route, so this is off-label use) [61, 62], and notes that the fall in homocysteine may take weeks; the reference it cites for B12 monitoring is a study of polymorphisms in pregnant women, not of severe deficiency [1]. Practice varies: in the 2016 Huemer cohort, 19 patients were receiving hydroxocobalamin, from 1 mg/day intramuscularly or orally to 1 mg intramuscularly per month, and 4 cyanocobalamin [6]; in the Yverneau cohort, 19 were receiving hydroxocobalamin and 4 cyanocobalamin [10]. Adverse effects of hydroxocobalamin, which are rare, include red discolouration of urine and skin, hypersensitivity, photosensitivity, nausea, injection-site reactions, and headache; doses up to 20 mg/day have been tolerated in cblC [17]. The hydroxocobalamin doses in Saudubray (1–2 mg/day or 5 mg/week intramuscularly; 10 mg once or twice daily orally) refer to the cobalamin defects and not to MTHFR deficiency [55].

7.6. L-methionine

Methionine is an essential amino acid, and the guideline recommends keeping it within the normal range, with oral supplementation if necessary (2017 recommendation 29, low quality) [17]. In MTHFR deficiency, the 2026 guideline recommends L-methionine supplementation in 2–3 daily doses if methionine is not safely maintained in the normal range on optimised betaine, without specifying a dose [2]. GeneReviews proposes 40–50 mg/kg/day, adjusted to keep plasma and CSF methionine in the upper normal range, with caution because of the potential increase in homocysteine [1]. Clinical improvement with methionine has been described in a very small number of patients [17, 63], including marked improvement after early introduction in one patient with severe deficiency [12]. In the 2016 Huemer cohort, 11 patients were receiving 100–1500 mg/day, and in the Yverneau cohort 7 patients were receiving a median of 8 mg/kg/day (4–12.5) [6, 10].

7.7. Other supplements and diet

Pyridoxine, riboflavin, cobalamin, and carnitine have no assessable efficacy in MTHFR deficiency with the available data [6, 17]. In Table 4 of the 2026 guideline, L-carnitine is not applicable to MTHFR deficiency [2]. Saudubray includes folic acid, reduced folates, methionine, pyridoxine, cobalamin, and carnitine among agents of unproven efficacy, possibly useful in isolated cases, and notes that most regimens without betaine were not effective [12]. High-dose pyridoxine can cause peripheral neuropathy (above 900 mg/day in adults) [55]. Protein restriction should not be used in any remethylation disorder (recommendation 28, moderate quality) [17]; conversely, GeneReviews advises against an excessively high-protein diet, which may worsen homocysteine [1].

7.8. Agents and situations to avoid

Nitrous oxide irreversibly oxidises the cobalt of cobalamin and inhibits methionine synthase, removing the only remethylation pathway still functional in the brain; the 2026 guideline strongly recommends against its use in any remethylation disorder (recommendation 33, high quality) [2, 33]. Folic acid should be avoided [2, 17]. GeneReviews advises caution with drugs that interfere with folate metabolism, such as methotrexate, and with anticonvulsants that may raise homocysteine, such as valproic acid, as well as avoiding prolonged periods without monitoring of B12, folate, and tHcy [1]. There is no evidence against specific antipsychotics (100% consensus) [2].

7.9. Supportive treatment and rehabilitation

Treatment is multidisciplinary [1]. For apnoeas, CPAP or oxygen is used, and a sleep study is performed when indicated; for feeding difficulties, feeding therapy, swallowing assessment with a low threshold for videofluoroscopy, and gastrostomy if problems persist [1]. Seizures are treated with conventional antiseizure medications, none of which has shown specific efficacy; valproate should be used with caution [1]. Motor rehabilitation includes orthopaedics, physical medicine and rehabilitation, physical therapy, and occupational therapy, with positioning and mobility devices; hypertonia and dystonia can be treated with baclofen, tizanidine, botulinum toxin, antiparkinsonian drugs, or orthopaedic surgery [1]. Stroke prevention and anticoagulation after thrombosis follow general indications, and betaine is suggested for patients with thromboembolism [1, 2]. Early intervention, educational support, social services, and, in severe cases, palliative and respite care complete the plan [1].

7.10. Regulatory status

Anhydrous betaine is the only medicinal product with an approved indication that explicitly names MTHFR deficiency [15, 16]. In the USA, the indication is the treatment of homocystinuria to decrease elevated blood homocysteine concentrations in paediatric and adult patients, including CBS deficiency, MTHFR deficiency, and defects in cobalamin cofactor metabolism [15]. In the EU, the indication is adjunctive treatment of homocystinuria involving deficiencies or defects in CBS, MTHFR, or cobalamin cofactor metabolism, as a supplement to other therapies such as vitamin B6, vitamin B12, folate, and a specific diet [16]. Hydroxocobalamin, calcium mefolinate, calcium folinate, and L-methionine have no approved indication for MTHFR deficiency in the sources consulted, and their use is off-label [2].

Table 7.2. Regulatory comparison of anhydrous betaine

The sources consulted do not include the reimbursement or access status of betaine in Portugal (INFARMED) or its registration in Brazil (ANVISA); the centralised European marketing authorisation applies to Portugal, but actual availability and dispensing arrangements were not documented.

7.11. Evidence from trials and interventional studies

There are no randomised clinical trials dedicated to severe MTHFR deficiency [17, 40]. The 2017 guideline found no randomised trial, and the high-quality evidence for betaine derived from a case meta-analysis showing a temporal relationship between treatment initiation and outcome [17]. Searches of ClinicalTrials.gov in October 2026 identified no dedicated interventional trials; the European trials register (EU CTR/CTIS) was not searched, and the absence of results does not prove that no studies exist [64, 65].

Table 7.3. Interventional studies and comparative evidence relevant to MTHFR deficiency

Yverneau's adjusted OR appears as 0.004 in the published abstract and as 0.04 in Table 4 and in the results text of the manuscript, with the same confidence interval (0.002–0.232); the OR in the univariate analysis was 0.039 (0.002–0.238) [10]. The studies in the table are not comparable with one another: the only randomised trial did not include patients with MTHFR deficiency in its publication and measured a biochemical outcome; the meta-analysis and the cohorts are observational and retrospective, with missing data and heterogeneous treatment, and Yverneau's presymptomatic group was identified because of an affected sibling, which may be associated with different surveillance and adherence [10, 27]. The percentage reductions in homocysteine described in the US label pool disorders with different mechanisms and provide no information on neurological outcome in MTHFR deficiency [15]. Study NCT04287881, registered as "MTHFR deficiency" in adults with ischaemic stroke and epilepsy, actually concerns the common polymorphisms and not severe deficiency [67].

7.12. Developments in the past 12 months and research

Table 7.4. Developments between October 2025 and October 2026

The searches carried out for this article identified no gene therapy, messenger RNA therapy, or pharmacological chaperone for MTHFR deficiency in preclinical studies published between 2024 and 2026; one result on "MTRFR", a mitochondrial release factor, corresponded to a different gene. In 2024, HCU Network America and HCU Network Australia awarded a research grant to ETH Zurich for MTHFR- and cblG-deficient brain organoid models, intended for testing folates and other therapies; this is preclinical research funded by patient organisations, with no published results in the sources consulted [69, 70].

7.13. Guideline recommendations and divergences

Table 7.5. Divergences among guidance sources

The 2026 guideline was developed with a patient representative at every stage and funded by the Society for the Study of Inborn Errors of Metabolism (SSIEM), an Italian cblC patient organisation, and HCU Network Australia; some authors received support from MetabERN, and since 2019 the E-HOD project has received industry funding (Aeglea, GAIN, Nutricia Metabolics, Recordati, SOBI, Travere) managed by the University Children's Hospital Zurich [2]. Delphi consensus was defined as 75–95% and strong consensus as greater than 95% [2].

7.14. Pregnancy, lactation, perioperative care, and transition

Pregnancy and lactation. As of February 2016, no pregnancy had been published in women with MTHFR deficiency or with remethylation disorders other than cblC, and the 2017 guideline made no recommendations [17]. The 2026 guideline describes prenatal treatment only in the combined cobalamin disorders [2], and the 2026 pregnancy series included only cobalamin disorders [68]. GeneReviews recommends metabolic stability, close monitoring of tHcy and methionine during pregnancy, and joint follow-up by a metabolic specialist and high-risk obstetrics, and notes that elevated homocysteine may increase the risk of preeclampsia, placental abruption, and thromboembolism [1]. Regarding betaine, the US label reports limited data with no identified risk, and the European label advises against its use unless clearly necessary; the 2017 guideline describes betaine use in pregnancy without adverse neonatal effects in a patient without MTHFR deficiency [15–17]. There are no data on breastfeeding [15]. None of the sources consulted recommends stopping betaine in pregnancy, and the decision is individual and shared with the patient, weighing the maternal risk of untreated hyperhomocysteinaemia described by GeneReviews [1].

Anaesthesia and surgery. Nitrous oxide is contraindicated (section 7.8). In Selzer's case, an infant with undiagnosed MTHFR deficiency was exposed to 60% nitrous oxide for 45 minutes during a biopsy and for 270 minutes during a resection, and died 46 days after the second operation [33]. The authors recommend that patients with severe MTHFR deficiency not receive nitrous oxide, that the same apply in emergencies to patients whose clinical picture suggests the diagnosis, and that the diagnosis be excluded beforehand in elective surgery [33, 71]. Propofol has been used safely for induction and maintenance in short procedures in 7 stable patients with cblC and 1 with cblG; there is no recommendation on the best anaesthetic approach in MTHFR deficiency [2, 72].

Transition to adult care. Saudubray emphasises that follow-up of early-onset patients reaching adulthood raises specific medical, dietary, and social issues, and that many adults with late-onset forms were diagnosed after years of symptoms [43]. In Marelli's adult series, the mean diagnostic delay was 14 years [11]. Transition should ensure continuity of betaine and folates, assessment of renal function, neuropsychiatric and ophthalmological surveillance, and reproductive counselling [2].

8. Reproductive Options and Family Counselling

8.1. Recurrence risks

Autosomal recessive inheritance determines the usual risks: the parents of an affected individual are presumed to be heterozygotes, and testing them is recommended to confirm the phase of the variants [1]. Each sibling of an affected individual has a 25% chance of being affected, a 50% chance of being an asymptomatic carrier, and a 25% chance of being neither affected nor a carrier [1]. The children of an affected individual are obligate heterozygotes, unless the other parent is a carrier or affected; the siblings of an affected individual's parents have a 50% chance of being carriers [1]. Heterozygotes are asymptomatic and are not at risk of developing the disease [1].

8.2. Special genetic situations

If only one parent is a carrier and paternity is confirmed, the second variant may have arisen de novo in the proband or result from postzygotic parental mosaicism [1]. Apparent homozygosity may be due to a deletion of one or more exons on the other allele, not detected by the method used, or to uniparental isodisomy, scenarios in which the recurrence risk is different [1]. These possibilities should be clarified before calculating the risk for other relatives.

8.3. Cascade testing and carrier detection

Carrier detection in relatives requires that the familial variants have been identified [1]. GeneReviews recommends testing the reproductive partners of carriers, especially when there is consanguinity or shared ancestry, as in Bukharan Jews and the Amish [1]. In Bukharan Jews, with a c.474A>T carrier frequency of 1:39, the authors of the founder study proposed population carrier screening to allow prenatal or preimplantation diagnosis [8]. Preconception and prenatal carrier screening panels that include remethylation disorder genes are already offered [2].

8.4. Prenatal and preimplantation diagnosis and the at-risk newborn

Once the variants are identified, prenatal diagnosis is performed on chorionic villi or amniotic fluid, and preimplantation genetic testing (PGT-M) is possible [1, 2]. Prenatal diagnosis has therapeutic as well as reproductive value: it allows treatment to start at birth, and early betaine treatment after prenatal diagnosis has given the best results [1, 12]. There is no evidence for prenatal treatment in MTHFR deficiency; maternal in utero therapy has been described only in the combined cobalamin disorders [2, 17]. For a newborn sibling of an affected individual without prenatal diagnosis, GeneReviews recommends immediate nutritional and metabolic management while awaiting tHcy, methionine, and the familial genotype [1]. The nine presymptomatic patients in the Yverneau cohort were diagnosed because they had an affected sibling, and in two cases the sibling treated late had a worse course [10].

8.5. Counselling of minors and communication

Carrier testing in asymptomatic minor siblings of an affected individual differs from diagnostic testing: a sibling at risk of being affected should be tested soon after birth or prenatally, because presymptomatic treatment changes the prognosis [1, 10], whereas determination of carrier status is mainly of reproductive interest. Planning before pregnancy and discussion of differing perspectives on prenatal diagnosis are part of counselling [1].

8.6. Common polymorphisms: thrombophilia, pregnancy loss, and neural tube defects

The 2013 ACMG guidance, reaffirmed by the Board of Directors on 27 April 2020 and reclassified as a clinical practice resource, sets out five points: genotyping of MTHFR polymorphisms should not be ordered as part of a thrombophilia or recurrent pregnancy loss evaluation; it should not be ordered for at-risk relatives; when a polymorphism is found, the geneticist should ensure that the patient has had a complete evaluation of their symptoms; in a c.665C>T homozygote, fasting tHcy may be ordered for more precise counselling; and the genotype does not change the recommendation that women of reproductive age take the standard dose of folic acid recommended for the general population to reduce the risk of neural tube defects [18, 19].

The data supporting these points are consistent. Thermolabile homozygotes with normal tHcy have no demonstrated increased risk of venous thromboembolism or recurrent pregnancy loss; with elevated tHcy, the risk is slightly increased (OR 1.27 for venous thromboembolism; pooled risk of 2.7 for recurrent pregnancy loss) [18]. In people with known thrombophilia (factor V Leiden, prothrombin variant c.*97G>A), the MTHFR genotype does not alter risk in a clinically significant way, and there is no contraindication to oral contraceptives [18]. B-vitamin supplementation lowers homocysteine without reducing cardiovascular events or mortality, and the genotype is not associated with mortality [18]. According to the ACMG, the American College of Obstetricians and Gynecologists (ACOG) does not recommend measuring homocysteine or MTHFR polymorphisms in the etiological evaluation of venous thromboembolism, and the British Committee for Standards in Haematology does not include MTHFR in its guidelines on inherited thrombophilia [18]. The Oxford Desk Reference cites elevated plasma homocysteine as an additional risk factor in factor V Leiden carriers, without recommending MTHFR genotyping [73].

For neural tube defects, women homozygous for c.665C>T have a modestly increased risk of affected offspring (OR 1.6), higher if the fetus is also homozygous [18]. The Oxford Desk Reference describes a risk about 1.8 times higher with the c.665C>T and c.1286A>C polymorphisms, detected only in non-Hispanic populations, according to the meta-analysis by the Latin American ECLAMC group (Latin American Collaborative Study of Congenital Malformations), and considers that analysis of predisposition genes is not used in clinical practice [74, 75]. The recommended prevention is independent of genotype: 400 µg/day of folic acid from at least one month before conception until week 12 for a first occurrence, and 5 mg/day over the same period when there is a history of a neural tube defect; in the 1991 Medical Research Council trial, 4 mg/day prevented about 70% of recurrences [74]. This recommendation, which applies to the general population, does not carry over to women with severe MTHFR deficiency, in whom folic acid should be avoided and replaced by methylfolate or folinic acid [2, 17].

Other clinical situations in which the genotype may be relevant include folate-dependent chemotherapy: the ACMG considers it reasonable to report the genotype before therapies such as methotrexate [18]. Daily B-complex vitamin supplementation, in a multivitamin or prenatal supplement, is optional and low-risk, but supplemental pyridoxine carries a risk of ataxia and sensory neuropathy [18].

9. Special Considerations: Population Diversity and Epidemiology

9.1. Prevalence and what each figure measures

The prevalence of severe MTHFR deficiency is unknown [3]. Orphanet assigns no numerical value [3], and the available estimates are counts of published cases: fewer than 200 cases by 2016 [6], 148 cases underlying the clinical characterisation in the 2017 guideline [17], 109 variants in 171 families in the Froese mutational review [7], more than 200 patients described according to Saudubray [12], and 150 individuals with biallelic variants recorded in HGMD according to GeneReviews [1]. GeneReviews estimates a prevalence of about 1/10,000–1/50,000, varying by population, but the references it cites for this figure deal mainly with the common polymorphisms and not with population screening or registries of the severe disease [1]. GeneGuidance adopts 2/100,000, the lower bound of that estimate, noting the conflict with Orphanet [40]. None of these estimates has been validated by population-based newborn screening in the sources consulted, and the apparent rarity also reflects underdiagnosis of late-onset forms, with mean diagnostic delays of 14 years in adults [11].

The frequencies of the common polymorphisms measure something else: the proportion of c.665C>T homozygotes, above 25% in Hispanic people and 10–15% in North American White people, and the allele frequency of about 38% in the Dutch and Canadian sample of the Frosst study bear no relation to the prevalence of the severe disease [18, 26].

9.2. Founder populations and geographic distribution

In Bukharan Jews, the c.474A>T carrier frequency is 1:39 [8]. In the Old Order Amish of Pennsylvania, the c.1129C>T p.(Arg377Cys) variant is common, and the Strauss series described five homozygous children [9, 35]. The seven Saudi patients of Aljassim were homozygous for c.680C>T p.(Thr227Met) [32]. The 2016 Huemer cohort illustrates the diversity of origins of patients studied in the European reference laboratory: 8 Turkish, 6 from Central Europe, 5 Pakistani, 5 from Southern Europe (Spain and Italy), 3 from Eastern Europe (Poland and Russia), 2 from Northern Europe (United Kingdom), 2 from the Near East, and 1 Japanese, with consanguinity in 13 families [6]. In the Yverneau early-onset cohort, 17 of 30 patients were White, 6 Asian, 4 Black, and 3 Arab [10]. There are also series from Pakistan, Turkey, and Ireland [42, 76, 77].

9.3. United States, Europe, Portugal, and Brazil

In the USA, MTHFR deficiency is not part of the RUSP; the federal advisory committee that maintained the panel was terminated in March 2025, and the list consulted is that of January 2023 [13]. Each state defines its own panel, and GeneReviews describes the positive-screen scenario with cutoffs set by each laboratory [1]. In Europe, the Yverneau cohort noted that few programmes detected the disease in 2022 [10], and the 2026 E-HOD guideline recommends screening [2].

In Portugal, the National Newborn Screening Programme includes classic homocystinuria and hypermethioninaemia due to MAT I/III deficiency and since 2017 has used second-tier markers for classic homocystinuria (methionine followed by tHcy) and for cblC/cblD (C3 and methionine, followed by MMA and tHcy), but MTHFR deficiency is not among the screened disorders [14]. For comparison, the 2022 report records a cumulative birth prevalence of 1:396,162 for classic homocystinuria (4 cases) [14]. The Coimbra and Porto centres participate in the E-HOD registry, and an author from Coimbra is a member of the 2026 guideline panel, but the sources consulted present no national data on patients with MTHFR deficiency [2, 41]. In Brazil, the Porto Alegre centre participates in the E-HOD registry, the reference centre of the Federal University of São Paulo is part of the Yverneau cohort consortium, and Brazilian colleagues are acknowledged in the 2026 guideline; no Brazilian screening or prevalence data for MTHFR deficiency were found in the sources consulted [2, 10, 41]. The 2020 Brazilian clinical protocol for homocystinuria refers to classic homocystinuria and not to MTHFR deficiency [40]. The ECLAMC meta-analysis on the c.665C>T polymorphism and neural tube defects is the main Latin American data point in the sources, and it does not concern the severe disease [75].

9.4. Special populations

Women of reproductive age with severe MTHFR deficiency need counselling about the absence of pregnancy data and about replacing folic acid with methylfolate or folinic acid (sections 7.4 and 7.14) [2, 17]. Adults with psychiatric symptoms or spastic paraparesis of unexplained cause are a population in which the disease is underdiagnosed [11, 36]. Newborns of mothers with B12 deficiency, especially strict vegans, are the main confounder in screening and early diagnosis [2, 17]. Patients undergoing anaesthesia or sedation with nitrous oxide are the population in which a missed diagnosis has the most serious consequences [33].

9.5. Reference networks and patient organisations

The relevant European network is MetabERN, the European Reference Network for Hereditary Metabolic Disorders (https://metab.ern-net.eu/), which supports some of the authors of the 2026 guideline [2, 20]. E-HOD maintains the international registry of patients with homocystinurias and methylation defects and produced both guidelines (https://e-hod.org/) [21]. Among patient organisations, HCU Network America (https://hcunetworkamerica.org/), in the USA, covers the homocystinurias, including severe MTHFR deficiency, with a dedicated committee, information on newborn screening, meetings, and a patient registry [69]; HCU Network Australia (https://www.hcunetworkaustralia.org.au/) focuses on classic homocystinuria but co-funds research on MTHFR deficiency [70]; and Metabolic Support UK (https://metabolicsupport.org/) supports patients with metabolic disorders in the United Kingdom [78]. The sources consulted identify no Portuguese or Brazilian organisations specific to MTHFR deficiency. The 2026 guideline survey of 24 patients or caregivers with remethylation disorders and 110 with classic homocystinuria showed that almost half considered treatment only partially satisfactory, with problems in access to medication, financial burden, and bureaucracy, and that few families had psychological support or information about support groups [2].

10. Illustrative Clinical Case

10.1. Neonatal-onset severe MTHFR deficiency due to a homozygous nonsense variant (Massadeh and colleagues, 2019)

In 2019, Massadeh and colleagues described a family followed in Riyadh, Saudi Arabia, with seven pregnancies: three deceased children, healthy twin girls, one miscarriage, and the proband, the seventh child (pedigree in Figure 1A of the article) [5]. The three deceased siblings had died at 40 days, 2 months, and 45 days of life, after becoming lethargic and hypoactive, with weak suck and prolonged somnolence; they were admitted to neonatal intensive care as cases of neonatal sepsis, computed tomography showed subarachnoid haemorrhage, and death was due to apnoea, disseminated intravascular coagulation, and cardiorespiratory arrest [5].

The proband had growth restriction in the last month of gestation; at birth, head circumference was 31.5 cm and weight 2980 g, with microcephaly documented in a photograph at 1 month (Figure 1B) [5]. At 14 days of life, apnoeas, bradypnoea, and swallowing difficulty appeared. Ophthalmological examination showed bilateral optic nerve hypoplasia and multiple retinal haemorrhages on the left, with a normal anterior segment; leucocytes and platelets were low, with lymphocytosis [5]. Brain MRI (Figure 1C) showed bilateral, symmetric T2 hyperintensity of the periventricular and deep frontal and subcortical temporal white matter, markedly dilated ventricles, white matter loss, and prominent extra-axial spaces [5]. tHcy was 163 µmol/L (reference 3.7–12.5) and methionine below 5 µmol/L (reference 10–39) [5].

Exome sequencing in five family members identified in the proband the homozygous variant NM_005957.5:c.154C>T p.(Arg52Ter), in exon 1, in the catalytic domain, with heterozygous parents; the variant was absent from the population databases consulted by the authors, and enzyme activity was not measured [5, 28]. Figure 2 of the article shows the gene structure, the protein domains, and the cross-species conservation of the Arg52 residue [5]. Treatment included betaine 100 mg/kg/day, methionine 50 mg/kg/day, oral vitamin B12 5 mg/day, and oral L-5-MTHF 15 mg/day [5]. At follow-up evaluation (Figure 1D), tHcy had fallen to 68 µmol/L and methionine was 54 µmol/L; the child was vocalising at 4–5 months, with a length of 79 cm, weight of 10.7 kg, and head circumference of 48 cm, all above the 95th percentile according to the authors, but optic atrophy and retinal haemorrhage persisted, with no visual fixation or tracking [5].

The figures consulted are available in open access: Figure 1, with pedigree, photographs, and MRI (https://pmc.ncbi.nlm.nih.gov/articles/PMC6491806/figure/F1/), and Figure 2, with gene and protein structure (https://pmc.ncbi.nlm.nih.gov/articles/PMC6491806/figure/F2/) [5].

10.2. Interpreting the case in light of the guidelines

The case brings together the features that guide early diagnosis: siblings who died in the neonatal period with a diagnosis of sepsis, central apnoea in the first weeks, microcephaly, white matter abnormalities, very high tHcy, and almost undetectable methionine [5, 17]. The nonsense variant in the catalytic domain belongs to the group with very low residual activity, associated with early onset and apnoea [6, 29]. The haematological abnormalities (low leucocytes and platelets) correspond to the neutropenia or pancytopenia described in 10–25% of early-onset patients [2]. The persistent severe eye disease is consistent with the 48% of ocular complications not improved by treatment in the Yverneau cohort [10].

The treatment regimen differs in part from current guidelines. Betaine at 100 mg/kg/day matches the European SmPC dose and lies at the lower limit of the 100–250 mg/kg/day range recommended in 2026 [2, 16]; the L-5-MTHF dose of 15 mg/day is lower than the 45 mg/day often used in children according to the guideline [2]; and oral vitamin B12 is in line with the recommendation for oral cobalamin in confirmed MTHFR deficiency [2]. Methionine at 50 mg/kg/day corresponds to the upper limit of the GeneReviews dose, and the methionine level of 54 µmol/L at follow-up was above the laboratory reference range, which illustrates the need to titrate methionine supplementation to the plasma concentration [1, 2]. The diagnosis, made after symptom onset and despite three deceased siblings, shows what prenatal diagnosis or treatment from birth of an at-risk sibling could have changed [1, 10].

10.3. Contrast: deterioration after nitrous oxide (Selzer and colleagues, 2003)

Selzer and colleagues described an infant who appeared normal until 3 months of age, when a mass was noted in the left leg; before surgery it was not known that the father and an uncle had tHcy above 20 and 30 µmol/L, respectively [33]. The biopsy, under halothane and 60% nitrous oxide for 45 minutes, revealed infantile fibrosarcoma, and the resection, four days later, was performed with halothane and 60% nitrous oxide for 270 minutes; the child was discharged on postoperative day 7, apparently well [33]. Seventeen days later, seizures and apnoeas appeared, with severe hypotonia, areflexia, and ataxic breathing; computed tomography showed generalised cerebral atrophy, plasma and urinary homocystine were elevated without MMA, methionine was 0.06 mg/dL (normal 0.48 ± 0.18), and B12 was normal [33]. The child died at 130 days of life, 46 days after the resection, of respiratory arrest; autopsy showed asymmetric cerebral atrophy and severe demyelination with astrogliosis in the midbrain, medulla, and cerebellum [33]. MTHFR activity in fibroblasts was severely reduced, and the genotype was atypical: c.1743G>A p.(Met581Ile) in cis with c.665C>T on the paternal allele and c.1286A>C on the maternal allele [28, 33]. This case underpins the high-quality recommendation against the use of nitrous oxide in any remethylation disorder [2, 71].

11. Prognosis and Long-Term Follow-Up

11.1. Prognosis

Prognosis depends mainly on age at onset and the timing of treatment [2]. Early-onset disease is associated with a high risk of death in the first months of life, and mortality is nil or very low in presentations in adolescence and adulthood [2]. In the Diekman meta-analysis, 31% of patients died, all untreated or treated late, whereas all those treated early survived with normal development [27]. In the Yverneau cohort, one of 32 early-onset patients died during follow-up (cardiomyopathy at 39 months), but 76% of those treated had a severe neurodevelopmental outcome; overall status was considered stable in 14/21 and improved in 4/21 [10]. The nine patients diagnosed before symptom onset all had normal cognitive development and neurological examination [10]. In late-onset adults, symptoms improved in 9 and stabilised in 4 of 13 treated patients [11], and three ambulatory patients from the Lossos families improved over 9–15 years of betaine [36]. GeneReviews states that, with early initiation of betaine, methionine, B12, and L-5-MTHF, many patients can expect a near-normal life span, and that stroke is an important cause of mortality [1].

Deficits established before treatment tend to persist. In most patients, microcephaly does not progress after treatment is started, but it rarely normalises [1]. Eye disease and cognitive impairment respond less to treatment than overall morbidity [10, 41], and psychiatric manifestations increase with age with no apparent influence of treatment [2]. Acute symptoms in adults (confusion, psychiatric decompensation, seizures, respiratory failure) are those that improve most with betaine [2, 11]. Quality of life has been little studied [2].

11.2. Surveillance

Table 11.1. Recommended surveillance in severe MTHFR deficiency

The instruments suggested by the 2026 guideline for standardised assessment, based on expert opinion, include Griffiths-III, WPPSI-III, WISC-IV, WAIS-IV, and Leiter-3 (nonverbal) for development and intelligence; NEPSY-II, MoCA, and MMSE for other cognitive functions; BRIEF, TMT, and WCST for executive functions; VMI and the Rey complex figure for visuoperceptual functions; Vineland and ABAS-II for adaptive behaviour; CBCL, ASR/ABCL (18–59 years), SCL-90-R, and K-SADS-PL for behaviour; PedsQL, DISABKIDS, and SF-36 for quality of life; VQoL, FVQ, PedEyeQ, and NEI VFQ-25 for vision; and PedsQL-FIM, Zarit, and PSI-SF for caregivers [2].

11.3. Long-term follow-up and quality of life

Follow-up is provided by a metabolic disease centre with a multidisciplinary team, preferably part of MetabERN, including neurology, ophthalmology, nutrition, psychology, and rehabilitation [1, 20]. Daily adherence to betaine in 2–3 divided doses and to folates is the basis of lifelong treatment, and the 2026 guideline survey shows that problems with access to medication, financial burden, and insufficient psychological support weigh on the experience of patients and families [2]. Enrolling patients in the E-HOD registry contributes to the data on treated natural history that underpin the recommendations [21, 41].

12. Summary and Concluding Remarks

Severe MTHFR deficiency is a rare autosomal recessive disorder of unknown prevalence, caused by biallelic loss-of-function variants in MTHFR, most of them private, with founder variants in Bukharan Jews and the Amish [7–9]. Lack of 5-methyltetrahydrofolate limits homocysteine remethylation and causes severe hyperhomocysteinaemia, low or normal methionine, and cerebral folate depletion, without methylmalonic aciduria and, in most patients, without megaloblastic anaemia [2, 12]. The early-onset form presents in the first weeks with feeding difficulties, lethargy, hypotonia, apnoeas, and seizures, with high mortality if untreated; the late-onset form presents from childhood to adulthood with spastic paraparesis, cognitive decline, epilepsy, psychosis, polyneuropathy, visual impairment, and thrombosis, often after years of diagnostic delay [10, 11].

Diagnosis begins with plasma total homocysteine, collected and processed with the preanalytical precautions described, and with pretreatment sampling for methionine, folate, B12, and MMA; it is confirmed by sequencing with parental testing and, if necessary, by fibroblast enzyme activity [2]. The main pitfalls are confusion with methionine synthase defects and with B12 or folate deficiency, including maternal deficiency in the newborn; the historical variant numbering, shifted by 12 nucleotides relative to NM_005957.5; and the most frequent variant in the disease, c.1530G>A, which is synonymous and acts through splicing [7, 17, 28].

Betaine is the treatment of choice and, when started before symptom onset, is associated with survival and normal neurodevelopment; the 2026 E-HOD guideline recommends 100–250 mg/kg/day in children and 6–9 g/day in adolescents and adults (the label specifies 100 mg/kg/day in 2 doses, titrated, in the European SmPC, or 6 g/day from age 3 years in the FDA label) [2, 15, 16], with methylfolate or folinic acid in all patients, oral cobalamin, L-methionine when methionine is not maintained in the normal range, and avoidance of folic acid and nitrous oxide [2]. The guideline also recommends newborn screening, still absent from the US and Portuguese panels, and the evidence for this recommendation comes from retrospective cohorts, with no dedicated randomised trials [2, 10, 13, 14]. The GeneReviews doses, the claim that no guidelines exist, ectopia lentis as a characteristic manifestation, the pseudogenes, and the prevalence of 1/10,000–1/50,000 are not supported by the primary sources consulted [1, 2, 4, 47].

The common polymorphisms c.665C>T and c.1286A>C are a different matter: they do not cause the disease, should not be tested in the evaluation of thrombophilia, pregnancy loss, or relatives, and do not change the recommendation of standard-dose folic acid to prevent neural tube defects [18, 19]. Patients with severe deficiency are followed in MetabERN reference centres, with biochemical monitoring every 3–4 months, regular neurological, ophthalmological, and neuropsychological assessment, and genetic counselling that includes prenatal diagnosis and treatment from birth of at-risk siblings [1, 2, 20].

13. References

1. Umair M, Alfadhel M. Homocystinuria due to Deficiency of N(5,10)-Methylenetetrahydrofolate Reductase Activity. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Published 29 May 2025; revised 4 Dec 2025. NBK615089.
2. Olivieri G, Bordugo A, Burnyte B, Costetti A, Couce ML, Diogo L, et al. First Revision of the Guidelines for the Diagnosis and Management of Remethylation Disorders. J Inherit Metab Dis. 2026;49(4):e70177. doi:10.1002/jimd.70177. PMID: 42231716.
3. Orphanet. Homocystinuria due to methylene tetrahydrofolate reductase deficiency. ORPHA:395 [Internet]. Paris: INSERM; Orphadata data of 23 Jun 2026. Disease page. Accessed 17 Oct 2026.
4. HUGO Gene Nomenclature Committee. MTHFR — symbol report (1p36.22; MANE Select NM_005957.5/ENST00000376590.9) [Internet]. genenames.org. Accessed 17 Oct 2026.
5. Massadeh S, Umair M, Alaamery M, Alfadhel M. A Novel Homozygous Non-sense Mutation in the Catalytic Domain of MTHFR Causes Severe 5,10-Methylenetetrahydrofolate Reductase Deficiency. Front Neurol. 2019;10:411. doi:10.3389/fneur.2019.00411. PMID: 31068897.
6. Huemer M, Mulder-Bleile R, Burda P, Froese DS, Suormala T, Zeev BB, et al. Clinical pattern, mutations and in vitro residual activity in 33 patients with severe 5, 10 methylenetetrahydrofolate reductase (MTHFR) deficiency. J Inherit Metab Dis. 2016;39(1):115–24. doi:10.1007/s10545-015-9860-6. PMID: 26025547.
7. Froese DS, Huemer M, Suormala T, Burda P, Coelho D, Guéant JL, et al. Mutation Update and Review of Severe Methylenetetrahydrofolate Reductase Deficiency. Hum Mutat. 2016;37(5):427–38. doi:10.1002/humu.22970. PMID: 26872964.
8. Ben-Shachar S, Zvi T, Rolfs A, Breda Klobus A, Yaron Y, Bar-Shira A, et al. A founder mutation causing a severe methylenetetrahydrofolate reductase (MTHFR) deficiency in Bukharian Jews. Mol Genet Metab. 2012;107(3):608–10. doi:10.1016/j.ymgme.2012.08.011. PMID: 22947400.
9. Strauss KA, Morton DH, Puffenberger EG, Hendrickson C, Robinson DL, Wagner C, et al. Prevention of brain disease from severe 5,10-methylenetetrahydrofolate reductase deficiency. Mol Genet Metab. 2007;91(2):165–75. doi:10.1016/j.ymgme.2007.02.012. PMID: 17409006.
10. Yverneau M, Leroux S, Imbard A, Gleich F, Arion A, Moreau C, et al. Influence of early identification and therapy on long-term outcomes in early-onset MTHFR deficiency. J Inherit Metab Dis. 2022;45(4):848–61. doi:10.1002/jimd.12504. PMID: 35460084. Full manuscript at HAL hal-03711581.
11. Marelli C, Lavigne C, Stepien KM, Janssen MCH, Feillet F, Kožich V, et al. Clinical and molecular characterization of adult patients with late-onset MTHFR deficiency. J Inherit Metab Dis. 2021;44(3):777–86. doi:10.1002/jimd.12323. PMID: 33089527.
12. Fowler B, Froese DS, Watkins D. Disorders of Cobalamin and Folate Transport and Metabolism. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 28. doi:10.1007/978-3-662-63123-2_28.
13. Health Resources and Services Administration. Recommended Uniform Screening Panel (primary and secondary conditions, January 2023 list; archived page, reviewed July 2024) [Internet]. hrsa.gov. Accessed 17 Oct 2026.
14. Vilarinho L, Garcia P, et al.; Comissão Executiva do Programa Nacional de Rastreio Neonatal. Programa Nacional de Rastreio Neonatal: relatório 2022 [National Newborn Screening Programme: 2022 report]. Porto: Instituto Nacional de Saúde Doutor Ricardo Jorge. INSA repository. Accessed 17 Oct 2026.
15. Recordati Rare Diseases. CYSTADANE (betaine anhydrous for oral solution): prescribing information. NDA 020576; SPL version of 24 Apr 2026. Silver Spring (MD): U.S. Food and Drug Administration; DailyMed. Label. Accessed 17 Oct 2026.
16. European Medicines Agency. Cystadane (betaine anhydrous): EPAR and summary of product characteristics. EU/1/06/379/001; authorised 15 Feb 2007; renewed 21 Nov 2016. EPAR. Accessed 17 Oct 2026.
17. Huemer M, Diodato D, Schwahn B, Schiff M, Bandeira A, Benoist JF, et al. Guidelines for diagnosis and management of the cobalamin-related remethylation disorders cblC, cblD, cblE, cblF, cblG, cblJ and MTHFR deficiency. J Inherit Metab Dis. 2017;40(1):21–48. doi:10.1007/s10545-016-9991-4. PMID: 27905001.
18. Hickey SE, Curry CJ, Toriello HV. ACMG Practice Guideline: lack of evidence for MTHFR polymorphism testing. Genet Med. 2013;15(2):153–6. doi:10.1038/gim.2012.165. PMID: 23288205.
19. Bashford MT, Hickey SE, Curry CJ, Toriello HV. Addendum: ACMG Practice Guideline: lack of evidence for MTHFR polymorphism testing. Genet Med. 2020;22(12):2125. doi:10.1038/s41436-020-0843-0. PMID: 32533132.
20. European Reference Network for Hereditary Metabolic Disorders (MetabERN) [Internet]. metab.ern-net.eu. Accessed 17 Oct 2026.
21. E-HOD — European network and registry for homocystinurias and methylation defects [Internet]. e-hod.org. Accessed 17 Oct 2026.
22. Mudd SH, Uhlendorf BW, Freeman JM, Finkelstein JD, Shih VE. Homocystinuria associated with decreased methylenetetrahydrofolate reductase activity. Biochem Biophys Res Commun. 1972;46(2):905–12. doi:10.1016/s0006-291x(72)80227-4. PMID: 5057914.
23. Freeman JM, Finkelstein JD, Mudd SH. Folate-responsive homocystinuria and "schizophrenia". A defect in methylation due to deficient 5,10-methylenetetrahydrofolate reductase activity. N Engl J Med. 1975;292(10):491–6. doi:10.1056/NEJM197503062921001. PMID: 1117892.
24. Saudubray JM, García-Cazorla Á. Clinical Approach to Inborn Errors of Metabolism in Paediatrics. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 1. doi:10.1007/978-3-662-63123-2_1.
25. Goyette P, Sumner JS, Milos R, Duncan AM, Rosenblatt DS, Matthews RG, et al. Human methylenetetrahydrofolate reductase: isolation of cDNA, mapping and mutation identification. Nat Genet. 1994;7(2):195–200. doi:10.1038/ng0694-195. PMID: 7920641.
26. Frosst P, Blom HJ, Milos R, Goyette P, Sheppard CA, Matthews RG, et al. A candidate genetic risk factor for vascular disease: a common mutation in methylenetetrahydrofolate reductase. Nat Genet. 1995;10(1):111–3. doi:10.1038/ng0595-111. PMID: 7647779.
27. Diekman EF, de Koning TJ, Verhoeven-Duif NM, Rovers MM, van Hasselt PM. Survival and psychomotor development with early betaine treatment in patients with severe methylenetetrahydrofolate reductase deficiency. JAMA Neurol. 2014;71(2):188–94. doi:10.1001/jamaneurol.2013.4915. PMID: 24323041.
28. 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):61–8. doi:10.1002/humu.23348. Service rest.variantvalidator.org (GRCh38, MANE Select) accessed 17 Oct 2026.
29. Burda P, Schäfer A, Suormala T, Rummel T, Bürer C, Heuberger D, et al. Insights into severe 5,10-methylenetetrahydrofolate reductase deficiency: molecular genetic and enzymatic characterization of 76 patients. Hum Mutat. 2015;36(6):611–21. doi:10.1002/humu.22779. PMID: 25736335.
30. Ibn-Salem J, Köhler S, Love MI, Chung HR, Huang N, Hurles ME, et al. Deletions of chromosomal regulatory boundaries are associated with congenital disease. Genome Biol. 2014;15(9):423. doi:10.1186/s13059-014-0423-1. PMID: 25315429.
31. National Center for Biotechnology Information. ClinVar [Internet]. Bethesda (MD): NCBI. Variation ID 3520, 3521, 3528, 800827, 466270, and 1066379. ClinVar. Accessed 17 Oct 2026.
32. Aljassim N, Alfadhel M, Nashabat M, Eyaid W. Clinical presentation of seven patients with Methylenetetrahydrofolate reductase deficiency. Mol Genet Metab Rep. 2020;25:100644. doi:10.1016/j.ymgmr.2020.100644. PMID: 32939339.
33. Selzer RR, Rosenblatt DS, Laxova R, Hogan K. Adverse effect of nitrous oxide in a child with 5,10-methylenetetrahydrofolate reductase deficiency. N Engl J Med. 2003;349(1):45–50. doi:10.1056/NEJMoa021867. PMID: 12840091.
34. Li YX, Wang MQ, Wang RY, Hao YL. Novel MTHFR variants manifesting with hereditary spastic paraplegia and recurrent pulmonary embolism: a case report and literature review of adult-onset severe MTHFR deficiency. Neurol Sci. 2026;47(1):178. doi:10.1007/s10072-025-08801-8. PMID: 41557084.
35. Strauss KA, Puffenberger EG. Genetics, medicine, and the Plain people. Annu Rev Genomics Hum Genet. 2009;10:513–36. doi:10.1146/annurev-genom-082908-150040. PMID: 19630565.
36. Lossos A, Teltsh O, Milman T, Meiner V, Rozen R, Leclerc D, et al. Severe methylenetetrahydrofolate reductase deficiency: clinical clues to a potentially treatable cause of adult-onset hereditary spastic paraplegia. JAMA Neurol. 2014;71(7):901–4. doi:10.1001/jamaneurol.2014.116. PMID: 24797679.
37. Weile J, Kishore N, Sun S, Maaieh R, Verby M, Li R, et al. Shifting landscapes of human MTHFR missense-variant effects. Am J Hum Genet. 2021;108(7):1283–300. doi:10.1016/j.ajhg.2021.05.009. PMID: 34214447.
38. Tsang BL, Devine OJ, Cordero AM, Marchetta CM, Mulinare J, Mersereau P, et al. Assessing the association between the methylenetetrahydrofolate reductase (MTHFR) 677C>T polymorphism and blood folate concentrations: a systematic review and meta-analysis of trials and observational studies. Am J Clin Nutr. 2015;101(6):1286–94. doi:10.3945/ajcn.114.099994. PMID: 25788000.
39. Nyhan WL, Hoffmann GF, Al-Aqeel AI, Barshop BA. Cobalamin C, D, F, G diseases; methylmalonic aciduria and variable homocystinuria. In: Atlas of Inherited Metabolic Diseases. 4th ed. Boca Raton: CRC Press; 2020. p. 39–40.
40. GeneGuidance. MTHFR — clinical curation of the gene and associated conditions [database; version of 26 Sep 2026].
41. Huemer M, Diodato D, Martinelli D, Olivieri G, Blom H, Gleich F, et al. Phenotype, treatment practice and outcome in the cobalamin-dependent remethylation disorders and MTHFR deficiency: Data from the E-HOD registry. J Inherit Metab Dis. 2019;42(2):333–52. doi:10.1002/jimd.12041. PMID: 30773687.
42. Karalar Pekuz OK, Teke Kisa P, Yildiz S, Kocabey M, Zubarioglu T, Arslan Gulten Z, et al. Clinical spectrum and genetic landscape of MTHFR deficiency: a cohort study including novel variants. Neurol Sci. 2026;47(10):833. doi:10.1007/s10072-026-09447-w. PMID: 42825826.
43. Mochel F, Sedel F. Inborn Errors of Metabolism in Adults: A Diagnostic Approach to Neurological and Psychiatric Presentations. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 2. doi:10.1007/978-3-662-63123-2_2.
44. Bönig H, Däublin G, Schwahn B, Wendel U. Psychotic symptoms in severe MTHFR deficiency and their successful treatment with betaine. Eur J Pediatr. 2003;162(3):200–1. doi:10.1007/s00431-002-1148-9. PMID: 12655429.
45. Pasquier F, Lebert F, Petit H, Zittoun J, Marquet J. Methylenetetrahydrofolate reductase deficiency revealed by a neuropathy in a psychotic adult. J Neurol Neurosurg Psychiatry. 1994;57(6):765–6. doi:10.1136/jnnp.57.6.765. PMID: 8006671.
46. Biesalski AS, Hoffjan S, Schneider R, Nguyen HP, Dekomien G, Lücke T, et al. Phoenix from the ashes: dramatic improvement in severe late-onset methylenetetrahydrofolate reductase (MTHFR) deficiency with a complete loss of vision. J Neurol. 2022;269(4):2206–9. doi:10.1007/s00415-021-10841-x. PMID: 34657180.
47. Couser NL, McClure J, Evans MW, Haines NR, Burden SK, Muenzer J. Homocysteinemia due to MTHFR deficiency in a young adult presenting with bilateral lens subluxations. Ophthalmic Genet. 2017;38(1):91–4. doi:10.3109/13816810.2016.1143017. PMID: 27046515.
48. Touati G, Mochel F, Artuch R. Diagnostic Procedures. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 3. doi:10.1007/978-3-662-63123-2_3.
49. Kožich V, Morris AAM, Blom HJ. Disorders of Sulfur Amino Acid Metabolism. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 20. doi:10.1007/978-3-662-63123-2_20.
50. Pickens CA, Courtney E, Isenberg SL, Cuthbert C, Petritis K. Multiplexing Homocysteine into First-Tier Newborn Screening Mass Spectrometry Assays Using Selective Thiol Derivatization. Clin Chem. 2023;69(5):470–81. doi:10.1093/clinchem/hvad007. PMID: 36920064.
51. Wong D, Tortorelli S, Bishop L, Sellars EA, Schimmenti LA, Gallant N, et al. Outcomes of four patients with homocysteine remethylation disorders detected by newborn screening. Genet Med. 2016;18(2):162–7. doi:10.1038/gim.2015.45. PMID: 25856670.
52. Schiff M, Mochel F, Dionisi-Vici C. Emergency Treatments. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 4. doi:10.1007/978-3-662-63123-2_4.
53. De Biase I, Gherasim C, La'ulu SL, Asamoah A, Longo N, Yuzyuk T. Laboratory evaluation of homocysteine remethylation disorders and classic homocystinuria: Long-term follow-up using a cohort of 123 patients. Clin Chim Acta. 2020;509:126–34. doi:10.1016/j.cca.2020.06.014. PMID: 32533987.
54. European Medicines Agency. Amversio (betaine anhydrous): EPAR and product information. EU/1/22/1640/001; SERB S.A.; authorised 5 May 2022. EPAR. Accessed 17 Oct 2026.
55. Morris AAM, Jones S. Medications Used in the Treatment of Inborn Errors of Metabolism. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 45. doi:10.1007/978-3-662-63123-2_45.
56. Valayannopoulos V, Schiff M, Guffon N, Nadjar Y, García-Cazorla A, Martinez-Pardo Casanova M, et al. Betaine anhydrous in homocystinuria: results from the RoCH registry. Orphanet J Rare Dis. 2019;14(1):66. doi:10.1186/s13023-019-1036-2. PMID: 30871635.
57. Schwahn BC, Scheffner T, Stepman H, Verloo P, Das AM, Fletcher J, et al. Cystathionine beta synthase deficiency and brain edema associated with methionine excess under betaine supplementation: Four new cases and a review of the evidence. JIMD Rep. 2020;52(1):3–10. doi:10.1002/jmd2.12092. PMID: 32154053.
58. Knowles L, Morris AA, Walter JH. Treatment with Mefolinate (5-Methyltetrahydrofolate), but Not Folic Acid or Folinic Acid, Leads to Measurable 5-Methyltetrahydrofolate in Cerebrospinal Fluid in Methylenetetrahydrofolate Reductase Deficiency. JIMD Rep. 2016;29:103–7. doi:10.1007/8904_2016_529. PMID: 26898294.
59. Knowles L, Morris AAM, Walter JH. Erratum to: Treatment with Mefolinate (5-Methyltetrahydrofolate), but Not Folic Acid or Folinic Acid, Leads to Measurable 5-Methyltetrahydrofolate in Cerebrospinal Fluid in Methylenetetrahydrofolate Reductase Deficiency. JIMD Rep. 2016;29:117. doi:10.1007/8904_2016_574. PMID: 27325425.
60. Teofarma. Lederfoline 15 mg comprimidos (folinato de cálcio) [Lederfoline 15 mg tablets (calcium folinate)]: summary of product characteristics. INFARMED, Infomed [cited 2026 Oct 26]. https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=4930&tipo_doc=rcm
61. Actavis Pharma. Hydroxocobalamin injection USP 1000 mcg/mL: prescribing information (ANDA 085998). Revised Jan 2017. DailyMed [cited 2026 Oct 26]. https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=a3087a39-f218-4f16-a1e7-17c39bb2eb69
62. RPH Pharmaceuticals AB. Neo-Cytamen 1000 micrograms/ml solution for injection (hydroxocobalamin): summary of product characteristics, PL36301/0011; revised 25 Jul 2024. electronic medicines compendium [cited 2026 Oct 26]. https://www.medicines.org.uk/emc/product/9746/smpc
63. Schiff M, Benoist JF, Tilea B, Royer N, Giraudier S, Ogier de Baulny H. Isolated remethylation disorders: do our treatments benefit patients? J Inherit Metab Dis. 2011;34(1):137–45. doi:10.1007/s10545-010-9120-8. PMID: 20490923.
64. ClinicalTrials.gov. Exploratory Clinical Study of Autologous Hematopoietic Stem Cell Transplantation for Neurological Damage Associated With Hereditary Homocysteine Remethylation Disorders. NCT07757685. First posted 11 Aug 2026. Record. Accessed 17 Oct 2026.
65. ClinicalTrials.gov. Betaine Metabolism of Patients With Homocystinuria. NCT02404337. Assistance Publique – Hôpitaux de Paris. Record. Accessed 17 Oct 2026.
66. Imbard A, Toumazi A, Magréault S, Garcia-Segarra N, Schlemmer D, Kaguelidou F, et al. Efficacy and pharmacokinetics of betaine in CBS and cblC deficiencies: a cross-over randomized controlled trial. Orphanet J Rare Dis. 2022;17(1):417. doi:10.1186/s13023-022-02567-4. PMID: 36376887.
67. ClinicalTrials.gov. Investigation the Incidence of Methylenetetrahydrofolate Reductase Deficiency in Ischemic Cerebrovascular Diseases Presenting With Epileptic Seizures in Adults. NCT04287881. Record. Accessed 17 Oct 2026.
68. Stepien KM, Sykut-Cegielska J, Sloan J, Piazzon F, Bick S, Bouchereau J, et al. Cobalamin-Related Remethylation Disorders: Pregnancy Outcomes and Prenatal Treatment — New Cases and a Literature Study. J Inherit Metab Dis. 2026;49(5):e70243. doi:10.1002/jimd.70243. PMID: 42644808.
69. HCU Network America [Internet]. Batavia (IL). hcunetworkamerica.org. Accessed 17 Oct 2026.
70. HCU Network Australia [Internet]. hcunetworkaustralia.org.au. Accessed 17 Oct 2026.
71. Erbe RW, Salis RJ. Severe methylenetetrahydrofolate reductase deficiency, methionine synthase, and nitrous oxide — a cautionary tale. N Engl J Med. 2003;349(1):5–6. doi:10.1056/NEJMp030093. PMID: 12840086.
72. Ktena YP, Ramstad T, Baker EH, Sloan JL, Mannes AJ, Manoli I, et al. Propofol administration in patients with methylmalonic acidemia and intracellular cobalamin metabolism disorders: a review of theoretical concerns and clinical experiences in 28 patients. J Inherit Metab Dis. 2015;38(5):847–53. doi:10.1007/s10545-015-9816-x. PMID: 25985870.
73. Firth HV, Hurst JA. Thrombophilia. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 530–3.
74. Firth HV, Hurst JA. Neural tube defects. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 500–5.
75. Amorim MR, Lima MA, Castilla EE, Orioli IM. Non-Latin European descent could be a requirement for association of NTDs and MTHFR variant 677C > T: a meta-analysis. Am J Med Genet A. 2007;143A(15):1726–32. doi:10.1002/ajmg.a.31812. PMID: 17618486.
76. Ahmed S, Akbar F, DeBerardinis RJ, Ni M, Afroze B. Evaluation of the clinical, biochemical, and genetic presentation of neonatal and adult-onset 5,10-methylene tetrahydrofolate reductase (MTHFR) deficiency in patients from Pakistan. J Pediatr Endocrinol Metab. 2023;36(8):761–71. doi:10.1515/jpem-2023-0083. PMID: 37440674.
77. Loftus E, Elbishari Z, Fitzsimons P, Howard C, Sheikh Y, Lynch B, et al. Expanding the Phenotypic Spectrum of Methylenetetrahydrofolate Reductase (MTHFR) Deficiency in Childhood: A Case Series. Clin Case Rep. 2025;13(7):e9660. doi:10.1002/ccr3.9660. PMID: 40575625.
78. Metabolic Support UK [Internet]. metabolicsupport.org. Accessed 17 Oct 2026.