KMT5B-Related Neurodevelopmental Disorder

Autosomal dominant disorder caused by heterozygous loss of function of KMT5B (SUV420H1), the histone H4 lysine 20 methyltransferase; designated in OMIM as intellectual developmental disorder, autosomal dominant 51 (MRD51/IDD AD 51, OMIM 617788), with developmental delay or intellectual disability, macrocephaly, hypotonia, autism spectrum disorder and variable multisystem involvement.

Literature and regulatory information updated: 28 October 2026.

0. Reference Summary Table

1. Introduction

1.1. Historical overview

KMT5B-related disease was recognised through the genetics of autism and neurodevelopmental disorders (NDD) before it had a clinical name. The first de novo loss-of-function variant in the gene, then called SUV420H1, was described in 2012 by Iossifov and colleagues in a proband with autism spectrum disorder (ASD); in the same year, the Simons Simplex Collection (Sanders and colleagues) added deleterious missense variants, and in 2014 the Autism Sequencing Consortium (De Rubeis and colleagues) identified two further de novo loss-of-function variants and classified the gene as significant, with a false discovery rate of 0.01 or less (indirect attribution, according to SFARI Gene) [17]. The SFARI Gene database currently assigns KMT5B score 1 (high confidence), with 35 reports, 22 of them on autism, and 103 rare variants recorded; the last change of score dates from 1 April 2021 [17]. Simons Searchlight, a family registry programme, places the description of the first case in 2012 [10].

The first study with clinical follow-up was published by Stessman and colleagues in 2017. Targeted sequencing by single-molecule molecular inversion probes (smMIP) of 208 candidate genes in more than 11,730 patients with NDD and more than 2,867 controls, from 15 international cohorts, identified 91 risk genes, 38 of them new; KMT5B was among the genes with the greatest excess of de novo truncating (likely gene-disrupting, LGD) variants and missense variants of high predicted pathogenicity (corrected p below 1 × 10⁻⁶) [18]. Seven probands with variants in the gene underwent clinical assessment: global developmental delay or intellectual disability (ID) in 7/7, ASD in 5/6, language delay in 3/4, motor delay in 3/5, febrile convulsions in 3/5 and attention deficit in three [18]. The same work showed that pan-neuronal knockdown of the Drosophila orthologue (Hmt4-20) produces a specific deficit in habituation learning [18]. According to the OMIM synopsis, of these seven patients four were new and three (patients 5 to 7) came from the autism exome study by Iossifov and colleagues published in 2014, which included no clinical data; ages ranged from 5 to 19 years, ID was mild to moderate and there were variable non-specific features, such as facial dysmorphism, cryptorchidism, foot deformities, sleep difficulties, a tendency to tall stature, enlarged ventricles on magnetic resonance imaging in two patients and, in one patient, common variable immunodeficiency [8]. Of the seven variants, four were truncating and three missense with predicted deleterious effect; five had occurred de novo, one missense was maternally inherited with no clinical information on the mother and in one segregation could not be studied, and none was studied functionally [8, 9].

In 2018, Faundes and colleagues analysed histone lysine methylases and demethylases (KMT and KDM) across human developmental disorders, with data from 4,293 trios of the Deciphering Developmental Disorders study (DDD, United Kingdom and Ireland), from more than 200 children of the Canadian CAUSES project and from more than 41,800 individuals in the DECIPHER database [13]. KMT5B was firmly associated with a dominant developmental disorder due to haploinsufficiency, with two individuals carrying de novo truncating variants and seven deletions including the gene, and the authors described the picture as an “overgrowth syndrome with ID” [13]. The two carriers of point variants, aged 13 and 19 years, had global delay with slightly late walking, ID and poor language, one with atypical autism and the other with learning difficulties, ASD and hypotonia; head circumference lay between +2 and +4 standard deviations, with facial asymmetry, frontal bossing, epicanthic folds, posteriorly rotated ears, proptosis and microtia, and one had epileptic seizures that ceased [8]. Among the deletions, two de novo deletions, of 399 kb and 839 kb, occurred in patients with a similar picture [8, 13]. The same analysis showed that about three quarters of the published pathogenic variants in KMT and KDM genes are heterozygous truncating variants [13].

The following milestones were regulatory and phenotypic. The ClinGen (Clinical Genome Resource) syndromic disorders curation panel classified the gene–disease association as definitive on 8 April 2022, and dosage sensitivity curation assigned it sufficient evidence of haploinsufficiency on 13 June 2023 [4]. Eliyahu and colleagues reviewed 16 patients in 2022 and observed that macrocephaly and overgrowth were not constant [19]. In 2023, Sheppard and colleagues published the largest cohort to date, with 43 patients, which redefined the phenotype around macrocephaly, hypotonia and congenital heart disease and removed autism from the status of a defining feature [5]. Between 2023 and 2026 there were descriptions of periventricular nodular heterotopia [20], of neuromuscular involvement with denervation on biopsy [11], of cerebral metabolic profiles on positron emission tomography [7], of psychiatric manifestations in adults [21] and of parkinsonism [22].

1.2. Nomenclature and reclassifications

The phenotype has the OMIM designation “intellectual developmental disorder, autosomal dominant 51” (#617788), which replaced the former “mental retardation, autosomal dominant 51” (MRD51), and belongs to the phenotypic series of numbered autosomal dominant intellectual disabilities [2, 3, 8]. The phenotype entry was created in November 2017, from the patients of Stessman and colleagues, and completed in 2018 with those of Faundes and colleagues [8]. The change follows the general evolution of terminology: in the United States, the term “mental retardation” was abandoned by federal law in favour of “intellectual disability”; ICD-11 (International Classification of Diseases, 11th revision) adopted “disorders of intellectual development”; and DSM-5 (Diagnostic and Statistical Manual of Mental Disorders, 5th edition) defines ID as a neurodevelopmental disorder with intellectual and adaptive deficits in the conceptual, social and practical domains, with severity determined by adaptive functioning and not by intelligence quotient (IQ) [23]. Orphanet uses a descriptive name, “intellectual disability–hypotonia–facial dysmorphism–macrocephaly syndrome” (ORPHA:684226), and records as a synonym “neurodevelopmental disorder due to KMT5B haploinsufficiency” [1]. The most recent articles use “KMT5B-related syndrome” or “KMT5B-related neurodevelopmental disorder” [7, 11].

The gene had the symbol SUV420H1 (suppressor of variegation 4-20 homolog 1, homologue of the Drosophila gene Su(var)4-20), replaced by KMT5B in the systematic nomenclature of lysine methyltransferases; OMIM also records the names “lysine-specific methyltransferase 5B” and CGI85 [4, 5, 9]. The paralogue KMT5C corresponds to the former SUV420H2 and is located at 19q13.42 [24]. One of the case reports refers to the protein encoded by KMT5B as “SUV420H2”; the correct designation is SUV420H1 [4, 20].

1.3. Definition and scope

KMT5B-related neurodevelopmental disorder is defined by the presence of a heterozygous pathogenic or likely pathogenic variant in KMT5B, or of a deletion including the gene, in an individual with developmental delay, ID, learning difficulties or ASD, with or without macrocephaly, hypotonia, facial dysmorphism and congenital anomalies [3, 5]. There are no consensus clinical criteria, and the diagnosis is molecular [3]. The entity is single: it does not belong to a numbered syndromic class with subtypes and has no described allelic disorders [3]. It lies, however, within two groupings with diagnostic implications: the disorders of chromatin regulators and, within these, of histone lysine methylases [13], and the syndromes of macrocephaly or overgrowth with ID caused by epigenetic regulators, such as Sotos, Weaver, Tatton-Brown-Rahman and Malan syndromes [25, 26].

2. Molecular Genetics and Pathophysiology

2.1. The gene and the protein

KMT5B is located at 11q13.2 (GRCh38 chr11:68,154,863–68,213,648; GRCh37 chr11:67,922,330–67,981,115) and has 13 exons [4, 5]. The MANE Select transcript (Matched Annotation from NCBI and EMBL-EBI) is NM_017635.5 (ENST00000304363.9), which encodes isoform 1, NP_060105.3, of 885 amino acids [4, 6]. Older articles use NM_017635.4, whose numbering coincides with that of .5 for the variants checked [6, 13]. Two published sources contain inconsistencies in these basic data: a 2024 case report attributes 11 exons to the transcript, and a 2025 review states sometimes 885 and sometimes 865 amino acids; the NCBI protein record gives 885 amino acids [6, 7, 27]. Historical data explain part of these discrepancies: the gene was cloned in 2001 by Twells and colleagues, under the name CGI85, during sequencing of a 400 kb candidate region at 11q13 associated with type 1 diabetes (IDDM4), with a deduced protein of 384 amino acids and a structure of at least 10 exons over 24 kb, figures superseded by the current annotation; the mouse orthologue, cloned in 2004, encodes 876 amino acids with the N-terminal SET domain [9].

The protein has a catalytic core formed by an N-terminal domain (pre-SET, rich in helices), by the SET domain (residues 193–308 according to Prosite) and by a post-SET domain with a zinc-binding CxCxxC motif, followed by a variable regulatory C-terminal region [5, 7, 24]. The MobiDB database predicts five intrinsically disordered regions (residues 1–62, 363–444, 590–655, 716–740 and 816–850), mainly in the C-terminal half [7]. The cryo-electron microscopy structure of residues 1–390 bound to the nucleosome (PDB 8JHG) shows that the N-terminal region contacts the DNA, the histone H4 tail and the “acidic patch” of the H2A–H2B dimer; residues R220, S255, R257, K258, S283, G285, R286, Y349, R352 and R357 are critical for this interaction, and the last 91 residues are required for binding to GRIP1 [7]. There are two isoforms: SUV420H1_i1, with the complete C-terminus and a PxVxL motif that binds the HP1α protein and directs it to pericentromeric heterochromatin, and SUV420H1_i2, with a truncated C-terminus, without HP1α binding and with pan-cellular distribution [24].

2.2. Function: methylation of histone H4 lysine 20

The KMT5 family methylates histone H4 lysine 20 (H4K20) in steps: KMT5A (SETD8) generates the monomethylated form (H4K20me1), and KMT5B and KMT5C convert it into di- and trimethylated forms (H4K20me2 and H4K20me3); in vitro, KMT5B also methylates unmethylated H4K20 [4, 5]. Increasing H4K20 methylation is associated with chromatin compaction, transcriptional repression and the building of pericentromeric constitutive heterochromatin; H4K20me3 is a mark of epigenetic repression [3]. In mouse embryonic fibroblasts, the Suv420h1 and Suv420h2 enzymes localise to pericentromeric heterochromatin and are required for the accumulation of H4K20me3 there, with no effect on H3K9me3; their SET domains are active on nucleosomes, poorly active on recombinant histone octamers and inactive on peptides and free histones [9]. Targeting of the enzyme depends, in part, on interaction with the RB1 family (RB1, RBL1 and RBL2) [3]. In the mouse brain, Kmt5c is not expressed and its knockout has no apparent phenotype, so KMT5B is the main source of H4K20me2/3 in this organ [24].

H4K20me2 takes part in the repair of DNA double-strand breaks by non-homologous end joining (NHEJ, non-homologous end joining), by recruiting the 53BP1 protein (TP53BP1) to the sites of damage [3, 5]. KMT5B is also involved in myogenesis: it maintains the quiescence of skeletal muscle satellite cells, through facultative heterochromatin and repression of MyoD, and silences the myogenic inhibitor Eid3 through H4K20me3 [3, 24]. Brain expression of KMT5B mRNA in humans (BrainSpan data) peaks before 20 weeks post-conception and falls after birth to a stable level, which places the window of vulnerability in the prenatal period [5].

2.3. Disease mechanism: haploinsufficiency

The established mechanism is haploinsufficiency. ClinGen assigns the gene the maximum haploinsufficiency score (3) and a triplosensitivity score of 0 [4]. In gnomAD v4 (Genome Aggregation Database), the gene is highly intolerant to loss of function: probability of loss-of-function intolerance (pLI) 1.00, LOEUF (upper bound of the confidence interval of the observed/expected ratio of loss-of-function variants) 0.20, with 10 loss-of-function variants observed against 83.2 expected; missense variation is partially constrained (observed/expected ratio 0.71; Z = 4.10) [28]. Sheppard and colleagues cited a missense Z of 2.79, calculated in an earlier version of gnomAD [5]. The DECIPHER predicted haploinsufficiency percentage is 20.8% [4].

Several lines of evidence support loss of function as the common pathway. Truncating variants predominate; the 399 kb and 839 kb deletions including the gene produce the same picture [13, 19]; there is no systematic phenotypic difference between truncating and missense variants [7]; and the mouse with loss of one allele reproduces aspects of the human phenotype [5, 29]. For missense variants, Eliyahu and colleagues left open, in 2022, the possibility of gain of function or a dominant-negative effect [19]. Subsequent studies point to loss of function in these cases too: the missense variants p.(Gly94Ser), in the N-terminal domain, and p.(Ala293Pro), in the SET domain, reduce protein stability and expression and, unlike the normal protein, do not correct the increase in p53, DDIT4 and γH2AX caused by knockdown of the gene in HEK293T cells [30]; and normal human KMT5B, but not the pathogenic de novo missense variants, corrects the reduced dendritic complexity of mouse cortical neurons with knockdown of the gene [31]. Three-dimensional modelling of 14 missense variants from the Sheppard cohort predicted that p.(Cys200Arg), p.(Ile271Met), p.(His273Arg) and p.(Glu302Lys) interfere with binding of the cofactor S-adenosylmethionine (SAM), with 37, 64, 161 and 3 interatomic contacts affected, respectively, and that p.(His273Arg) may also disturb zinc binding; another seven surface-exposed variants, with a small destabilising effect (|ΔΔG| < 2) but at conserved residues, are likely to alter protein interactions [5]. In the Eliyahu family, the variant p.(Asn278Ile) lies next to cysteine 275, which binds zinc, and to phenylalanine 281, whose F281A mutation abolishes catalytic activity, and loses three of the five hydrogen bonds of the normal residue [19].

In the nonsense variant p.(Ser832Ter), located in the last exon, the mutant allele remains detectable in RNA (about 36–37% of reads), as expected from the absence of nonsense-mediated mRNA decay (NMD), and total KMT5B expression in blood is reduced [7]. The same authors raise, without demonstration, the hypothesis of an antagonistic effect of the short truncated product of the variant p.(Arg187Ter) [7].

2.4. Pathophysiology: cellular data and animal models

The downstream mechanisms of KMT5B loss have been studied in patient cells and in animal models; none has yet been validated as a therapeutic target. Fibroblasts from two patients, with p.(His273Arg) and p.(Phe154SerfsTer4), grow more slowly than those of two controls (P < 0.05 at 48 hours and P < 0.005 at 72 hours), without increased cell death [5]. RNA sequencing (RNA-seq) of patient lymphoblasts and of embryonic brains of heterozygous mice (embryonic day 14.5) shows alteration of pathways of nervous system development, axon guidance, TGF-β (transforming growth factor beta) signalling and extracellular matrix remodelling (collagens Col1a1 and Col1a2, Hapln1, Lum), and a four- to five-fold increase in the choroid plexus marker Ttr, with no change in circulating thyroxine; the authors propose, without demonstration, that KMT5B regulates cerebrospinal fluid dynamics [5]. In the blood of two patients with nonsense variants, RNA-seq showed 157 shared overexpressed genes (activation of kinases and transcription factors, TGF-β, nervous system and striated muscle cell development) and 230 shared underexpressed genes (antiviral response and regulation of nuclear division), with overexpression of DDIT4 (REDD1) in both; one of the published expression change values (log2FC 180.89) is implausible and should be read with caution [7].

The DNA damage response pathway links these data. Knockdown of Kmt5b in the mouse prefrontal cortex reduces H4K20me2, compromises 53BP1-mediated repair, activates p53 and Ddit4, decreases AMPA and NMDA receptor subunits (GluR2, NR2A, NR2B) and PSD-95, and produces isolated social deficits with dysfunction of glutamatergic transmission [24, 32]. Politano and colleagues propose the sequence loss of H4K20me2/3 → reduced recruitment of 53BP1 → impaired NHEJ → activation of p53 and DDIT4, with suppression of mTOR → synaptic dysfunction [7]. In utero knockdown reduces proliferation of cortical progenitors, accelerates neuronal migration and disorganises cortical lamination [24, 31]. In human cortical organoids with more than 745,000 cells, haploinsufficiency of KMT5B, ARID1B and CHD8 converges on an alteration in the development of GABAergic interneurons and deep-layer projection neurons [24].

In the mouse, the homozygous knockout is lethal in the embryonic period or produces animals that are smaller at birth (P < 0.001) without a proportional reduction of the brain, which constitutes relative macrocephaly; there is no increase in brain proliferation (Ki67, NeuN) and there are more pyknotic cells in the corpus callosum [5, 13, 27]. The heterozygote has persistently lower body weight and length, more marked in males, late eye opening, weaker reflexes, alterations in anxiety, fear and extinction, repetitive grooming in females and vertebral anomalies [13, 29]. In muscle, the heterozygous mouse has lower slow-muscle mass before puberty and, in the young adult, lower strength, lower muscle weight, smaller myofibres and fragmentation of the neuromuscular junction, which provides an experimental basis for hypotonia and for the human neuromuscular phenotype [5, 33]. Deletion of Kmt5b in mouse cardiopulmonary progenitors causes a phenotype resembling pulmonary hypertension with right ventricular dilatation, through oxidative stress [24]. In Xenopus, knockdown of KMT5B, but not of KMT5C, reduces transcription of ciliary genes and the cilia of epidermal multiciliated cells, an effect corrected by catalytically active PHF8 demethylase and independent of changes in chromatin accessibility [34]. In zebrafish, the orthologue has 52% identity with the human protein and its expression becomes progressively restricted to the central nervous system during embryonic development [5].

2.5. Variant spectrum and nomenclature

In the 43-patient cohort, all variants were heterozygous: 9 frameshift, 11 nonsense, 3 splicing, 2 partial gene deletions, 1 single amino acid deletion and 16 missense, most of the latter in the SET domain, at conserved residues [5]. In the review of 64 patients, the distribution was 24 missense (38%), 19 frameshift (29.5%), 12 nonsense (19%), 4 splicing (6%), 3 in-frame deletions (4.5%) and 2 partial gene deletions (3%) [7]. A Chinese family has a rare mutational mechanism, a 314 base pair insertion in exon 5, composed of a 4 bp duplication and 310 bp of an AluYa8 element, described by the authors as c.438_439ins[TCTT;KT192064.1:1_310], p.(Asp147SerfsTer33) [27]. By 2024 about 170 variants in the gene had been recorded across HGMD, ClinVar and the literature [27].

In ClinVar, the search of 28 October 2026 returned 317 records for the gene: 148 of uncertain significance (VUS, about 47%), 59 likely benign, 5 benign, 2 benign/likely benign, 47 likely pathogenic, 33 pathogenic, 6 pathogenic/likely pathogenic, 8 with conflicting classifications and 9 unclassified; among the pathogenic and likely pathogenic, frameshift (31) and nonsense (16) variants predominate, followed by missense (14) and splicing (7) variants [35]. These counts reflect the volume of laboratory submissions and do not constitute clinical frequency data [3].

Table 2.1 brings together the variants cited in this article, normalised in VariantValidator against the MANE transcript. The table of one of the founding articles of the missense spectrum contains an error: Eliyahu and colleagues describe in the text the variant of proband B as c.833A>T p.(Asn278Ile), but the table presents it as c.764A>T p.(Asn255Ile); VariantValidator rejects this second form, because the reference nucleotide at position c.764 is a cytosine and not an adenine, and confirms c.833A>T p.(Asn278Ile) as the valid description [6, 19].

Table 2.1. KMT5B variants cited, normalised (NM_017635.5; NP_060105.3; GRCh38)

The variants p.(Cys200Arg), p.(His273Arg), p.(Asp222Tyr), p.(Phe154SerfsTer4), p.(Tyr758CysfsTer5) and p.(Ala74Thr) from the Sheppard cohort have no ClinVar record found by text search and are kept with the authors' protein notation [5, 35]. The 2024 report of the Alu insertion has other inaccuracies (the abstract writes “ins[ASD; …]” and the text describes residue Asp147 as asparagine), which do not alter the classification of the variant as likely pathogenic [27].

2.6. Founder variants and presence in the population

No founder variants or recurrent variants associated with a population have been described [5, 7]. Three variants from the Sheppard cohort are in gnomAD: p.(Tyr758CysfsTer5) and p.(Ala74Thr), once each in the subset of participants with neuropsychiatric disease (which the authors interpret as possible missed diagnoses), and p.(Pro444Arg), twice [5]. In a disorder of variable expressivity, the presence of a variant in gnomAD does not exclude its pathogenicity [5].

2.7. Genotype–phenotype correlations, penetrance and expressivity

Predicted loss-of-function and missense variants produce overlapping pictures, with some differences in frequency between the two groups of the 43-patient cohort for which no statistical test was published (Table 3.1): hypotonia, joint hyperlaxity and congenital heart disease were more frequent in the loss-of-function group, and epileptic seizures in the missense group [5]. The 2022 review of 16 patients and the 2025 review of 64 patients also found no consistent correlation with variant type [7, 19]. There is no established predictor of severity [3].

Sex is the only modifier with a statistical association: in the Sheppard cohort, macrocephaly occurred in 81% of males and 44% of females (P = 0.0357, Fisher's exact test), and tall stature in 32% of males and in a single female (P = 0.0571, not significant) [5]. Published cases show an excess of male patients (35 males and 23 females in the 2025 review; 25 males and 17 females in the 2023 cohort), without statistical significance; the authors of the review raise the hypothesis of higher penetrance in males [5, 7]. The heterozygous mouse also shows sex-dependent effects [29].

Intrafamilial expressivity is variable. In at least five families, the transmitting parent has a milder picture than the child: a mother with mild ID and hand stereotypies whose son has severe ID without speech [20]; a father with ID and obsessive-compulsive disorder recognised only in adulthood [21]; a father with delayed walking and speech in childhood, apparently attenuated in adulthood [27]; and two parent–child pairs in the Sheppard cohort [5]. Penetrance has not been quantified; detection of the episignature in healthy carriers, noted by the laboratory that markets it, suggests that it may be incomplete [14].

2.8. Allelic disorders, dual diagnoses and modifiers

No allelic disorders have been described for KMT5B [3]. Reanalysis of the exomes of 4,351 German patients with NDD identified one patient with a dual molecular diagnosis, TMEM237 and KMT5B, a reminder that identification of a variant in KMT5B does not exclude another concomitant genetic cause [36]. In the Croatian report, the authors do not exclude a paternal polygenic contribution, since the father had schizophrenia [20]. KMT5B interacts with the FRG1 protein, implicated in facioscapulohumeral muscular dystrophy, but this link has no described clinical expression in patients with KMT5B variants [24].

3. Clinical Spectrum and Manifestations by System

3.1. Data sources and their limits

Clinical knowledge rests on three partially overlapping series and on case reports. The international cohort of Sheppard and colleagues, published in 2023, brought together 43 patients recruited through a clinical exome laboratory in the United States (GeneDx), the GeneMatcher platform and contacts between centres, with participation of centres in North America, Europe, Singapore and Australia; 26 patients had predicted loss-of-function variants (pLOF) and 17 had missense variants or the deletion of one amino acid [5]. The review by Eliyahu and colleagues (2022) brought together 16 published patients, three of them Israeli and new [19]. The review by Politano and colleagues brought together 64 patients with clinical characterisation published up to 9 March 2025, with 23 females, 35 males and 6 of unstated sex, and ages between 5.5 weeks and 50 years (median close to 11 years) [7]. In all of them, denominators vary by feature, because not all patients were assessed for all features; a percentage refers to the patients with information and not to the whole cohort. The series are mostly paediatric, cross-sectional and non-systematic, and do not describe untreated natural history: patients received early intervention, rehabilitation therapies and symptomatic treatments whose effects were not analysed [5, 7]. Frequencies published before 2023 come mainly from autism cohorts, which inflated the proportion of ASD [5].

Table 3.1. Phenotype in the 43-patient cohort (Sheppard 2023), overall and by variant type

Source: [5]. Subgroup sums are calculations from the published denominators.

Table 3.2. Literature review of 64 patients (Politano 2025)

Source: [7]. The sum of some denominators in the original table does not equal 64; the percentages are those published.

3.2. Psychomotor development and cognition

Developmental delay or ID is present in practically all patients described [5, 7]. ID severity ranges from mild to profound: in the 2018 series, two patients with deletions had mild or mild-to-moderate ID, one with a frameshift had moderate ID and one with a nonsense variant had severe ID [13]; a 35-year-old adult and a 16-year-old adolescent have profound and severe ID, without speech [19, 20]. Motor delay is usual in the first years: head control at 5–5.5 months, rolling at 7–12 months, sitting at 8–9 months and walking between 18 months and 3 years in the cases described in detail [7, 11, 19, 20, 27]. Language is the most affected domain: first words between 26 months and 4 years in cases with acquisition, and absence of speech at 7 and at 16 years in others [19, 20, 27].

The spectrum includes intelligence within normal limits. Two 13-year-old Italian girls, with de novo nonsense variants, had full-scale IQs of 104 and 96, with developmental delay in the first years and later diagnoses of ASD, developmental coordination disorder (DCD), executive dysfunction, expressive language deficit and internalising problems above the 98th percentile on the CBCL (Child Behavior Checklist) [7]. The authors stress that these profiles, a minority in the literature, justify domain-by-domain neuropsychological assessment even when IQ is normal [7].

3.3. Behaviour and psychiatric manifestations

ASD is present in about two thirds of patients with information (62% in the 2023 cohort; 65% in the 2025 review), a lower proportion than in the autism series from which the gene emerged, in which almost all cases had ASD [5, 7]. The 2022 review found ASD or autistic traits in 7/16 patients (44%) [19]. Attention deficit was described in three of the first seven patients, and attention deficit hyperactivity disorder (ADHD) or ADHD traits in 11% of patients in the 2025 review [7, 18]. About a quarter of patients have other behavioural or psychiatric disorders, including self-injury and self-harm; two cases of psychosis (4%) and three of regression (5%) have been described [7]. In adulthood, obsessive-compulsive disorder was reported in a carrier father with ID recognised only in adulthood, and episodes of unmotivated laughter in a child [19, 21].

3.4. Growth: macrocephaly and stature

Macrocephaly is the growth parameter that best signals the disorder. It affects 63% (24/38) of patients in the 2023 cohort and 57% (30/53) of those in the 2025 review, with male predominance (81% of males versus 44% of females) [5, 7]. Tall stature is less frequent (21%, 9/43) and, when present, is associated with macrocephaly: in the 43-patient cohort, 8 had macrocephaly and tall stature, 22 macrocephaly without tall stature and only 1 tall stature without macrocephaly [5]. The authors conclude that the core feature is macrocephaly and not systemic overgrowth, with the limitation that they did not collect parental height [5]. This conclusion qualifies the initial description of an “overgrowth syndrome with ID”, based on few individuals, some with macrocephaly of +4.43 standard deviations (SD) and height of +2.91 SD and others with normal head circumference and height (Table 3.3) [13]. A 2026 review states that macrosomia affects 30 of 31 patients and attributes the figure to Faundes 2018; the primary data do not support it [5, 13, 24]. Three patients with missense variants (18%) had short stature without microcephaly [5], and other cases have normal head circumference and height [11].

Overgrowth may become more marked later. A 35-year-old Israeli man had accelerated growth from the age of 10 and reached a height of 205 cm, a weight of 150 kg and a head circumference of 62 cm, all above the 97th percentile, with morbid obesity; he is the first adult described with overgrowth above 6 SD [19]. In infancy, poor weight gain and weak suck were reported in the series before 2023 and in isolated cases [5, 11].

Table 3.3. Individuals with KMT5B in the Faundes 2018 series (NM_017635.4)

Source: [13].

3.5. Nervous system

Hypotonia is the main neurological finding (81% in the 2023 cohort; 58% in the 2025 review) and manifests from the neonatal period, sometimes with weak suck [5, 7, 11]. Next come motor incoordination and DCD, which account for two thirds of the movement disorders in the 2025 review [7]. In 2026, a report of parkinsonism in KMT5B-related disease was published in Spain, presented as an expansion of the spectrum; the full text and abstract were not accessible, so its clinical features are not described here [22].

Epileptic seizures occur in 23% (12/52) to 31% (9/29) of patients [5, 7]. In the first patients described, febrile convulsions predominated (3/5) [18]. There is no characteristic electroencephalographic pattern or seizure type [7]. In the cases described in detail, generalised seizures began at 8 and 10 years and were controlled with one antiepileptic drug, in one case carbamazepine [19, 20].

Neuroimaging is abnormal in about 37–40% of patients studied, with non-specific and sporadic findings: ventricular dilatation, white matter changes, unspecified leukodystrophy, Chiari malformation, syringomyelia, subdural hygroma, hydrocephalus, hypoplasia or dysgenesis of the corpus callosum, periventricular nodular heterotopia and perisylvian polymicrogyria [5, 7, 19]. Bilateral periventricular nodular heterotopia next to the caudate nucleus was first described in 2023 [20]. Fluorodeoxyglucose positron emission tomography (FDG-PET) in two adolescents with normal magnetic resonance imaging showed left medial temporal (uncus, fusiform or parahippocampal gyrus) and cerebellar hypometabolism and, in one of them, right parietal hypermetabolism; it is a research examination, with no established diagnostic role [7].

3.6. Neuromuscular involvement

A German case, published in December 2025, describes a girl with reduced fetal movements, hypotonia and weak suck at birth, walking at 2.5 years and, from childhood, episodes of fatigue, hypersomnia and transient stiffness of one lower limb with falls, followed by progressive weakness and fatigability [11]. At 14 years, walking was limited to 500 m, with exertional cramps, choking, exertional headache, waddling gait, mild scapular winging, distal hypotonia, postural tremor and distal hyperlaxity; she used a wheelchair for long distances and orthoses [11]. Post-exercise lactate rose to 57.2 mg/dL (10.4 mg/dL at rest), nerve conduction was normal and quadriceps biopsy at 16 years showed variation in fibre calibre, angular fibres, fibre-type grouping, predominance of fast fibres and lipid accumulation, suggestive of chronic denervation, with a normal mitochondrial respiratory chain; proteomics identified 77 dysregulated proteins (sarcomeric, mitochondrial and glycolytic, acute-phase, complement and extracellular matrix) [11]. The picture is consistent with the heterozygous mouse, which loses muscle strength and mass and has fragmentation of the neuromuscular junction [33]. Reduced muscle mass was also recorded in the Croatian adolescent [20].

3.7. Heart and vessels

Congenital heart disease affects 27% (8/30) of patients in the 2023 cohort, with atrial septal defect, ventricular septal defect and patent ductus arteriosus, more frequent in the loss-of-function group (33%) than in the missense group (10%) [5]. It was first described as frequent in that cohort [5]. In the 2025 review, non-cerebral congenital malformations (about 17–21%) are mainly cardiovascular and genitourinary [7]. One of the cases had surgical repair of an atrial septal defect diagnosed at 2 years [11]. The pulmonary hypertension of the mouse with deletion of the gene in cardiopulmonary progenitors has no described human counterpart [24].

3.8. Craniofacial

Facial dysmorphism is present in about four fifths of patients (79% in the 2023 cohort; 78% in the 2025 review), without a facial pattern recognisable at first sight [5, 7, 19]. The features most cited are a long face, arched or thick eyebrows, widely spaced eyes with upslanting palpebral fissures, prominent ears and mild prognathism [5]. The 2022 review found a broad forehead in 31% of patients and, recurrently, ptosis, epicanthus, high palate and long or tapering fingers [19]. Individual cases add hypertelorism, a broad nasal root, a bulbous nasal tip, diastema of the incisors, gingival hypoplasia or hypertrophy, low-set ears and occipital protrusion [7, 19, 27]. Three patients with metopic craniosynostosis have been described, and an exome analysis of 526 trios with syndromic craniosynostosis identified KMT5B among the 13 genes with a significant excess of damaging de novo variants [7, 37].

3.9. Musculoskeletal system

Joint hyperlaxity affects half of patients with loss of function and 17% of patients with missense variants [5]. Scoliosis has been described from lumbar forms of 7° to severe thoracic forms operated on in adolescence [11, 13, 20]. Other findings are pectus excavatum, flat feet, talipes equinovarus, genu valgum, large hands and feet with long digits, and joint and bone pain [7, 11, 13, 19, 27].

3.10. Eyes and hearing

Ocular anomalies affect 22% of patients in the 2025 review [7]. Strabismus is the most frequent (5/16, 31%, in the 2022 review) and may require surgery; ptosis, astigmatism and hypermetropia corrected with glasses from 18 months have also been described [19]. In the neuromuscular case, visual evoked potentials showed a prolonged P100 latency on one side [11]. One adolescent had central-type hearing loss [20].

3.11. Sleep, digestive, endocrine and other systems

Sleep disorders were reported in 12% of patients in the 2025 review, a figure the authors consider underestimated; they include sleep-onset insomnia, night terrors and sleepwalking, the latter in three of the six patients described in detail [7]. Gastrointestinal complaints include constipation, gastro-oesophageal reflux and lactose intolerance [7, 19]. In adolescence and adulthood, obesity with insulin resistance and low HDL cholesterol have been described, and, in one adult, low testosterone with an incidentally found 6 × 8 mm non-functioning pituitary lesion [7, 19]. Isolated findings, without established association, are common variable immunodeficiency in one patient (1/7 patients included in an immunology review of chromatinopathies), diabetes mellitus in one patient with an 839 kb deletion, multiple lentigines, neutropenia, lymphadenopathy and cryptorchidism [13, 19, 38]. One child had vitamin D deficiency [27].

3.12. Course by age

In the prenatal and neonatal period, the findings described are reduced fetal movements, intrauterine growth restriction in one case, variable birth weight (1,825 g to 4,500 g), hypotonia and weak suck [7, 11, 19]. In the first year, hypotonia, motor delay, lack of eye contact and increasing head circumference predominate [19]. At preschool and school age, language delay, ASD, DCD, sleep problems and, sometimes, seizures, which may begin between 8 and 10 years, become evident [7, 19, 20]. In adolescence, scoliosis, obesity, internalising problems and, in the neuromuscular case, progressive weakness are added [7, 11, 20]. Adulthood is poorly described: the oldest published patient was 50 years old, and the adults reported have overgrowth and obesity, obsessive-compulsive disorder, psychosis or parkinsonism [7, 19, 21, 22]. There are no data on mortality or life expectancy [3].

4. Clinical Recognition and Red Flags

4.1. When to suspect

There is no pathognomonic sign or recognisable facial gestalt, and the core phenotype (developmental delay, macrocephaly, hypotonia, ASD and dysmorphism) is shared by many disorders of chromatin modifiers; clinical suspicion serves to choose a broad genomic test and to interpret the result, not to request a test targeted at one gene [3, 19]. The contexts in which the sources recommend considering the disorder are: global developmental delay or ID with absolute or relative macrocephaly, with or without tall stature [5, 19]; ASD, including children with normal IQ but with DCD, executive dysfunction or expressive language deficit [7]; hypotonia with septal congenital heart disease or patent ductus arteriosus and non-specific dysmorphism [5]; syndromic metopic craniosynostosis [7, 37]; and combined neurodevelopmental and neuromuscular pictures, with distal hypotonia, weakness, fatigability and distal hyperlaxity, seen in neuromuscular disease clinics [11].

Table 4.1. Contexts of suspicion by age

4.2. Measurement and interpretation of head circumference

Macrocephaly thresholds vary between sources: the Oxford Desk Reference uses more than 3 SD above the mean in the macrocephaly chapter and more than 2 SD (approximately the 98th percentile) in the overgrowth chapter, and the GeneReviews chapters on Sotos and Tatton-Brown-Rahman syndromes use 2 SD or more [25, 39–41]. In the assessment, head circumference is compared with length or height, to distinguish absolute from relative macrocephaly; parental head circumference is measured and corrected for height; and spurious causes, such as hairstyle or craniosynostosis, are excluded [39]. In KMT5B, macrocephaly is often relative or isolated, without tall stature, and is less frequent in females, so its absence does not exclude the diagnosis [5, 19].

4.3. Red flags and emergency signs

In the child with an established diagnosis, the sources describe manifestations that require targeted assessment: epileptic seizures (present in about a quarter to a third of patients, with possible onset at school age) [5, 7]; developmental regression, described in 5% of patients and, in the neuromuscular case, in the form of episodes of fatigue, hypersomnia and transient stiffness [7, 11]; progressive weakness, falls, choking or exertional cramps, suggestive of a neuromuscular component [11]; signs compatible with Chiari malformation, syringomyelia or hydrocephalus, changes described in the cohort [5, 19]; progression of a scoliosis [20]; and new psychiatric changes, including psychosis, self-injury and obsessive-compulsive symptoms [7, 21]. In children with ASD and ID, macrocephaly is associated with a higher risk of seizures (odds ratio 7.7 in an Israeli community study of more than 4,000 children with neurodevelopmental disability) [39].

4.4. Newborn screening

There is no newborn screening for this disorder; the defect is epigenetic and produces no measurable metabolite [3].

5. Diagnosis: Criteria, Testing and Algorithms

5.1. Criteria

No clinical diagnostic criteria have been published [3]. The diagnosis is established by identification of a heterozygous pathogenic or likely pathogenic variant in KMT5B, or of a deletion including the gene, in an individual with a compatible phenotype; the methylation episignature suggests the diagnosis, but GeneReviews recommends confirming it by molecular testing targeted at the gene, which also enables family studies [2, 3]. There is no biochemical or metabolic biomarker [3].

5.2. Diagnostic algorithm

The current sequence in the child with developmental delay or ID follows the 2021 American College of Medical Genetics and Genomics (ACMG) guideline, which strongly recommends, by the GRADE method (Grading of Recommendations Assessment, Development and Evaluation), exome or genome sequencing as a first- or second-line test in patients with congenital anomalies with onset before 1 year or with developmental delay or ID with onset before 18 years, because of the higher diagnostic yield and possible cost-effectiveness when requested early [12]. The 2014 American Academy of Pediatrics (AAP) guideline placed chromosomal microarray and fragile X testing in the first line and the exome as an emerging test; it remains a reference for general clinical assessment (inborn errors of metabolism, magnetic resonance imaging in selected patients), but has been superseded by the ACMG regarding the place of the exome [12, 42]. In the child with ASD, the AAP recommends genetic aetiological assessment as part of the initial evaluation [43].

The pathway of the published patients illustrates the earlier sequence: in the three Israeli probands and in the Croatian adolescent, microarray and fragile X testing were normal before the exome; the Croatian adolescent also had a karyotype and subtelomeric and microdeletion MLPA (multiplex ligation-dependent probe amplification), all negative, and was diagnosed by clinical exome (TruSight One panel) [19, 20]. In the 43-patient cohort, the diagnosis was made by clinical exome or by an autism and ID panel [5].

Table 5.1. Practical algorithm when KMT5B-related disease is suspected

5.3. Molecular tests, sensitivity and limits

The clinical sensitivity of the exome or genome for this disorder has not been published; the fraction of patients with a compatible phenotype who have a KMT5B variant is also unknown [3]. General yields serve as context: in cases of non-specific ID, high-resolution microarray has a mean yield of about 18% as first line, and the exome reaches 50% in sporadic severe ID; in the DDD study, 94 genes showed enrichment of damaging de novo variants [23]. Trio analysis facilitates interpretation of VUS; female sex, paternal age and absence of consanguinity increase the probability of a de novo dominant variant [23].

The limits relevant to KMT5B are as follows. Exome sequencing may fail to detect deletions of one or a few exons, deep intronic variants and mobile element insertions: the 314 bp Alu insertion in exon 5 of the Chinese family escapes conventional analysis pipelines that do not call mobile insertions [27]. Large deletions including the gene are detected by microarray [13]. The exome does not detect repeat expansions, so fragile X is tested separately [23]. Low-level mosaicism in blood may go unnoticed, both by routine sequencing and by the episignature [2]. A negative result does not exclude the diagnosis: periodic exome reanalysis increased the yield in a series of 4,351 German patients and revealed, among others, KMT5B diagnoses [36].

5.4. Variant interpretation and report content

Classification follows the ACMG/AMP criteria (American College of Medical Genetics and Genomics / Association for Molecular Pathology). Truncating variants before the last exon receive the PVS1 criterion for loss of function in a gene with established haploinsufficiency [4, 27]. In the variant c.2495_2497del p.(Ser832Ter), located in the last exon, the allele escapes NMD; the authors applied PVS1 and classified it as likely pathogenic; applying PVS1 at full strength to a variant that does not trigger NMD is debatable [7]. Missense variants require caution: about 47% of ClinVar records for the gene are VUS, mainly missense [35]. Missense variants in the SET domain (residues 193–308), a region in which UniProt records 18 missense or in-frame variants (8 pathogenic, 10 uncertain and none benign), justify the PM1 criterion, and should not be presumed benign because they are not truncating [3, 7, 20]. The presence of a variant once or twice in gnomAD does not exclude it as the cause, as shown by three variants from the 43-patient cohort [5].

For VUS, family segregation is of limited value, because carrier parents may have mild pictures [20, 21]; the episignature test and clinical assessment of the parent help the decision [14]. The report should state the gene, the transcript with version (NM_017635.5), the HGVS description at DNA and protein level, zygosity, origin (de novo, inherited or undetermined), the ACMG/AMP class with the criteria applied and, for deletions, the coordinates and genes included [6, 13]. ClinGen and GenCC record two “definitive” and three “strong” classifications for the gene–disease association; KMT5B is not on the ACMG secondary findings list (version 3.2) [4].

5.5. DNA methylation episignature

KMT5B is included in Table 3 of the GeneReviews chapter on episignatures, with reference to the signatures derived by Aref-Eshghi (2020) and by Levy (2021–2022) on 450,000 and EPIC methylation arrays [2, 44]. The assay compares the methylation pattern at about 1,000 loci across the genome, in leucocytes, with controls and with the signatures of each gene [2]. The clinical EpiSign v5 panel (Greenwood Genetic Center), with more than 100 conditions in 2025, includes the disorder as a moderate-strength episignature; the test uses 3–5 mL of blood in EDTA, has a turnaround of about 4 weeks and the laboratory notes that healthy carriers and individuals with incomplete penetrance also show the signature [14, 44].

GeneReviews considers the episignature a second-line test: useful when first-line molecular testing has not established the diagnosis, when there is a VUS, when the method used does not detect large deletions or duplications, non-coding or mosaic variants, or when the phenotype suggests more than one disorder with a known signature; it is not suitable as a first-line test because it covers a limited number of disorders [2]. The limits are tissue type (a signature characterised in blood may not apply to DNA from amniocytes, chorionic villi or fibroblasts), low-level mosaicism in leucocytes and the need to confirm a positive signature molecularly [2]. The sources consulted do not describe the sensitivity and specificity of the KMT5B signature; the articles that derived it present these data in supplementary tables [44].

5.6. Other examinations

Blood RNA-seq has been used in research to confirm reduced expression of the gene and NMD escape of a last-exon variant [7]. FDG-PET and muscle biopsy with proteomics are research examinations or examinations for individual indication, with no routine diagnostic role [7, 11]. Non-specific complementary assessment includes developmental assessment, serial head circumference, echocardiogram (heart disease affects 27% of patients), EEG if seizures are suspected and brain magnetic resonance imaging in the presence of micro- or macrocephaly and neurological signs [3, 5, 23]. In macrocephaly with delay without a diagnosis, the Oxford Desk Reference also includes metabolic screening (urinary organic acids, mucopolysaccharidoses) and, with strong suspicion of PI3K/AKT pathway mosaicism, high-depth sequencing (above 1,000×) [39].

5.7. Prenatal and preimplantation diagnosis

Prenatal diagnosis and preimplantation genetic testing are only possible when the proband's variant is known and classified as pathogenic or likely pathogenic [23]. Molecular methods targeted at the variant are used; the episignature is not validated in amniocytes or chorionic villi [2]. There is no characteristic prenatal ultrasound sign; the prenatal findings described (reduced fetal movements, intrauterine growth restriction) are isolated and non-specific [7, 11].

6. Differential Diagnosis

6.1. Principles

The differential diagnosis of KMT5B-related disease is that of the child with developmental delay or ID and macrocephaly, with or without overgrowth. Most known overgrowth syndromes belong to two molecular families, the epigenetic regulators and the PI3K/mTOR pathway [25]. In practice, clinical distinction between these entities is limited and the discriminating test is broad genomic analysis (exome or genome), complemented by tests that the exome does not cover (FMR1 expansions, 11p15 methylation) and, in the presence of VUS, by the episignature, available for several of the entities in the table (NSD1, DNMT3A, EZH2, EED, CHD8, SETD2, SETD5, KAT6A, KMT2D/KDM6A, FMR1) [2, 26, 39]. GeneReviews recommends also considering the OMIM phenotypic series of autosomal dominant, autosomal recessive and X-linked ID [26].

Table 6.1. Differential diagnosis

6.2. Overgrowth syndromes due to epigenetic regulators

Sotos syndrome is the classic differential. It shares ID, ASD, macrocephaly, seizures, heart disease and hyperlaxity, but has a characteristic face, advanced bone age and global overgrowth as part of the cardinal triad [40]. In KMT5B-related disease, the face is non-specific and tall stature affects only one fifth of patients [5]. Weaver and Tatton-Brown-Rahman syndromes are distinguished by their own facial features, advanced bone age, camptodactyly and doughy skin (Weaver) or by obesity and a round face (Tatton-Brown-Rahman), in addition to specific tumour risks (neuroblastoma and acute myeloid leukaemia, respectively) [41, 45]. SUZ12 is, with EZH2 and EED, one of the core subunits of Polycomb repressive complex 2, and its heterozygous variants cause Imagawa-Matsumoto syndrome, identified since 2017 in patients previously classified as Weaver or Weaver-like [46]. In the 13 patients gathered by 2019, excess height became more marked with age (mean of +1.99 SD for birth length, +2.53 SD in childhood and +3.75 SD for final height), head circumference at birth was almost normal (mean of +0.85 SD) and two patients never exceeded 2 SD; in the 2019 series of 10 patients, 8 had developmental delay and 6 ID, and 6 of the 8 boys had genitourinary anomalies [46]. Missense variants appeared to be associated with greater overgrowth than truncating variants, in a sample the authors considered small for correlations [46]. Advanced bone age, large hands and feet and the progression of height set it apart from the KMT5B picture, in which tall stature is a minority feature [5]. Malan syndrome, whose main differential is Sotos syndrome, has a slender habitus, ocular changes and an anxious profile [47].

CHD8-related disorder is the closest to the KMT5B picture: macrocephaly in the first year, late tall stature, ASD and mild to moderate ID; sleep disturbance and constipation are prominent and hypotonia is less frequent (about 30%) [26]. The convergence of the two entities has an experimental correlate, because haploinsufficiency of KMT5B and of CHD8 alters the same neuronal populations in cortical organoids [24]. SETD2-related disorders combine macrocephaly with ventriculomegaly, Chiari malformation and obesity with advanced bone age [48].

6.3. Other causes of macrocephaly with delay

PTEN hamartoma tumour syndrome should be considered in any child with macrocephaly and ASD; macrocephaly is usually more marked (often above 4.5 SD), and the diagnosis has oncological consequences for the patient and the family [39, 49]. In one of the Israeli probands with KMT5B, PTEN sequencing had been normal before the exome [19]. Fragile X should be tested separately, by repeat analysis, because the exome does not detect it [23]. Benign familial macrocephaly requires measurement of parental head circumference, bearing in mind that PTEN and PTCH1 may give pictures with few signs [39]. Leukodystrophies with macrocephaly (Canavan, Alexander), the mucopolysaccharidoses, especially type III, and glutaric aciduria type 1 are distinguished by regression, imaging and biochemical tests (urinary organic acids) [39]. Neurofibromatosis type 1, the RASopathies (cardiofaciocutaneous, Costello), skeletal dysplasias and Proteus syndrome complete the differential of macrocephaly [39].

6.4. Differential of the neuromuscular phenotype

In the presence of distal hypotonia, progressive weakness, fatigability and raised exertional lactate in a patient with ID, the differential includes mitochondrial myopathies and neuromuscular junction disorders; in the published case, the respiratory chain was normal and the biopsy showed chronic denervation [11]. The authors recommend including KMT5B in the panels and exome analysis of patients in neuromuscular clinics with a combined neurodevelopmental and neuromuscular phenotype [11].

7. Treatment and Risk Management Strategies

7.1. General principles

There is no specific or disease-modifying treatment, nor a dedicated management guideline; GeneReviews has no dedicated entry and, therefore, there is no gene-specific management section, surveillance table or list of agents to avoid [3]. No drug has been studied in patients with KMT5B variants, and there is no randomised trial, interventional study or guideline that names the gene [3, 15]. Management is symptomatic, organised by problem and based on the general guidelines for ID, ASD, insomnia in ASD, ADHD and epilepsy. The doses presented come from the official labels (FDA, EMA and INFARMED) for analogous indications; their use in this disorder is an extrapolation and, when the label indication does not match the patient's problem, the use is off-label. For the child with ASD, the AAP recommends management of comorbidities (sleep, feeding, gastrointestinal symptoms, obesity, convulsions, ADHD, anxiety and elopement), evidence-based behavioural interventions and decisions shared with the family [43].

7.2. Neurodevelopment, behaviour and rehabilitation

Early intervention, speech therapy, occupational therapy and physiotherapy were the interventions used in the cases described in detail [20, 27]. Politano and colleagues advocate a comprehensive neuropsychological assessment, of global cognition and by domain (attention, executive functions, memory, social perception, sensorimotor functions, language and behaviour), to guide individualised rehabilitation, even in patients with normal IQ [7]. Evidence-based behavioural interventions are the basis of ASD treatment [43]. DCD and motor incoordination, present in two thirds of patients with movement disorders, justify targeted physiotherapy and occupational therapy [7].

7.3. Sleep disorders

The 2020 American Academy of Neurology (AAN) guideline on insomnia in children and adolescents with ASD recommends, at level B, assessing drugs and comorbidities that contribute to insomnia, using behavioural sleep hygiene strategies as first line and offering melatonin if these are insufficient, starting with a low dose, preferring pharmaceutical-grade melatonin and informing about adverse effects and the lack of long-term safety data; there is no evidence for weighted blankets or special mattresses [50]. The melatonin doses in the guideline were not extracted, because only the abstract was accessible; therefore, the reference dose is that of the European label.

Prolonged-release melatonin (Slenyto®, 1 mg and 5 mg tablets) was authorised by the European Commission on 20 September 2018 for insomnia in children and adolescents aged 2 to 18 years with ASD and/or Smith-Magenis syndrome, with renewal on 5 June 2023. The Commission decision of 26 August 2024 extended the indication to neurogenetic disorders with aberrant daytime melatonin secretion and/or nocturnal awakenings, and that of 17 March 2025 added insomnia from 6 to 17 years with ADHD, whenever sleep hygiene measures are insufficient [51]. In ASD and neurogenetic disorders, the recommended starting dose is 2 mg, increased to 5 mg if the response is inadequate, up to a maximum of 10 mg; in ADHD, the starting dose is 1–2 mg, adjustable to 5 mg and, if clinically necessary, to 10 mg [51]. It is taken once daily, 0.5 to 1 hour before bedtime, with or after food, and the tablet is swallowed whole, without breaking, crushing or chewing [51]. After about 3 months the effect is assessed and it is stopped if there is no clinically relevant benefit; thereafter, the patient is reassessed at least every 6 months, with withdrawal attempts at least yearly when the effect is uncertain; safety data are available for up to 2 years [51]. It is not recommended in hepatic impairment and requires caution in renal impairment; the most frequent adverse effects (1–10%) are somnolence, fatigue, mood changes, headache, irritability, aggression and a hangover feeling [51]. Use in KMT5B-related disease falls within the European indication when ASD coexists, or within the category of neurogenetic disorders with nocturnal awakenings; sleep disorders affect 12% of published patients, with insomnia, night terrors and sleepwalking [7].

7.4. Irritability, aggression and self-injury

Risperidone has different indications in the United States and in the European Union. The FDA (Food and Drug Administration) label approves it for irritability associated with autistic disorder from 5 to 17 years, including aggression, deliberate self-injury, temper tantrums and quickly changing moods [52]. Below 20 kg, it is started at 0.25 mg/day; from 20 kg, at 0.5 mg/day; after at least 4 days, it is increased to 0.5 mg/day (below 20 kg) or 1 mg/day (from 20 kg), maintained for at least 14 days; subsequent increases, of 0.25 mg/day or 0.5 mg/day, are made at intervals of at least 2 weeks; the effective range is 0.5 to 3 mg/day, in one or two doses, with no maximum dose defined in the source beyond the upper limit of that range, and there are no dose data below 15 kg [52]. In Portugal, the SmPC (summary of product characteristics) of Risperdal® oral solution 1 mg/mL limits the paediatric indication to short-term (up to 6 weeks) symptomatic treatment of persistent aggression in conduct disorder, in children over 5 years and adolescents with below-average intellectual functioning or mental retardation, as part of a programme with psychosocial and educational interventions and prescribed by a specialist [53]. From 50 kg, it is started at 0.5 mg/day, with increases of 0.5 mg no more often than every other day, and the usual optimal dose is 1 mg/day (0.5–1.5 mg); below 50 kg, it is started at 0.25 mg once daily, with increases of 0.25 mg, and the usual optimal dose is 0.5 mg/day (0.25–0.75 mg); the SmPC does not define a maximum dose beyond these ranges, and risperidone is not recommended below 5 years [53]. The two posologies diverge in indication, weight threshold (20 kg versus 50 kg) and duration (no limit versus up to 6 weeks); for irritability associated with ASD, use in Portugal is off-label.

Aripiprazole (Abilify®) is FDA-approved for irritability associated with autistic disorder from 6 to 17 years: it is started at 2 mg/day, increased to 5 mg/day and, if necessary, to 10 or 15 mg/day, with adjustments of up to 5 mg/day at intervals of no less than 1 week; the recommended range is 5 to 15 mg/day, in a single dose, and half the dose is used in CYP2D6 poor metabolisers [54]. The sources consulted did not include the European indication of aripiprazole in this situation. Atypical antipsychotics require monitoring of weight, glycaemia, lipids, prolactin, somnolence and extrapyramidal symptoms [52, 54]. This point carries particular weight in patients with obesity or signs of metabolic syndrome, described in adolescence and adulthood [7, 19].

7.5. Attention deficit and hyperactivity

Methylphenidate (Rubifen®, 5, 10 and 20 mg tablets) is indicated in Portugal for ADHD from 6 years, as part of a comprehensive treatment programme and under specialist supervision [55]. Before starting, blood pressure, heart rate and history of sudden death are assessed, and height and weight are recorded; it is started at 5 mg once or twice daily (for example, at breakfast and lunch), with weekly increments of 5 to 10 mg, up to a maximum of 60 mg/day; it is stopped if there is no improvement after 1 month of adjustment, and a break is made at least yearly for reassessment [55]. Blood pressure and pulse are checked at each adjustment and at least every 6 months, and height, weight and appetite every 6 months [55]. Methylphenidate may lower the seizure threshold in patients with a history of seizures and should be stopped if seizure frequency increases or new seizures appear, a warning relevant in a disorder with seizures in 23–31% of patients [5, 7, 55]. ADHD or ADHD traits affect 11% of published patients [7].

7.6. Epilepsy

There are no data on the response to specific antiepileptic drugs in this disorder, nor a characteristic electroencephalographic pattern or seizure type to guide the choice [7]. The choice follows the seizure type and general epilepsy standards. In the published cases, the generalised tonic-clonic seizures of an adolescent were controlled with carbamazepine, and those of an adult with an unspecified antiepileptic drug [19, 20]. As an example of a label dose, for the drug used in the Croatian case, the US label of Tegretol® (carbamazepine) indicates it for complex partial seizures, generalised tonic-clonic seizures and mixed seizures, not absences; below 6 years it is started at 10–20 mg/kg/day, in two or three doses (tablets) or four (suspension), with weekly increases, and response is generally obtained with less than 35 mg/kg/day; from 6 to 12 years it is started at 200 mg/day (100 mg twice daily), with weekly increases of up to 100 mg/day, maintenance of 400 to 800 mg/day and a usual maximum of 1,000 mg/day; from 12 years it is started at 400 mg/day, with weekly increases of up to 200 mg/day, maintenance of 800 to 1,200 mg/day and a maximum of 1,000 mg/day from 12 to 15 years and 1,200 mg/day above 15 years [56]. The label has a boxed warning for serious skin reactions (toxic epidermal necrolysis and Stevens-Johnson syndrome), associated with the HLA-B*15:02 allele, which is tested for before starting treatment in patients with ancestry from at-risk areas of Asia, and for aplastic anaemia and agranulocytosis, with a mandatory baseline blood count [56]. This warning is relevant for the Chinese patients described [27, 30].

7.7. Orthopaedic, cardiac and ophthalmological problems and surgery

Severe thoracic scoliosis was treated surgically in adolescence in one case, and another patient had a 7° lumbar scoliosis under surveillance [11, 20]. An atrial septal defect was surgically repaired in childhood [11]. Divergent strabismus led to planned surgery in the first year of life, and refractive errors were corrected with glasses from 18 months [19]. None of these surgeries had reported complications, and there are no specific data on anaesthesia [11, 19, 20].

7.8. Neuromuscular component

In the patient with weakness and fatigability, recognition of neuromuscular involvement makes it possible to adapt physiotherapy and rehabilitation to weakness and contractures, select medication and organise early social support; a wheelchair for long distances, ankle-foot orthoses and a bath lift were used [11].

7.9. Reference dosing

Table 7.1. Drugs used in symptomatic treatment (label doses; use in this disorder by extrapolation)

7.10. Comparative regulatory status

Table 7.2. Regulatory status of symptomatic drugs

7.11. Clinical trials of symptomatic drugs

There are no trials in patients with KMT5B variants. The trials below support the label indications and were carried out in populations with ASD or conduct disorder with ID.

Table 7.3. Pivotal trials cited in the labels

These trials did not include patients with a KMT5B diagnosis identified as such, and transposing the results assumes that the response of insomnia or irritability does not depend on the genetic aetiology, an untested assumption. The trials are short (6–13 weeks, with a 2-year open-label extension for melatonin), do not allow indirect comparisons between drugs and report mean differences without confidence intervals in the sources consulted. In patients with hypotonia, somnolence and obesity, the adverse effects of antipsychotics may weigh more than in the trial populations.

7.12. Developments and research (past 12 months)

Between November 2025 and October 2026, published research was phenotypic or preclinical, with no product in clinical development. In human studies, the neuromuscular phenotype with denervation on biopsy (December 2025) [11], FDG-PET and RNA-seq profiles in two adolescents (October 2025) [7], a family with variable expressivity and obsessive-compulsive disorder in adulthood (2026) [21], two de novo missense variants with disturbance of the DNA damage response (2026) [30] and parkinsonism (June 2026) [22] were described. In a preclinical study, it was shown in Xenopus that the catalytic activity of KMT5B promotes ciliogenesis without altering global chromatin accessibility [34].

The compound A-196 is the first potent and selective inhibitor of KMT5B and KMT5C (50% inhibitory concentration of 25 nM and 144 nM, respectively); it increases H4K20me1 and decreases H4K20me2/3 and has been studied, with analogues, to increase frataxin expression in Friedreich ataxia [24]. Because it inhibits the enzyme, it has no therapeutic application in a loss-of-function disorder; its interest is mechanistic [24].

On ClinicalTrials.gov, the search of 28 October 2026 for “KMT5B”, “SUV420H1” and “intellectual developmental disorder 51” found only the observational Simons Searchlight study (NCT01238250, recruiting since October 2010, which includes the gene) and an out-of-scope hepatology study (NCT06332677); there is no therapeutic trial [15]. A public fundraising campaign in Poland (Fundacja Siepomaga), started on 12 June 2026 and planned to run until 1 January 2027, seeks to fund research to develop a gene therapy; on 28 October 2026 it had raised 84,024 zlotys (1.97% of the target), the promoters state that research will only begin when one million euros is reached and the research group is not identified [58]. This is a fundraising initiative, not a published preclinical study or a registered trial. No regulatory applications, orphan designations or authorisations were identified; the FDA and EMA orphan designation databases were not specifically consulted, so the absence of results does not prove non-existence.

7.13. Agents to use with caution, pregnancy, perioperative care and transition

There is no disease-specific list of contraindicated agents [3]. Precautions arise from comorbidities: methylphenidate lowers the seizure threshold [55]; antipsychotics worsen metabolic risk in patients with obesity [52, 54]; and carbamazepine requires HLA-B*15:02 testing in patients of at-risk Asian ancestry [56].

Several women with KMT5B variants and a mild picture have had affected children [5, 20], but there are no data on disease-specific obstetric complications; one pregnancy with an affected fetus had intrauterine growth restriction from 28 weeks and caesarean delivery at 34 weeks and 6 days, and another had reduced fetal movements, isolated events [7, 11]. For the symptomatic drugs, the Slenyto® SmPC advises, as a precaution, avoiding melatonin in pregnancy, owing to lack of data, and notes that exogenous melatonin probably passes into breast milk, with unknown effect on the infant [51]. The Risperdal® SmPC advises against risperidone in pregnancy unless clearly necessary, describes extrapyramidal and withdrawal symptoms in newborns exposed in the third trimester, recommends not stopping it abruptly and states that it passes into milk in small amounts [53]. The Rubifen® SmPC describes, in a cohort of about 3,400 pregnancies exposed in the first trimester, a small increase in cardiac malformations (adjusted relative risk 1.3; 95% CI 1.0–1.6), corresponding to 3 additional cases per 1,000 pregnancies, and does not recommend methylphenidate in pregnancy unless delaying treatment poses a greater risk [55].

There are no disease-specific data on anaesthesia or surgery; the scoliosis, atrial septal defect and strabismus surgeries took place without reported complications [11, 19, 20]. The authors of the Turkish report recommend prolonged clinical surveillance and lifelong neuropsychiatric assessment, because psychiatric symptoms may appear or be recognised only in adulthood, which should form part of the transition to adult care [21].

7.14. Guideline recommendations and divergences

Table 7.4. General guidelines applied and their relationship to the disorder

8. Reproductive Options and Family Counselling

8.1. Inheritance pattern and origin of variants

Inheritance is autosomal dominant [2, 3]. Most variants arise de novo: 90% of the 33 patients with both parents tested in the 43-patient cohort and 86.5% (45/52) of patients with known origin in the review of 64 patients [5, 7]. The 2025 review records 6 inherited variants and 1 case of probable germline mosaicism [7]. In the Sheppard cohort, the variants p.(Asp222Tyr) and p.(Arg395Ter) were inherited from the mother and p.(Pro444Arg) from the father, and in both missense cases the transmitting parent is also included in the cohort as a patient; the variant p.(Arg220Ter) was found in two non-twin siblings and was not detected in the blood of either parent, which indicates probable germline mosaicism in one of them [5]. In the case reports, the variant was inherited from a mother with mild ID [20], from a father with ID and obsessive-compulsive disorder [21] and from a father with childhood delay, in whom the variant had arisen de novo, since it was absent in the paternal grandparents [27].

8.2. Recurrence risks

Table 8.1. Recurrence risks by scenario

8.3. Parental testing and cascade testing

After diagnosis, both parents are tested for the variant. Because the transmitting parents described are often more mildly affected (mild ID, stereotypies, learning difficulties or psychiatric symptoms recognised only in adulthood), the parental study includes a clinical assessment and not only segregation testing [20, 21, 27]. If a parent is a carrier, testing is extended to siblings and other relatives at risk, with counselling on variable expressivity [5]. The episignature test also detects healthy carriers and people with incomplete penetrance, which may help interpret a variant of uncertain significance in a parent without evident signs [14].

8.4. Prenatal diagnosis and preimplantation genetic testing

When the familial variant is known and classified as pathogenic or likely pathogenic, prenatal diagnosis by amniocentesis or chorionic villus sampling and preimplantation genetic testing are technically possible, by targeted molecular tests [23]. The episignature should not be used on prenatal samples, because the signature was characterised in peripheral blood [2]. After a child with a de novo variant, prenatal diagnosis may be discussed because of the risk of germline mosaicism [5]. A positive prenatal result does not make it possible to predict severity, which ranges from normal intelligence with ASD to profound ID [7, 20].

8.5. Content of counselling

Counselling covers the dominant mechanism, the predominance of de novo variants, the possibility of transmission by mildly affected parents, variable expressivity within the same family and the absence of predictors of severity [3, 5]. The parents in a Chinese family reported that the genetic diagnosis helped them understand their son's problem but increased the psychological burden, an aspect to address in counselling [27]. In asymptomatic minors, predictive testing is justified only when it influences present care; in practice, siblings with signs of delay should be assessed clinically and tested as patients. In adults with ID, the decision on testing and on reproduction should respect decision-making capacity and the support needed. Family registries (Simons Searchlight) and the ERN ITHACA network are resources for families and professionals [10, 16].

9. Special Considerations: Population Diversity and Epidemiology

9.1. Prevalence: what exists and what does not

There is no published estimate of prevalence or incidence; OMIM, ClinGen and GenCC give no frequencies, and Orphanet records no prevalence for the entity [1, 3, 8]. The figures available are case counts: 43 patients in the largest cohort (2023), at least 59 people described by 2024 and 64 patients with clinical characterisation up to March 2025 [5, 7, 10]. These counts reflect ascertainment effort and not frequency in the population [3]. Large-scale sequencing studies do not provide a prevalence either: 7 probands with clinical follow-up among more than 11,730 patients with NDD sequenced for 208 genes [18], 2 de novo truncating variants in 4,293 DDD trios and 7 deletions in more than 41,800 individuals in DECIPHER [13] and diagnoses in the reanalysis of 4,351 German exomes [36]. The presence of two cohort variants in the neuropsychiatric subset of gnomAD suggests underdiagnosis [5]. As a general reference, ID affects 1–2.5% of the population and severe ID (IQ below 50) 0.3–0.5%, with a male:female sex ratio of 1.3:1 [23].

9.2. Geographical distribution and ancestry

Published cases come mainly from the United States and Europe. The 43-patient cohort was recruited through a US laboratory and centres in several countries (United States, France, Belgium, Germany, the Netherlands, Spain, Italy, Switzerland, Norway, the United Kingdom, Singapore and Australia, according to the authors' affiliations), with no published distribution of patients by country [5]. Individual reports and small series come from Israel (one Jewish patient of mixed ancestry, one Ashkenazi and one Muslim Arab) [19], the United Kingdom and Ireland [13], Italy [7], Germany [11, 36], Croatia [20], Spain [22], Turkey [21] and China [27, 30, 31]. No founder populations or recurrent variants associated with an ancestry have been described [7]. The sources consulted include no cases from Portugal or Brazil; absence from the sources does not prove absence of patients in these countries.

9.3. Sex ratio

Published cases show an excess of male patients (58% in the 2025 review; 25 males and 17 females in the 2023 cohort), without statistical significance [5, 7]. Macrocephaly is more frequent in males, and the hypothesis of higher male penetrance has not been tested in series with sufficient power [5, 7].

9.4. Special populations

In adults, the phenotype is poorly described, with cases up to 50 years [7]. The published adults include a 35-year-old man with overgrowth, morbid obesity, low testosterone and profound ID [19], carrier parents with mild ID or obsessive-compulsive disorder recognised only in adulthood [20, 21] and a report of parkinsonism [22]. Adolescents may develop obesity with insulin resistance [7]. In patients of Asian ancestry, HLA-B*15:02 testing before carbamazepine is mandatory according to the label [56].

9.5. Reference networks and patient associations

The relevant European reference network is ERN ITHACA (rare malformation syndromes, intellectual disability and other neurodevelopmental disorders) [16]. For the neuromuscular component and for epilepsy, the EURO-NMD and EpiCARE networks are relevant. Simons Searchlight maintains a page on the gene, with family stories, quarterly data reports (the latest, “KMT5B-2026Q3”, from October 2026) and a community group on a social network [10]. SFARI Gene and the NDD GeneHub maintain records of the gene for research [17]. No foundation dedicated exclusively to KMT5B was identified; the Polish fundraising campaign for gene therapy is the only patient initiative with a therapeutic aim found [58].

10. Illustrative Clinical Case

The case was published in 2023 in the Croatian Medical Journal by Odak and colleagues and describes a Croatian family with an in-frame deletion in the SET domain inherited from a mildly affected mother [20].

A boy was born at term, without perinatal complications, weighing 3,500 g, measuring 52 cm and with an Apgar score of 10/10, to non-consanguineous parents; the mother had mild ID and autistic traits, with hand stereotypies, and the father had schizophrenia, with no other NDD in the family [20]. Somatic growth was normal, but the infant had hypotonia and motor, speech and language delay; he walked at 18 months and never acquired speech [20]. He followed an early intervention programme, speech therapy, occupational therapy and physiotherapy. At 2 years, assessment showed no visceral anomalies, routine biochemical, haematological and metabolic tests were normal, vision was normal and hearing assessment suggested central-type hearing loss; at 3 years, global developmental delay was confirmed [20].

At 10 years, generalised tonic-clonic seizures appeared, treated with carbamazepine. Brain magnetic resonance imaging showed bilateral periventricular nodular heterotopia next to the caudate nucleus and asymmetry of the caudate itself (Figure 1 of the original article, coronal section with the heterotopias marked by arrows, https://pmc.ncbi.nlm.nih.gov/articles/PMC10668041/) [20]. In adolescence, he developed severe thoracic scoliosis, treated surgically [20].

At 16 years, he had severe ID, tall stature, thoracic scoliosis and facial dysmorphism with a long face, thick eyebrows, upslanting palpebral fissures, a broad and elongated nasal root and a narrow mouth (Figure 2 of the original article); he had not acquired sphincter control or speech, and autistic behaviour manifested as lack of social skills, poor emotional control and ADHD [20]. Muscle mass was reduced, and there was no macrocephaly or congenital heart disease [20].

The genetic work-up, summarised in the timeline of Figure 3 of the original article, included conventional karyotype, chromosomal microarray, subtelomeric and microdeletion MLPA and fragile X testing, all negative [20]. Clinical exome sequencing (TruSight One panel), with analysis guided by Human Phenotype Ontology terms, identified in exon 7 of KMT5B the heterozygous in-frame deletion NM_017635.4:c.715_717del p.(Glu239del) (Figure 4 of the original article), also present in the mother and absent from gnomAD [20]. On normalisation to the current MANE transcript, the variant corresponds to NM_017635.5:c.715_717del, NC_000011.10:g.68171649_68171651del [6]. The authors classified it as likely pathogenic by the criteria PM1 (SET domain, with 18 missense or in-frame variants recorded in UniProt, 8 pathogenic, 10 uncertain and none benign), PM4 (protein length change outside a repetitive region), PM2 at supporting strength and PP5 [20].

The case illustrates several points of this article. The deletion of a single amino acid in the SET domain produced a severe picture, whereas the mother, with the same variant, had mild ID and was almost independent, which exemplifies variable expressivity and the importance of clinically assessing the parents [20]. The absence of macrocephaly and heart disease shows that these features, present in 27–63% of patients, are not necessary for the diagnosis [5]. Periventricular nodular heterotopia was described here for the first time in the disorder, and the hypotonia and reduced muscle mass, absent in the mother, are consistent with the murine models of muscle involvement [20, 33]. The authors do not exclude a paternal polygenic contribution, given the history of schizophrenia, and designate the protein as “SUV420H2”, when the correct designation is SUV420H1 [4, 20]. Carbamazepine was the authors' individual choice for generalised seizures and does not constitute a disease-specific recommendation.

11. Prognosis and Long-Term Follow-up

11.1. Prognosis

Prognosis is determined by the severity of ID and comorbidities, which range from normal intelligence with ASD and DCD to profound ID without speech [7, 19, 20]. There is no established predictor of severity, either by variant type or by its location [3, 7]. Some patients do not acquire speech by 7 or 16 years [20, 27]; others, with nonsense variants, have IQs of 96 and 104 in adolescence [7]. The transmitting parents described reached adulthood with mild pictures and, in one family, the signs of childhood delay attenuated in adulthood [20, 27]. Conversely, the neuromuscular case had progressive weakness in adolescence [11], and in adulthood overgrowth with obesity, psychosis, obsessive-compulsive disorder and parkinsonism have been described [7, 19, 21, 22]. Life expectancy has not been studied; the oldest published patient was 50 years old, and there are no mortality data [3, 7].

11.2. Surveillance

There is no disease-specific surveillance guideline, nor a GeneReviews chapter with a surveillance table [3]. The frequencies of congenital heart disease (27%) and seizures (23–31%) justify cardiac and neurological assessments at diagnosis, but no source defines the modality or interval of follow-up [3, 5, 7]. The following table brings together the items with a source; intervals are given only when a source defines them (drug labels and the review of undiagnosed ID); otherwise, the interval is clinical and individual.

Table 11.1. Surveillance

11.3. Transition to adulthood

Because psychiatric symptoms may appear or be recognised only in adulthood and the adult phenotype is poorly described, the authors of the Turkish report recommend prolonged clinical surveillance and lifelong neuropsychiatric assessment [21]. Transition should ensure continuity of metabolic, orthopaedic, neurological and mental health surveillance, and review of genetic counselling when the patient reaches reproductive age [5, 7].

12. Summary and Final Considerations

KMT5B-related neurodevelopmental disorder (IDD AD 51, OMIM 617788; ORPHA:684226) results from haploinsufficiency of a methyltransferase that converts H4K20me1 into H4K20me2 and H4K20me3, marks of heterochromatin and of DNA break repair. The gene–disease association is definitive and haploinsufficiency has sufficient evidence in ClinGen. Truncating variants, deletions including the gene and missense variants, mainly in the SET domain, produce overlapping pictures; the most recent functional studies indicate that missense variants also act through loss of function. In the largest cohort (43 patients), all patients with information had developmental delay or ID, and hypotonia (81%), non-specific facial dysmorphism (79%), macrocephaly (63%, more frequent in males), ASD (62%), seizures (31%), septal congenital heart disease or patent ductus arteriosus (27%) and tall stature (21%) complete the phenotype; macrocephaly, and not systemic overgrowth, is the core growth feature. The spectrum includes normal intelligence with ASD and DCD, neuromuscular involvement with denervation, periventricular nodular heterotopia and, in adults, psychiatric manifestations and parkinsonism. There is no published prevalence; about 64 patients have been described, mainly from the United States and Europe, without founder populations.

Diagnosis is molecular, by exome or genome (with microarray for deletions), and the DNA methylation episignature serves as a second-line test for variants of uncertain significance, which are frequent in this gene. The table of one of the reference articles contains an invalid notation (c.764A>T p.(Asn255Ile)), corrected to c.833A>T p.(Asn278Ile). About 86–90% of variants are de novo, but there is transmission by mildly affected parents and probable germline mosaicism, which makes it necessary to assess the parents clinically and to inform about recurrence risk. There is no specific treatment, clinical trial or dedicated guideline; management is symptomatic, with rehabilitation, treatment of sleep, irritability, ADHD and epilepsy by extrapolation from labels and general guidelines, and with cardiac, neurological, ophthalmological, orthopaedic, metabolic and psychiatric surveillance whose intervals the sources do not define. Therapeutic research is at the preclinical stage, centred on the DNA damage response, the p53–DDIT4 pathway, muscle and cilia.

13. References

1. Orphanet. Intellectual disability-hypotonia-facial dysmorphism-macrocephaly syndrome (ORPHA:684226) [Internet]. Paris: INSERM; 2026 [record of 2026-07-02; accessed 2026-10-28]. Available from: https://www.orpha.net/en/disease/detail/684226
2. Wallace SE, Mirzaa GM, Bean LJH. Resources for Genetics Professionals — Epigenetic Signature Analysis. 2019 Dec 5 [revised 2022 Jun 30]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK550348/
3. GeneGuidance. Gene and condition curation: KMT5B [database]. Version of 26 September 2026; accessed 28 Oct 2026.
4. ClinGen. KMT5B (HGNC:24283) — curation results: gene-disease validity (Syndromic Disorders GCEP, 2022-04-08) and dosage sensitivity (2023-06-13) [Internet]. Clinical Genome Resource [accessed 2026-10-28]. Available from: https://search.clinicalgenome.org/kb/genes/HGNC:24283
5. Sheppard SE, Bryant L, Wickramasekara RN, Vaccaro C, Robertson B, Hallgren J, et al. Mechanism of KMT5B haploinsufficiency in neurodevelopment in humans and mice. Sci Adv. 2023;9(10):eade1463. doi:10.1126/sciadv.ade1463
6. Freeman PJ, Hart RK, Gretton LJ, Brookes AJ, Dalgleish R. VariantValidator: accurate validation, mapping, and formatting of sequence variation descriptions. Hum Mutat. 2018;39(1):61-8. doi:10.1002/humu.23348
7. Politano D, Borgatti R, Borgonovi G, Cistaro A, Danesino C, Fania P, et al. Bridging genotype to phenotype in KMT5B-related syndrome: evidence from RNA-seq, 18FDG-PET, clinical deep phenotyping in two new cases, and a literature review. Genes (Basel). 2025;16(10):1174. doi:10.3390/genes16101174
8. OMIM® Online Mendelian Inheritance in Man. #617788 INTELLECTUAL DEVELOPMENTAL DISORDER, AUTOSOMAL DOMINANT 51; MRD51. Johns Hopkins University; last edited 27/04/2022. https://omim.org/entry/617788. Accessed 28 October 2026.
9. OMIM® Online Mendelian Inheritance in Man. *610881 LYSINE METHYLTRANSFERASE 5B; KMT5B. Johns Hopkins University; last edited 27/04/2022. https://omim.org/entry/610881. Accessed 28 October 2026.
10. Simons Searchlight. KMT5B [Internet]. New York: Simons Foundation [accessed 2026-10-28]. Available from: https://www.simonssearchlight.org/research/what-we-study/kmt5b/
11. Aksel Kilicarslan O, Gangfuß A, Kölbel H, Muhmann D, Polavarapu K, Thompson R, et al. Combined histological and proteomic analysis reveals muscle denervation in KMT5B-related neurodevelopmental disorder: a case report. J Clin Med. 2025;14(24):8636. doi:10.3390/jcm14248636
12. Manickam K, McClain MR, Demmer LA, Biswas S, Kearney HM, Malinowski J, et al. Exome and genome sequencing for pediatric patients with congenital anomalies or intellectual disability: an evidence-based clinical guideline of the American College of Medical Genetics and Genomics (ACMG). Genet Med. 2021;23(11):2029-37. doi:10.1038/s41436-021-01242-6
13. Faundes V, Newman WG, Bernardini L, Canham N, Clayton-Smith J, Dallapiccola B, et al. Histone lysine methylases and demethylases in the landscape of human developmental disorders. Am J Hum Genet. 2018;102(1):175-87. doi:10.1016/j.ajhg.2017.11.013
14. Greenwood Genetic Center. EpiSign v5 [test brochure]. Greenwood (SC): GGC; 2025 [accessed 2026-10-28]. Available from: https://ggc.org/wp-content/uploads/2025/10/EpiSign-v5-brochure_Final.pdf
15. ClinicalTrials.gov. Search “KMT5B”, “SUV420H1” and “intellectual developmental disorder 51” (API v2) [Internet]. Bethesda (MD): National Library of Medicine [accessed 2026-10-28]. Available from: https://clinicaltrials.gov/search?term=KMT5B
16. ERN ITHACA — European Reference Network for Rare Malformation Syndromes, Intellectual and Other Neurodevelopmental Disorders [Internet] [accessed 2026-10-28]. Available from: https://ern-ithaca.eu/
17. Simons Foundation. SFARI Gene — KMT5B [Internet]. New York: Simons Foundation [accessed 2026-10-28]. Available from: https://gene.sfari.org/database/human-gene/KMT5B
18. Stessman HA, Xiong B, Coe BP, Wang T, Hoekzema K, Fenckova M, et al. Targeted sequencing identifies 91 neurodevelopmental-disorder risk genes with autism and developmental-disability biases. Nat Genet. 2017;49(4):515-26. doi:10.1038/ng.3792
19. Eliyahu A, Barel O, Greenbaum L, Zaks Hoffer G, Goldberg Y, Raas-Rothschild A, et al. Refining the phenotypic spectrum of KMT5B-associated developmental delay. Front Pediatr. 2022;10:844845. doi:10.3389/fped.2022.844845
20. Odak L, Vulin K, Meašić AM, Šamadan L, Tripalo Batoš A. Neurodevelopmental disorder caused by an inherited novel KMT5B variant: case report. Croat Med J. 2023;64(5):334-8. doi:10.3325/cmj.2023.64.334
21. Özcan S, Yeter B. A novel KMT5B frameshift variant presenting with autism and psychiatric features: intrafamilial phenotypic variation — a case report. Mol Syndromol. 2026;17(4):434-40. doi:10.1159/000549246
22. Rábano-Suárez P, Sánchez-Ferro Á, Arteche-López A, Quesada JF, Pérez de la Fuente R, Morales M, et al. Parkinsonism in KMT5B-related disorders: expanding the spectrum. Mov Disord Clin Pract. 2026;13(10):2617-20. doi:10.1002/mdc3.70701
23. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 188–193 (Intellectual disability).
24. Liu J, Guo X, Wang Y, Ma W. KMT5B in cancerous and noncancerous diseases: clinical and mechanical considerations. Int J Med Sci. 2026;23(5):1720-37. doi:10.7150/ijms.128796
25. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 272–275 (Overgrowth).
26. Mitchel MW, Myers SM, Heidlebaugh AR, Taylor CM, Rea H, Neuhaus E, et al. CHD8-Related Neurodevelopmental Disorder with Overgrowth. 2022 Oct 27. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK585456/
27. Tong J, Chen X, Wang X, Men S, Liu Y, Sun X, et al. Novel KMT5B variant associated with neurodevelopmental disorder in a Chinese family: a case report. Heliyon. 2024;10(7):e28686. doi:10.1016/j.heliyon.2024.e28686
28. gnomAD. KMT5B (ENSG00000110066) — constraint metrics, GRCh38 [Internet]. Broad Institute; accessed via the GraphQL API on 2026-10-28. Available from: https://gnomad.broadinstitute.org/gene/ENSG00000110066?dataset=gnomad_r4
29. Wickramasekara RN, Robertson B, Hulen J, Hallgren J, Stessman HAF. Differential effects by sex with Kmt5b loss. Autism Res. 2021;14(8):1554-71. doi:10.1002/aur.2516
30. Hu X, Tao R, Wang Y, Lin X, Ye C, Yang L, et al. Novel KMT5B de novo variants disrupt DNA damage response in intellectual disability. QJM. 2026;119(6):463-72. doi:10.1093/qjmed/hcaf327
31. Chen G, Han L, Tan S, Jia X, Wu H, Quan Y, et al. Loss-of-function of KMT5B leads to neurodevelopmental disorder and impairs neuronal development and neurogenesis. J Genet Genomics. 2022;49(9):881-90. doi:10.1016/j.jgg.2022.03.004
32. Wang ZJ, Rein B, Zhong P, Williams J, Cao Q, Yang F, et al. Autism risk gene KMT5B deficiency in prefrontal cortex induces synaptic dysfunction and social deficits via alterations of DNA repair and gene transcription. Neuropsychopharmacology. 2021;46(9):1617-26. doi:10.1038/s41386-021-01029-y
33. Hulen J, Kenny D, Black R, Hallgren J, Hammond KG, Bredahl EC, et al. KMT5B is required for early motor development. Front Genet. 2022;13:901228. doi:10.3389/fgene.2022.901228
34. Tait J, Marthen C, Hoelscher B, Straub T, Hsam O, Rupp RA. Catalytic activity of KMT5B promotes ciliogenesis without affecting global chromatin accessibility. Life Sci Alliance. 2026;9(6):e202503455. doi:10.26508/lsa.202503455
35. National Center for Biotechnology Information. ClinVar — KMT5B[gene] [Internet]. Bethesda (MD): NCBI [accessed 2026-10-28]. Available from: https://www.ncbi.nlm.nih.gov/clinvar/?term=KMT5B%5Bgene%5D
36. Trinh J, Kandaswamy KK, Werber M, Weiss MER, Oprea G, Kishore S, et al. Novel pathogenic variants and multiple molecular diagnoses in neurodevelopmental disorders. J Neurodev Disord. 2019;11(1):11. doi:10.1186/s11689-019-9270-4
37. Timberlake AT, McGee S, Allington G, Kiziltug E, Wolfe EM, Stiegler AL, et al. De novo variants implicate chromatin modification, transcriptional regulation, and retinoic acid signaling in syndromic craniosynostosis. Am J Hum Genet. 2023;110(5):846-62. doi:10.1016/j.ajhg.2023.03.017
38. Jans D, Al B, Bouman A, Geelen JM, Kleefstra T, Placek K. Clinical immunology in chromatinopathies: a scoping review. Front Immunol. 2026;17:1773284. doi:10.3389/fimmu.2026.1773284
39. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 224–226 (Macrocephaly).
40. Ocansey S, Cole TRP, Rahman N, Tatton-Brown K. Sotos Syndrome. 2004 Dec 17 [updated 2025 Jun 5]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1479/
41. Ostrowski PJ, Tatton-Brown K. Tatton-Brown-Rahman Syndrome. 2022 Jun 30. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK581652/
42. Moeschler JB, Shevell M; Committee on Genetics. Comprehensive evaluation of the child with intellectual disability or global developmental delays. Pediatrics. 2014;134(3):e903-18. doi:10.1542/peds.2014-1839
43. Hyman SL, Levy SE, Myers SM; Council on Children with Disabilities, Section on Developmental and Behavioral Pediatrics. Identification, evaluation, and management of children with autism spectrum disorder. Pediatrics. 2020;145(1):e20193447. doi:10.1542/peds.2019-3447
44. Alegret-García A, Cáceres A, Sevilla-Porras M, Pérez-Jurado LA, González JR. Analysis methods for diagnosing rare neurodevelopmental diseases with episignatures: a systematic review of the literature. Biomedicines. 2025;13(12):3043. doi:10.3390/biomedicines13123043
45. Ocansey S, Tatton-Brown K. EZH2-Related Overgrowth. 2013 Jul 18 [updated 2024 Mar 21]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK148820/
46. OMIM® Online Mendelian Inheritance in Man. #618786 IMAGAWA-MATSUMOTO SYNDROME; IMMAS. Johns Hopkins University; last edited 24/12/2021. https://omim.org/entry/618786. Accessed 28 October 2026.
47. Priolo M. NFIX-Related Malan Syndrome. 2024 Aug 1. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK605494/
48. Pappas J, Rabin R. SETD2 Neurodevelopmental Disorders. 2021 Dec 30 [revised 2022 Sep 22]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK575927/
49. Tramontana T, Yehia L. PTEN Hamartoma Tumor Syndrome. 2001 Nov 29 [updated 2026 Jul 28]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1488/
50. Williams Buckley A, Hirtz D, Oskoui M, Armstrong MJ, Batra A, Bridgemohan C, et al. Practice guideline: treatment for insomnia and disrupted sleep behavior in children and adolescents with autism spectrum disorder: report of the Guideline Development, Dissemination, and Implementation Subcommittee of the American Academy of Neurology. Neurology. 2020;94(9):392-404. doi:10.1212/WNL.0000000000009033
51. European Medicines Agency. Slenyto (melatonin) prolonged-release tablets: EPAR — product information (summary of product characteristics), sections 4.1 and 4.2. EMEA/H/C/004425; first authorisation 20 Sep 2018; renewal 5 Jun 2023; extension to neurogenetic disorders (II/0025) 26 Aug 2024; ADHD indication (II/0028) 17 Mar 2025; product information updated on 10 Jun 2026 [accessed 7 Nov 2026]. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/slenyto
52. Janssen Pharmaceuticals. Risperidone (RISPERDAL) tablets, oral solution and orally disintegrating tablets: prescribing information. NDA020588. Silver Spring (MD): FDA/DailyMed; revision 2026-08-20 [accessed 2026-10-28]. Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=701e1cd8-2868-4d29-85a9-a2212c88165f
53. INFARMED. Risperdal 1 mg/ml solução oral — Resumo das Características do Medicamento. Janssen Farmacêutica Portugal [Internet]. Lisbon: INFARMED (Infomed) [accessed 2026-10-28]. Available from: https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=7607&tipo_doc=rcm
54. Otsuka America Pharmaceutical. ABILIFY (aripiprazole) tablets: prescribing information. NDA021436. Silver Spring (MD): FDA/DailyMed; revision 2025-01-29 [accessed 2026-10-28]. Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=c040bd1d-45b7-49f2-93ea-aed7220b30ac
55. INFARMED. Rubifen 5 mg, 10 mg e 20 mg comprimidos — Resumo das Características do Medicamento [Internet]. Lisbon: INFARMED (Infomed); text revised September 2025 [accessed 2026-10-28]. Available from: https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=36077&tipo_doc=rcm
56. Novartis Pharmaceuticals. TEGRETOL (carbamazepine) tablets, chewable tablets, extended-release tablets and suspension: prescribing information. NDA016608/NDA018927/NDA020234. Silver Spring (MD): FDA/DailyMed; revision 2026-08-20 [accessed 2026-10-28]. Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=8d409411-aa9f-4f3a-a52c-fbcb0c3ec053
57. Agência Nacional de Vigilância Sanitária. Dados abertos — medicamentos registados (DADOS_ABERTOS_MEDICAMENTOS.csv) [Internet]. Brasília: ANVISA [accessed 2026-10-28]. Available from: https://dados.anvisa.gov.br/dados/DADOS_ABERTOS_MEDICAMENTOS.csv
58. Fundacja Siepomaga. Terapia KMT5B — financing research to develop gene therapy [Internet]. Poznań: Siepomaga [accessed 2026-10-28]. Available from: https://www.siepomaga.pl/en/terapia-kmt5b