CHD Gene-Related Disorders (Snijders Blok-Campeau, Sifrim-Hitz-Weiss, and Pilarowski-Bjornsson)

Neurodevelopmental disorders caused by heterozygous variants in the genes of the CHD (chromodomain helicase DNA-binding) family of ATP-dependent chromatin remodellers: Snijders Blok-Campeau syndrome (CHD3), Sifrim-Hitz-Weiss syndrome (CHD4), and Pilarowski-Bjornsson syndrome (CHD1), together with Parenti-Mignot neurodevelopmental syndrome (CHD5); with comparative mention of the CHD2, CHD7, and CHD8 entities.

Literature and regulatory information updated: 18 October 2026.

0. Reference Summary Table

1. Introduction

1.1 The CHD family and its classification

CHD (chromodomain helicase DNA-binding) proteins are SNF2-superfamily chromatin remodellers that depend on adenosine triphosphate (ATP) hydrolysis and are distinguished by two tandem chromodomains capable of binding methylated histones [24]. The human genome contains nine CHD genes, grouped into three subfamilies according to the architecture of their accessory domains: subfamily I (CHD1 and CHD2), with a C-terminal AT-rich DNA-binding domain; subfamily II (CHD3, CHD4, and CHD5), with two PHD (plant homeodomain) motifs, zinc fingers that read the state of histone tails; and subfamily III (CHD6, CHD7, CHD8, and CHD9), with BRK (Brahma and Kismet) or SANT domains [22, 28]. The subfamily II members serve, in mutually exclusive fashion, as the ATPase subunit of the NuRD (nucleosome remodelling and deacetylase) complex, which couples nucleosome remodelling to histone deacetylation by the deacetylases HDAC1 and HDAC2 and to methylated-DNA binding by the MBD2 protein [4, 13].

The first CHD gene linked to a human disease was CHD7, the cause of CHARGE syndrome (OMIM 214800); it was followed by CHD1 (Pilarowski-Bjornsson syndrome, OMIM 617682), CHD2 (epileptic encephalopathy, OMIM 615369), CHD4 (Sifrim-Hitz-Weiss syndrome, OMIM 617159), CHD8 (autism spectrum disorder with macrocephaly, OMIM 615032), and, in 2018, CHD3 (Snijders Blok-Campeau syndrome, OMIM 618205) [22]. In 2021, CHD5 was linked to the neurodevelopmental disorder later named Parenti-Mignot syndrome (OMIM 619873) [28, 29]. CHD6 has a proposed association with Hallermann-Streiff syndrome, and CHD9 has, to date, no pathogenic variants associated with a Mendelian disease [24]. This family of disorders shares a clinical core (developmental delay, speech and language disorder, intellectual disability of variable severity, craniofacial dysmorphism and, in several members, abnormal head circumference) and a common mechanistic basis, dysregulated nucleosome positioning, which justifies treating the entities as a class and using the DNA methylation profile as a molecular phenotype to distinguish them [24, 25]. The reference textbook on inherited metabolic diseases lists variants in NuRD components (CHD3, CHD4, and GATAD2B) among the genetic causes of macrocephaly with intellectual disability, alongside the overgrowth syndromes and the leukodystrophies [37].

1.2 Snijders Blok-Campeau syndrome (CHD3)

The syndrome was delineated starting from a child with childhood apraxia of speech (CAS) included in a whole-genome sequencing cohort of 19 unrelated children with CAS, in whom a de novo variant was identified in the helicase domain of CHD3, one of the few documented interaction partners of FOXP2 [21]. The same index case was published by Eising and colleagues in 2019, with the variant described on a different transcript (p.(Arg1228Trp), corresponding to p.(Arg1169Trp) on the MANE transcript) [1, 38]. Before this description, studies of simplex autism families had recorded two de novo nonsynonymous CHD3 variants, and the UK DDD (Deciphering Developmental Disorders) study a further eight [21]. Through the GeneMatcher platform, Snijders Blok, Rousseau, Twist, and colleagues, under the supervision of Fisher and Campeau, assembled 35 probands with de novo variants and published the delineation in Nature Communications in November 2018; the names of the first and last authors gave the syndrome its eponym [21]. The article received two errata in 2019, one of them to remove the images of one individual from the facial figure [1, 21].

A second cohort of 24 patients from eight countries (Australia, Canada, China, France, Germany, Italy, the Netherlands, and the United States [USA]) brought the total to 59 and extended the molecular mechanisms to truncating variants, variants outside the helicase domain, and whole-gene copy-number changes [22]. In 2022, van der Spek and colleagues, from member centres of ERN ITHACA and the Solve-RD project, described 21 families with inherited variants, mostly from mildly affected or unaffected parents, which recast the syndrome from an essentially de novo disorder into a dominant disorder with variable expressivity and reduced penetrance [12]. A third series of 20 patients from 17 families (Spain, India, Italy, and France, partly from the Spanish SOGRI overgrowth registry at Hospital Universitario La Paz) and a review of 63 published patients brought the total to 83 in 2023 [19]. In 2025 and 2026, publications described the neurobehavioural profile of 38 patients, a pain survey, two distinct phenotypes, the DNA methylation episignature, and a mouse model with postnatal gene correction [24, 39–42].

OMIM names the entity "Snijders Blok-Campeau syndrome" (SNIBCPS), with the alternative title "intellectual developmental disorder with macrocephaly, speech delay, and dysmorphic facies" (IDDMSF); the entry was created on 28 November 2018 and last edited on 13 June 2023 [1]. Orphanet uses a descriptive name (ORPHA:599082) and retains "Snijders Blok-Campeau syndrome" as a synonym; its definition refers to "moderate to severe" intellectual disability, a wording not supported by the primary cohorts, in which severity ranges from borderline to severe and includes patients without intellectual disability [2, 12, 19]. The eponym should not be confused with Snijders Blok syndrome, an X-linked intellectual developmental disorder caused by DDX3X variants (OMIM 300160); a 2025 case report titled as Snijders Blok-Campeau syndrome with pulmonary arterial hypertension actually describes a female infant with a DDX3X variant and does not constitute evidence of pulmonary hypertension in CHD3-related disease [43].

1.3 Sifrim-Hitz-Weiss syndrome (CHD4)

CHD4-related disease was characterised independently by two groups in 2016. Sifrim, Hitz, and colleagues, in exome sequencing of 518 trios with syndromic congenital heart disease, identified five patients with de novo CHD4 variants, a significant enrichment (p = 2.28×10⁻⁷) [3, 44]. Weiss and colleagues, using reverse phenotyping, described five individuals with de novo variants, intellectual disability and distinctive dysmorphic features, hypogonadotropic hypogonadism in all three males, cervical vertebral fusions, hearing loss, and a congenital stroke with moyamoya [3, 45]. Both groups noted the overlap with CHARGE syndrome [3]. In 2020, Weiss and colleagues described 32 individuals together with the clinical spectrum, genotype–phenotype correlations, and functional effects of the variants [13], and Weiss and Lachlan published the GeneReviews chapter "CHD4 Neurodevelopmental Disorder" (submitted 28 January 2020, published 3 September 2020, not subsequently updated) [4]. Pinard and colleagues extended the phenotype to moyamoya angiopathy [46], and in 2025 Karimi and colleagues described the methylation episignature and clinical-molecular subtypes associated with truncating variants [25]. GeneReviews is out of date relative to these data (episignature, truncating-variant subtypes, aortic root dilatation, intestinal malrotation, and optic nerve anomalies) [4, 25].

1.4 Pilarowski-Bjornsson syndrome (CHD1)

In 2018, Pilarowski, Bjornsson, and colleagues at the Epigenetics and Chromatin Clinic of Johns Hopkins University described, via GeneMatcher and GeneDx, six female individuals with heterozygous missense CHD1 variants (three confirmed de novo), with developmental delay, speech apraxia, autism, and hypotonia [23]. Previously, three de novo CHD1 variants had been recorded in patients with autism in large genomic surveys [23]. The initial absence of boys led to the hypothesis of male lethality [14]. In September 2026, Anderson and colleagues published the largest cohort (56 individuals from 14 countries, with an excess of males) and a mouse model showing androgen-mediated sexual dimorphism; that article reports four de novo cases in the original series, whereas the 2018 article states three confirmed [14, 23]. The OMIM entry was last edited on 16 January 2024 and does not include this cohort [6]. ClinGen classified the CHD1–complex neurodevelopmental disorder relationship as "limited" in April 2024, before these data [18].

1.5 Other entities in the class and scope of the article

This article covers in depth the three subfamily I and II entities (CHD3, CHD4, and CHD1) and, given its membership in NuRD, Parenti-Mignot syndrome (CHD5), whose first cohort of 16 patients was published in 2021 [28]. The CHD2-related disorder (DEE94), the CHD7-related disorder (CHARGE syndrome and hypogonadotropic hypogonadism) and the CHD8-related disorder are included in the class comparison. These entities appear here in the description of the class and in the differential diagnosis. The association of CHD6 with Hallermann-Streiff syndrome is listed in Orphanet with a loss-of-function mechanism, but without ClinGen curation and with limited evidence in the sources consulted [18, 24, 30]. Orphanet assigns OMIM 615369 (DEE94) to Lennox-Gastaut syndrome (ORPHA:2382) and to epilepsy with myoclonic-atonic seizures (ORPHA:1942), rather than to an entity of its own; citing ORPHA:2382 as the code for CHD2-related DEE, as a 2026 article does, is therefore inaccurate [24, 30, 32]. GeneGuidance links OMIM 615032 (CHD8) to 14q11.2 microduplication syndrome (ORPHA:261229), but the correct Orphanet entity is CHD8 overgrowth syndrome (ORPHA:642675) [36, 47, 48].

1.6 Current classifications and class table

Three complementary classifications are in use: OMIM, by phenotype, with a separate number for each gene; Orphanet, with descriptive names and ICD-10 classification codes; and ClinGen, by gene–disease validity with the corresponding MONDO entity (for example, MONDO:0032600 for Snijders Blok-Campeau syndrome) [1, 2, 18, 24]. There is no accepted clinical subtype classification for any of the entities; the proposed clinical-molecular subtypes (two opposing phenotypes in CHD3 according to variant location; a truncating-variant subtype in CHD4) are hypotheses from small series [25, 41].

AD: autosomal dominant; HH5: hypogonadotropic hypogonadism 5; IDDAM: intellectual developmental disorder with autism and macrocephaly.

2. Molecular Genetics and Pathophysiology

2.1 Shared architecture and the NuRD complex

CHD genes encode proteins with two N-terminal chromodomains, an SNF2-type ATPase/helicase core made up of an ATP-binding lobe (helicase ATP-binding) and a C-terminal helicase lobe, and accessory domains specific to each subfamily [22, 28]. ATP hydrolysis in the SNF2 core slides nucleosomes along the DNA and alters the accessibility of promoters and enhancers. Within the NuRD complex, the remodelling activity of the CHD3, CHD4, or CHD5 subunit increases nucleosome density at promoters and enhancers, which can displace the transcriptional machinery and allow new factors to bind, with a repressive or activating effect depending on context [49]. In the mouse cortex, the three subunits act sequentially rather than interchangeably: CHD4 regulates early proliferation of basal progenitors, CHD3 late radial migration and cortical layer specification, and CHD5 neuronal differentiation, radial migration, and neuronal identity; loss of one subunit is not compensated by overexpression of another [21, 28]. CHD3 and CHD4 are ubiquitously expressed, whereas CHD5 is expressed mainly in the brain and testis [28]. CHD1, a subfamily I member, reads histone H3 lysine 4 trimethylation and dimethylation (H3K4me3 and H3K4me2) through its chromodomains, slides nucleosomes, participates in the SAGA complex, supports transcription by RNA polymerases I and II and deposition of histone H3.3, regulates chromatin opening and embryonic stem cell pluripotency, and has its highest brain expression in the cerebellum and basal ganglia [11, 23].

2.2 CHD3: gene, transcripts, and protein

CHD3 maps to 17p13.1 (GRCh38 17:7,884,796–7,912,755) and has 40 exons and three alternative transcripts [1, 50]. The MANE Select transcript is NM_001005273.3 (NP_001005273.1; ENST00000330494.12); transcript NM_001005271.3 (ENST00000380358) encodes an isoform 59 amino acids longer, which shifts the protein numbering so that the same variant appears in the literature under two names [9]. NM_001005273 is the most highly expressed transcript in brain tissue, according to data cited by Gao and colleagues [50]. The protein has two PHD domains, chromodomain 1 (amino acids 494–595), chromodomain 2 (631–673), the helicase ATP-binding and helicase C-terminal lobes (predicted helicase domain between amino acids 879 and 1258 of NP_001005273.1), ATP-binding residues 761–768, two domains of unknown function (DUF), and a C-terminal domain 2 (CHDCT2) [21, 22]. Beyond its role in NuRD, CHD3 anchors pericentrin at the centrosome and contributes to mitotic spindle organisation [8]. In BrainSpan, expression is high in the striatum, anterior cingulate cortex, hippocampus, cerebellar cortex, medial prefrontal cortex, and primary motor cortex; in the mouse, brain protein levels decline from the embryonic period to adulthood [42].

The gene is highly intolerant of variation: in ExAC, the probability of loss-of-function intolerance (pLI) is 1.0 and the missense Z score is +7.15; in gnomAD, pLI is 1, the observed/expected loss-of-function ratio is 0.09 (0.05–0.15), the missense Z score is 6.15, and the observed/expected missense ratio is 0.5 (0.46–0.53) [19, 21]. The DECIPHER haploinsufficiency index for CHD3 was 16.9 in 2020, a value that does not strongly support haploinsufficiency [13].

2.3 CHD3: variant spectrum

In the founding cohort, the 35 probands had 23 different de novo variants, none present in gnomAD, all missense except four [21]. Seventeen of the 19 distinct missense variants lay in the ATPase/helicase domain, at residues conserved across species and among CHD3, CHD4, and CHD5; all were predicted pathogenic (PolyPhen-2 and SIFT) and had a CADD (Combined Annotation Dependent Depletion) score above 24 [21]. Residue Arg985 was a hotspot: p.(Arg985Trp) in six children from five families (two siblings with a mosaic mother) and p.(Arg985Gln) in two more [15, 21]. The remaining variants were an in-frame deletion, p.(Gly1109del); two truncating variants (p.(Glu457Ter) and p.(Phe1935GlufsTer108), the latter at the extreme C-terminus, with escape from nonsense-mediated mRNA decay [NMD] demonstrated by RNA sequencing with and without cycloheximide); and a splice variant, c.4073-2A>G, predicted to cause in-frame skipping of exon 27 (22 amino acids, at the start of the second DUF domain) [21]. Proband 15 had, in addition to de novo p.(Asp1120His), a de novo truncating CIC variant (NM_015125.3:c.1444G>T p.(Glu482Ter), published designation; = MANE NM_001386298.1:c.4171G>T p.(Glu1391Ter), NC_000019.10:g.42289931G>T), raising the possibility of a dual contribution to the phenotype [21]. The missense variants cluster significantly within the transcript (P = 0.0017), an argument against simple haploinsufficiency [21].

The second cohort (24 patients) included 14 missense variants or in-frame deletions in or near the helicase domain (14 of 19 between amino acids 886 and 1262) and 10 alterations of other types: one whole-gene duplication and one whole-gene deletion, three truncating variants, one splice variant, and four missense variants outside the helicase domain (residues 569, in monozygotic twins, 1415, and 1955) [22]. The exon 27 splice-acceptor variant was interpreted as causing frameshifting exon skipping, contrary to the in-frame skipping predicted in 2018 for c.4073-2A>G [22]. The 0.5 Mb deletion (chr17:g.7394419_7871060del, hg19) encompassed CHD3, MPDU1, TP53, and WRAP53; the TP53 deletion carries its own implications (Li-Fraumeni syndrome) but does not explain the neurodevelopmental phenotype, and DECIPHER lists another 37 patients with deletions encompassing the whole of CHD3 [22]. The 6.5 Mb duplication (chr17:g.7339633_7902885dup, hg18) involving CHD3, CHRNB1, MPDU1, TP53, and WRAP53 occurred in a patient with a negative 1162-gene intellectual disability panel and in an equally affected sibling; the authors consider the evidence insufficient to assert pathogenicity of CHD3 duplications [22]. None of the recurrent 2018 variants reappeared except p.(His886Arg) and p.(Arg1172Gln), and a new recurrent residue, Arg966, emerged in four unrelated individuals; the variants were deposited in LOVD [22].

In the 2023 series, 16 of the 20 variants were missense, with one nonsense variant, one frameshift variant upstream of C-terminal domain 2, and two in-frame deletions; 8 were classified as pathogenic and 12 as likely pathogenic under the American College of Medical Genetics and Genomics and Association for Molecular Pathology (ACMG/AMP) criteria, none had been previously described or was present in population databases, and four lay in constrained coding regions in the 90th–100th percentile [19]. Of the 13 ATPase-domain variants in that review, half affected arginines (p.(Arg985Trp), p.(Arg1044Trp), p.(Arg1070Gln), p.(Arg1169Trp)) [19]. That article mixes transcripts NM_001005271.3 and NM_001005273.3 in the same table (position chr17:7903962 appears as c.3865G>A p.(Ala1289Thr) on one and as c.4042G>A p.(Ala1348Thr) on the other), and some coordinates stated as hg38 appear to be hg19, so its positions must be converted before comparison [9, 19]. In four Chinese patients, c.3592_3606del p.(Ala1198_Met1202del), c.1708-1G>T (with aberrant splicing demonstrated in a minigene assay), c.2954G>C p.(Arg985Pro), and c.3371C>T p.(Ala1124Val) were identified, three of them novel [50]. In the 2026 European cohort, 63% of patients had missense variants; truncating or frameshift variants clustered in the N- and C-terminal domains, whereas those in the chromodomains and the central helicase core were exclusively missense [24]. Two patients in that cohort had the synonymous variant NM_001005273.3:c.5754G>A, with a probable splicing effect confirmed by RNA studies in one of them [24].

Inherited variants have a different profile. In the 21 van der Spek families, 13 were missense and eight truncating (seven single-nucleotide variants and one intragenic deletion; 40% of variants with data); no inherited truncating variant was predicted to escape NMD, whereas de novo truncating variants accounted for 4 of 55 (7.3%) in earlier cohorts [12]. Inherited missense variants did not cluster in the ATPase/helicase domain, and in three-dimensional modelling de novo missense variants lay closer to the ATP-binding site than inherited ones; the inherited variant p.(Ile983Val), the closest to the de novo variants, occurred in the family whose transmitting parent had the most pronounced phenotype [12]. The p.(Arg1342Gln) variant, described as de novo in the founding cohort, was found inherited from an unaffected parent [12].

The dual numbering requires normalisation. Drivas's recurrent variant at residue Arg966 is the same as p.(Arg1025Trp) in the mouse model and the 2025 clinical trial, and the variant in Eising's 2019 index case (p.(Arg1228Trp)) is the same as p.(Arg1169Trp) in Snijders Blok 2018 [9, 15]. The conversions below were obtained with VariantValidator (GRCh38), not by manual calculation.

The other CHD3 variants listed in OMIM, all on NM_001005273.2, comprise, in addition to those already given, c.4073-3_4078del, a 76-base-pair deletion that removes the exon 27 acceptor (de novo, Drivas patient 21) [15]. The founding-cohort variants were annotated on GRCh37 and NM_001005273.2 (6003-base-pair coding region) and deposited in ClinVar (SCV000787629–SCV000787651) [21]. GeneGuidance gives the transcript without a version number; the current MANE version is .3 [8, 9].

2.4 CHD3: mechanisms

Functional assays in 2018 of six variants, using protein purified from HEK293 cells, showed heterogeneous effects [21]. ATPase activity was reduced for p.(Arg1121Pro) (motif V) and p.(Arg1172Gln) (motif VI) with both nucleosomes and double-stranded DNA, reduced only with DNA for p.(Asn1159Lys), increased for p.(Leu915Phe) (SNF2 motif III), and unchanged for p.(Arg1187Pro) and p.(Trp1158Arg) [21]. In remodelling assays (restriction enzyme accessibility), p.(Arg1121Pro) was severely impaired, p.(Arg1172Gln) partially impaired, p.(Leu915Phe) hyperactive, p.(Trp1158Arg) and p.(Asn1159Lys) severely impaired despite preserved ATPase activity, and only p.(Arg1187Pro) equivalent to wild-type protein [21]. Equimolar mixing of wild-type and mutant protein showed no biochemical interference, and subcellular localisation was normal [21]. The authors invoked the analogy with SMARCA4, in which ATPase-core variants act as dominant negatives whether they increase or decrease activity [21]. In 2026 structural modelling, using the AlphaFold model superimposed on structures of nucleosome-bound CHD1 and CHD4, residues Thr768, Arg1169, and Arg1172 interact directly with ATP in the active site, and variants associated with a severe phenotype lie mainly in the helicase domain [24].

Inherited variants point to loss of function. In van der Spek family 1, the p.(Trp1158Ter) variant was found in the proband, mother, and grandmother; in lymphoblastoid lines, cycloheximide increased the mutant allele (an NMD target), and CHD3 transcript and protein (about 227 kDa) were similarly reduced in all three carriers, affected or not, compared with the noncarrier sibling [12]. Compensation by the wild-type allele therefore does not explain the variable expressivity, at least in blood cells; extrapolation to neurons is uncertain, because CHD3 undergoes neuron-specific alternative splicing [12]. Chromatin-binding (p.(Ser477Phe)) and GATAD2B-binding (p.(Arg1342Gln), p.(Glu1837Lys), p.(Gln1888Arg)) assays with inherited missense variants showed no abnormalities [12]. The authors propose loss of function as the main mechanism in inherited cases and a possible dominant-negative effect for de novo ATPase-domain missense variants [12]. Other explanations for the variable expressivity include a second hit on the other allele, genomic mutational burden, and common regulatory variants, and four probands had an additional copy-number variant (one 22q11.2 duplication and three of uncertain significance) [12].

The p.(Arg1025Trp) variant (= p.(Arg966Trp) on MANE), located between the helicase ATP-binding lobe and the helicase C-terminal lobe, accelerates protein degradation via the ubiquitin–proteasome pathway: protein is reduced with normal mRNA, levels recover with a proteasome inhibitor, coexpression with wild-type protein does not reduce the total amount (arguing against a dominant-negative effect), and any aromatic amino acid at this position reduces the protein [42]. Chd3 knockdown in cortical neurons reduces dendritic, but not axonal, length and branching, and the defect is rescued by wild-type but not mutant protein [42]. In human induced pluripotent stem cells differentiated into cranial neural crest cells, heterozygous CHD3 loss had a modest effect (36 differentially expressed genes), but complete loss (1516 genes) prevented neural crest specification, with loss of accessibility at regulatory elements linked to bone morphogenetic protein (BMP)-responsive factors such as the DLX and MSX families, an imbalance between the BMP and Wnt pathways, and a shift towards early mesodermal identity that was partially corrected by Wnt titration [49]. These data suggest a mechanism for the craniofacial dysmorphism, in cellular models [49].

2.5 CHD3: genotype–phenotype correlations, penetrance, and allelic forms

In the founding cohort there was no evident correlation between variant type or location and severity; the only patient with epilepsy was the only one with a missense variant in the C-terminal domain [21]. In the second cohort, comparison of helicase-domain variants (n = 14) with others (n = 10) showed only one difference with p < 0.05 that did not survive Bonferroni correction (moderate or mild-to-moderate intellectual disability more frequent with helicase variants); lateral incisor agenesis occurred only in the helicase group (p = 0.1186), and composite facial masks did not differ [22]. In the 2023 series, variants inside (8) and outside (12) the ATPase/helicase domain were not associated with phenotypic differences [19]. Enomoto and colleagues proposed two opposing phenotypes according to location (helicase C-terminal, with macrocephaly and overgrowth; SNF2 domain or CHDCT2, with microcephaly and growth restriction), by analogy with types 1 and 2 of SPOP-related Nabais Sá-de Vries syndrome, with the caveat that earlier microcephalic patients with p.(Gly961Glu), p.(Arg966Trp), and p.(Phe944Tyr) had hypertelorism [41]. Carriers of p.(Arg966Trp) and p.(Arg966Gln) in the 2026 cohort lacked macrocephaly and had more severe intellectual disability than the rest of the cohort, with dysmorphism mainly affecting the ears; three of ten variants of uncertain significance were arginine substitutions [24]. The same p.(Arg1172Gln) variant was associated with differing severity in three patients [24].

Penetrance is incomplete and expressivity variable: all 17 heterozygous parents with data had at least one feature of the syndrome, but 5 of them (29%) had only one or two, and 15 of 19 (79%) had no history of intellectual disability [12]. Transmission was maternal in 15 of 21 families (71%; P = 0.0392), single-nucleotide truncating variants were maternally transmitted in six of seven, and the only father transmitting a loss-of-function variant had mild intellectual disability; probands showed no sex bias and no greater severity in boys, and the authors raise the possibility of a female protective effect [12]. In gnomAD, 12 of the 15 carriers of CHD3 loss-of-function variants among 198,800 individuals were women (P = 0.0173) [12]. In the UK Biobank (white British ancestry), rare missense variants at intolerant sites were not associated with fluid intelligence (n = 77,998), educational attainment (n = 120,596), or intracranial volume (n = 18,254), and the four carriers of rare loss-of-function variants had larger intracranial volume only at a nominal level (P = 0.018; Bonferroni P = 0.054) [12].

Nineteen rare CHD3 missense variants have been associated with Chiari I malformation in individuals without features of the syndrome (Sadler and colleagues, indirect attribution) [12]. Two individuals with a homozygous in-frame insertion, c.5384_5389dup p.(Arg1796_Phe1797insTrpArg), had a more severe phenotype than heterozygotes (more marked dysmorphism and severe intellectual disability), according to Goldfarb Yaacobi and Sukenik Halevy (2024, indirect attribution) [49]. One patient in the 2026 cohort with a severe phenotype and craniosynostosis was the child of a consanguineous couple, and the authors raise the possibility of a contribution from another genetic factor [24].

2.6 CHD4

CHD4 (12p13.31; GRCh38 12:6,570,082–6,607,379) is the core component of NuRD, with ubiquitous expression and roles in stem cell differentiation, embryonic development, the cell cycle, and DNA repair [3, 4]. In the 2020 cohort, 91% of variants were missense or in-frame indels and 82% lay in the SNF2-type region (approximately amino acids 724–1281, mainly in its C-terminal portion), with six in the N-terminal ATPase, 16 in the C-terminal helicase, and six in other conserved domains (CHDNT, PHD2, second chromodomain, DUF1087), with a similar phenotype [13]. Almost half clustered in a hotspot between amino acids 1127 and 1192 (SNF2 motifs V, Vb, and VI), with recurrent residues (Arg1127, Arg1173, Arg1183) and systemic and cognitive variability among carriers of the same variant (p.(Arg1127Gln) in one patient with moderate intellectual disability and in another with a normal intelligence quotient [IQ]) [13]. No genotype–phenotype correlation has been established [4, 13]. The gene has a pLI of 1 and a missense Z score of 7.05; ClinGen assigns it a haploinsufficiency score of 1 (little evidence) and a triplosensitivity score of 0, and whole-gene deletion is not evidence for this diagnosis [10, 46].

Assays of five variants in 2020 showed variant-specific effects: p.(Cys467Tyr) (PHD2) and p.(Arg645Trp) (chromodomain) reduced ATPase activity by 30–50% as well as remodelling; p.(Trp1148Leu) (helicase) and p.(Arg1340Cys) (DUF1087) did not alter ATPase activity but reduced remodelling; p.(Met954Val) (N-terminal ATPase) slightly increased remodelling, without statistical significance [13]. The authors propose a dominant-negative or gain-of-function mechanism for nontruncating variants, in which the mutant protein is incorporated into NuRD with aberrant function, by analogy with SMARCA4 and SMARCA2 [13]. GeneReviews regards the dominant-negative mechanism as suspected but unproven [4]. The three truncating variants in the 2020 cohort (9%) were associated with a phenotype involving fewer systems and a different facial gestalt: c.2143C>T p.(Gln715Ter) with delay, macrocephaly, and Pierre Robin sequence; c.2662delC p.(Val888TyrfsTer2) with delay and mild ventriculomegaly; and c.5608C>T p.(Arg1870Ter), C-terminal and probably escaping NMD, with delay, a thin corpus callosum, and an additional diagnosis of FLNA-related periventricular nodular heterotopia [13]. In 2025, the seven truncating variants studied had a distinct, less pronounced methylation profile, were mostly inherited, were extremely rare in population databases, and were associated with autism, suggesting a clinical-molecular subtype with incomplete penetrance; two individuals with compound heterozygous variants and an overlapping phenotype leave open the possibility of rare recessive inheritance [25]. In the 2025 clinical cohort of 27 patients, the 26 variants were de novo in 8, inherited in 7, of unknown inheritance in 11, and inherited in compound heterozygosity in one case [25]. CHD4 interacts with pericentrin through its C-terminal region, where three of the moyamoya-associated variants lie [46]. No other allelic germline phenotypes are known [4].

2.7 CHD1

The variants in the original series were missense; five of the six involved loss of an arginine (enrichment p = 9.3×10⁻⁵, adjusted for the protein's 7% arginine content), including p.(Arg618Gln), adjacent to the Walker B motif, and positions Arg141, Arg460, and Arg1708 [23]. Subject 5 had de novo p.(Asp857Gly) and compound heterozygosity in WDR62 (a cause of primary microcephaly 2), so this subject's phenotype was not attributed to CHD1 [23]. Fibroblasts from the subject with p.(Arg618Gln) showed a global increase in histone H3 lysine 27 trimethylation (more closed chromatin) [23]. A 9-year-old boy with a de novo deletion of RGMB and the last nine exons of CHD1 had only clubfoot and asthma, Chd1+/− mice are healthy and fertile, and the seven DECIPHER deletions encompassing the gene are large (2.95–12.10 Mb) and lack a consistent neurological phenotype; hence the initial hypothesis of a dominant-negative mechanism [11, 23].

The 2026 cohort comprised 52 variants in 56 individuals (38 males and 18 females; male excess p = 0.0065) [14]. Thirty-three had 29 missense variants (de novo in 52%, inherited in 18%, unknown in 30%) and 20 had truncating loss-of-function variants (de novo in 60%, inherited in 20%, unknown in 20%), with one de novo stop-loss variant and two possible noncanonical splice variants [14]. Under ACMG/AMP criteria, 24 of the 52 (46%) were likely pathogenic; 83% of missense variants were of uncertain significance, and AlphaMissense classified 41% of missense variants as deleterious, 14% as uncertain, and 45% as benign [14]. Recurrent variants were c.2090G>A p.(Arg697Gln) (two males), c.3335G>T p.(Arg1112Leu) (two females), and c.5123G>A p.(Arg1708Gln) (two sisters and one male; also present in unaffected parents and possibly benign) [14]. Truncating variants occurred in 15 males and 5 females, and in a separate laboratory cohort no female had nonsense or frameshift variants [14]. The p.(Arg618Gln) variant, immediately downstream of the Walker B motif (DEAHR motif II) and in contact with nucleosomal DNA, abolishes remodelling in vitro, and the authors conclude in favour of a loss-of-function mechanism, which makes the 2018 dominant-negative hypothesis less likely [14].

In the Chd1 R616Q mouse model (equivalent to p.(Arg618Gln)), no homozygotes were born among 96 pups from 18 litters, and heterozygotes were fewer than expected at weaning (57 versus 78; p = 2.72×10⁻⁹) but not between E14.5 and E18.5, with no reduction in mRNA or protein [14]. Phenotypes were restricted to females: low body weight, abnormal righting reflex on day 6 (hypotonia), anxiety-like behaviour in the open field, and adult motor deficits, without deficits in spatial memory or vocalisations [14]. Neural progenitors showed reduced proliferation (prolonged G1 phase), hypothalamic oxytocin was elevated, long genes were dysregulated, and there was an ID6 indel signature (homologous recombination deficiency) [14]. Dihydrotestosterone (10 nM) reversed the transcriptional dysregulation in neural progenitors (overlap of 425 genes; r = −0.916), an effect partly dependent on the androgen receptor and attenuated by enzalutamide; CHD1 binding sites were enriched for the androgen receptor motif [14]. Testosterone given to pregnant dams for ten days corrected neuronal proliferation at E18.5, orchiectomy on day 15 unmasked low body weight in mutant males, and a testosterone implant on day 15 corrected low body weight in females [14]. In gnomAD v4.1.0, rare CHD1 missense variants were more frequent in males (p = 3.65×10⁻¹⁰), a finding replicated in the UK Biobank (p = 0.019) but not in All of Us (p = 0.15); 33 other highly constrained autosomal genes showed an excess of missense variants in healthy males [14]. The authors interpret these data as an androgen-mediated male protective effect and possibly greater female embryonic lethality [14]. Somatic CHD1 deletions in prostate cancer (with SPOP and loss of PTEN) do not imply cancer predisposition in germline carriers [11].

2.8 CHD5 and the remaining genes in the class

CHD5 (1p36.31, within the critical region of 1p36 monosomy) is a tumour suppressor gene frequently deleted in neuroblastoma and a neuronal NuRD component that binds H3K27me3 and unmethylated H3K4 [13, 28, 51]. The cohort of 16 patients had 11 missense variants, one frameshift variant, c.612dup p.(Ser205LeufsTer88), two nonsense variants (c.940G>T p.(Glu314Ter) and c.1786C>T p.(Arg596Ter)), and two splice variants (c.4079-3C>G and c.4171+1G>C, the latter with in-frame exclusion of exon 27 in blood, p.(Glu1360_Ser1391delinsGly)), on NM_015557.3 [28]. The missense variants were c.577C>T p.(Arg193Trp), c.578G>A p.(Arg193Gln), c.1279G>A p.(Glu427Lys), c.2735C>T p.(Ser912Phe), c.3250G>A p.(Asp1084Asn), c.3371C>T p.(Pro1124Leu), c.3407G>A p.(Arg1136His), c.3419A>T p.(Asn1140Ile), c.4257C>G p.(Ile1419Met), c.4463A>T p.(Asp1488Val), and c.5141A>G p.(Glu1714Gly) [28]. Twelve were de novo (ten missense and two splice variants), two of them mosaic (p.(Asn1140Ile) and p.(Ile1419Met), with fewer than 25% of reads in blood), three were inherited (p.(Arg193Trp), p.(Glu314Ter), and p.(Arg596Ter)), and one was a de novo variant of uncertain significance (c.815C>T p.(Ala272Val)) [28]. The affected residues are paralogous to those in CHD3 (CHD5 Asp1084 with CHD3 Asp1120; Arg1136 with Arg1172) [28]. The gene has a haploinsufficiency probability of 0.93, a pLI of 1, a loss-of-function observed/expected upper bound fraction (LOEUF) of 0.16, and a missense Z score of 5.32; 25 of 141,456 gnomAD individuals have loss-of-function variants, some with allelic imbalance suggestive of somatic status [28]. Missense variants, mainly in the helicase domain, tended to be associated with severe early-onset epilepsy (three of four with West syndrome), in a small sample [28]. ClinGen has no validity curation and assigns a haploinsufficiency score of 1 [18, 51].

In the remaining genes of the family, the predominant mechanism is loss of function: haploinsufficiency through truncating variants and deletions in CHD2 (ClinGen score 3), with a brain-restricted phenotype [31, 52]; loss or reduction of function in CHD7, which binds more than 10,000 sites in the genome, with missense variants generally less severe [33]; and haploinsufficiency in CHD8, a gene highly intolerant of loss of function and a negative regulator of the Wnt/β-catenin pathway, with increased head circumference as the most replicated finding in humans and models [35]. The contrast between the predominance of loss of function in subfamily I and III genes and the predominance of missense variants clustered in the ATPase core of CHD3 and CHD4 had already been noted in 2018 [23].

2.9 ClinGen gene–disease validity classification

None of these genes has a ClinGen actionability assessment, paediatric or adult [8, 52, 53].

3. Clinical Spectrum and Manifestations by System

3.1 Shared syndromic description of the class

The subfamily II entities and CHD1 share a picture of global developmental delay with disproportionate impairment of speech and expressive language, intellectual disability (ID) of variable severity that may be absent, infantile hypotonia, mild craniofacial dysmorphism with a prominent forehead, widely spaced eyes, and a pointed chin, and abnormal head circumference (HC) [13, 21, 23, 28]. What distinguishes the entities are the organ anomalies: heart disease, hearing loss, hypogonadism, and vertebral anomalies in CHD4; visual abnormalities, joint laxity, and hypersociability in CHD3; autism and growth deficiency in CHD1; and epilepsy and craniosynostosis in CHD5 [4, 14, 22, 28]. All available data come from case series recruited through data-sharing platforms and diagnostic laboratories, with ascertainment bias towards more severe forms, mostly paediatric ages, and denominators that vary by feature. There are no prospective natural history studies for any of the entities, and the frequencies describe patients without specific treatment, because no disease-modifying therapy exists.

3.2 Snijders Blok-Campeau syndrome: cohorts and frequencies

The founding cohort (35 probands with de novo variants, mostly in the first and second decades, including a 59-year-old man, a pair of twins, and a pair of siblings) defined the triad of ID or developmental delay, macrocephaly, and speech and language disorder [1, 21]. The following table reproduces the frequencies from that study, with denominators varying according to the information available.

Source: [21].

The second cohort (24 patients; mean age 9.3 years, median 8, range 1–31 years) confirmed the picture and added systematically assessed facial features [22]. The comparison with the initial cohort and the combined figures are shown in the following table.

Source: [22]. In the original table, the initial-cohort atrial and ventricular septal defect values contain a typographical error, and initial-cohort seizures appear as 4/34, whereas the 2018 text reports three cases (one with epilepsy and two with neonatal seizures) [21, 22].

Comparing probands with de novo variants, probands with inherited variants, and heterozygous parents reveals the gradient of expressivity [12].

Source: [12].

The 2023 review pooled 83 patients (63 published and 20 new) and calculated frequencies using the total number of patients as the denominator, rather than those assessed for each feature, which underestimates features not investigated in all series; the original table also has inconsistent sums (for example, hernia with 5 + 6 = 6) [19]. With that caveat, the overall frequencies were: global developmental delay 81 (98%), speech disorder or delay 76 (92%), ID 61 (73%), hypotonia 59 (71%), hypertelorism 52 (63%), visual abnormality 46 (55%), macrocephaly 46 (55%), prominent forehead 43 (52%), central nervous system anomaly 35 (42%), frontal bossing 35 (42%), pointed chin 27 (33%), broad nasal bridge 26 (31%), thin upper lip 26 (31%), joint laxity 25 (30%), autism spectrum disorder (ASD) 23 (28%), strabismus 23 (28%), full cheeks 22 (27%), laterally sparse eyebrows 19 (23%), motor delay 19 (23%), low-set ears 17 (20%), deep-set eyes 16 (19%), delayed walking 16 (19%), narrow palpebral fissures 15 (18%), telecanthus 13 (16%), prominent nose 12 (14%), midface hypoplasia 12 (14%), male genital anomaly 12 (14%), congenital heart disease 11 (13%), seizures 11 (13%), feeding difficulties in infancy 11 (13%), posteriorly rotated ears 10 (12%), broad forehead 10 (12%), epicanthus 10 (12%), missing teeth 8 (10%), attention-deficit/hyperactivity disorder (ADHD) 7 (8%), hernia 6 (7%), broad nasal tip 6 (7%), dental anomalies 6 (7%), stereotyped behaviour 5 (6%), high palate 5 (6%), foot deformities 4 (5%), ventriculomegaly 4 (5%), depressed nasal root 4 (5%), rosy cheeks 3 (4%), hearing loss 3 (4%), dysarthria 3 (4%), microcephaly 3 (4%), unsteady gait 2 (2%), café-au-lait macules 2 (2%), and happy facial expression 2 (2%); autism with high cognitive abilities, renal anomaly, synophrys, short philtrum, narrow mouth, sagittal craniosynostosis, flat foot, atopic dermatitis, and thin eyebrows were each described once [19]. Frequencies from the three earlier cohorts, using the number assessed as denominator, are preferable for counselling [12, 19, 22].

3.3 Snijders Blok-Campeau syndrome: manifestations by domain

Development, speech, and language. All 33 patients in the founding cohort aged at least 2 years had delayed speech and language milestones; the mean age at first words was 2 years and 10 months (range 1.5–5.5 years, excluding six patients who were nonverbal at last assessment) [21]. Expressive language is more affected than receptive language, intelligibility is often reduced, and dysarthria, apraxia of speech, oromotor problems, and stuttering coexist [21]. Walking is acquired late, usually by age 5, and the gait is often broad-based or unsteady [1, 21].

Cognition. Full-scale IQ ranged from 70–85 (borderline intellectual functioning) to below 35 (severe ID) [21]. In the second cohort, ID was present in 20 of 21 assessable patients, with an even distribution of severity (four severe, three moderate to severe, six moderate, four mild to moderate, and four mild) [22]. In the 2023 series, almost half of patients aged 5 years or older did not have ID, although they had neurodevelopmental problems [19]. Among probands with inherited variants there was no severe ID (0/14), whereas among de novo probands severe ID reached 23% [12]. The appropriate formulation is ID of highly variable severity, from borderline to severe, with some patients having no ID; Orphanet's "moderate to severe" designation does not match the primary cohorts [2, 12].

Behaviour and adaptive functioning. Many patients are described as cheerful, sociable, and pleasant [1]. Friendly or sociable behaviour was recorded in 11 of 28 patients (39%) in the founding cohort, and Enomoto and colleagues propose hypersociability as a suggestive feature, based on the documented interaction between CHD3 and GTF2I, the principal gene of Williams syndrome [41]. The neurobehavioural profile of 38 patients (25 male; mean age 9.17 years, range 2 months to 47.5 years; 16 preschoolers, 18 school-age children, and 4 adults) was compared with cohorts with fragile X syndrome (n = 46) and level 1 ASD (n = 49) [39]. In this cohort, the recorded comorbidities were ASD in 5, ID in 13, epilepsy in 4, ADHD in 5, and cerebral palsy in 3, and medications included antipsychotics in 3, antidepressants in 3, anxiolytics in 1, psychostimulants in 4, and anticonvulsants in 4 [39]. On the Vineland-3, 67% (16/24) had profound global impairment of adaptive functioning and 33% (8/24) were at the borderline level; daily living skills were at a low adaptive level in 71% (17/24); on the ABAS-II, the general adaptive composite was extremely low in 95% (18/19) of patients aged 5 to 21 years; and on the ASQ-3, all 14 preschoolers were below the cutoff in at least one domain [39]. Communication was at a low adaptive level in 71% (17/24) and at a moderate level in 29%, with no patient at a normal level, and clinical delay in 64% (9/14) of preschoolers [39]. Thirteen of 14 (93%) had gross motor delay and 9 of 14 (64%) fine motor delay; the DCDQ suggested developmental coordination disorder in all 17 assessed, and 72% (13/18) were at a low adaptive level in the motor domain [39]. Socialisation was the relative strength: low in 58% (14/24), moderate in 33%, and adequate in 8%, and higher than the overall adaptive composite in 79% (19/24) [39]. On the SRS-2 (n = 31), 26% were in the normal range, 23% mild, 32% moderate, and 19% severe, with 4 of 6 preschoolers in the normal range [39]. On the Conners-3 (n = 18), the global index was in the clinical range in 56% (10/18), DSM-5 inattention in 50% (9/18), and hyperactivity-impulsivity in 61% (11/18); the academic area was of concern in 69% (11/16), and conduct and oppositional disorders were mostly within normal limits [39]. On the CBCL or ASR, the total score was within normal limits in 65% (22/34) and in the clinical range in 26% (9/34), and the ABC-C did not distinguish the syndrome from ASD or fragile X [39]. Overall adaptive functioning and communication were comparable to fragile X and lower than in level 1 ASD, with less inappropriate speech than in fragile X [39].

Sensory profile and pain. On the sensory profile, 46% (12/26) had extremely low registration of stimuli, 62% (16/26) had typical sensory avoiding, and 64% (14/22) of those under 14 years had altered body position processing; hypotonia and impaired proprioception were described as core features [39]. In an anonymous survey of 15 caregivers (July to September 2024, with the CHD3 Foundation; median age of children 5 years, range 1 to 15; median age at diagnosis 3 years; 73% boys; 80% white), 8 (53%) reported early painful experiences (neonatal intensive care unit, infections, needles, injuries), with reduced pain behaviours in 4 (27%) and normal pain behaviours in 3 (20%) [40]. Seven (47%) had undergone surgery (testis in 2, eye in 1, ear in 2), and 4 of those 7 showed no postoperative pain behaviours [40]. Even on a good day, 3 of 13 (23%) had mild pain, and 3 of 13 (23%) had recurrent pain (constipation with nighttime pain, ankle pain from hypermobility and repeated sprains, pain on toothbrushing, earache with colds) [40]. On the sensory questionnaire (n = 13), 3 (23%) were insensitive to deep pressure and hard impact, and several were hypersensitive to innocuous stimuli (drafts 4/13, rough surfaces 4/13, cold 5/13), with an aversive reaction [40]. The sample is small, the response rate was 20%, and reporting was by proxy [40].

Growth and head circumference. Macrocephaly affects about half of patients (51% in the combined cohort), with microcephaly in 5% [22]. In the second cohort, the mean HC Z-score was +0.9 at birth (standard deviation 1.5) and +1.3 at the last measurement (standard deviation 2.7), with 50% of patients within ±2 standard deviations [22]. The patients with variants in the helicase C-terminal domain described by Enomoto had macrocephaly and overgrowth, whereas those with phenotype 2 had microcephaly, prenatal growth restriction, and poor weight gain [41]. Two patients in the 2026 cohort with the synonymous variant c.5754G>A had short stature (one with malabsorption and the other with microcephaly and very short stature) [24].

Central nervous system. Enlarged cerebrospinal fluid spaces were described in 33% of magnetic resonance imaging (MRI) reports in the founding cohort, and the ventricles may be enlarged even in the absence of macrocephaly [1, 21]. In the second cohort, 9 of 23 patients with imaging (39%) had structural anomalies: enlarged extra-axial spaces (4/9), delayed myelination (3/9) and, in one patient each, low white matter volume, cerebral dysgenesis, periventricular polycystic structures on ultrasound, and a short corpus callosum; extra-axial spaces and ventriculomegaly did not correlate with macrocephaly [22]. Seizures occurred in 16% of the combined cohort (21% in the second cohort: a single absence seizure, three febrile seizures in one patient, and three without further detail), with no correlation with variant type or location [22]. In the 2026 cohort, one patient with p.(Arg1172Gln) had epilepsy treated with sodium valproate, episodes of aggression, and brachycephaly [24]. One patient in the founding cohort developed parkinsonism later in life, an isolated finding [21].

Vision. Visual abnormality is among the most frequent features (72% in the combined cohort): strabismus, cortical visual impairment, astigmatism, myopia, hypermetropia and, in isolated cases, dissociated vertical deviation, exotropia, amblyopia, and esotropia [22]. Among probands with inherited variants, the frequency was 25% [12].

Face and evolution with age. The gestalt includes widely spaced eyes (hypertelorism or telecanthus), a broad forehead with frontal bossing, periorbital fullness, narrow palpebral fissures, deep-set eyes, epicanthal folds, a prominent supraorbital ridge, midface hypoplasia, laterally sparse eyebrows, low-set ears, sometimes posteriorly rotated and simple with thick helices, a broad nasal base, a prominent nose, a bifid nasal tip, and a pointed chin [21]. The square face and full cheeks are more evident at younger ages and become less marked, whereas frontal bossing, the prominent and broad nose and, above all, the pointed chin persist and become defining [22]. Patients and parents with inherited variants show the same features, with a prominent nose mainly in adults [12]. Other descriptors include a thin upper lip, downturned corners of the mouth, a high palate, and dental anomalies [1, 24].

Musculoskeletal system, skin, and teeth. Hypotonia is present in 75–92% and joint laxity in 33–40%, with dislocations in some patients [21, 22]. Scoliosis occurs in some, and mild foot deformities have been described [1, 19]. Absence of permanent teeth was found in 5 patients in the second cohort (21% of the total and 33% of those examined: lateral incisors in 3, upper molars in 1, unspecified tooth in 1) [22]. Café-au-lait macules, nuchal folds, dry skin, thick scars, nevus flammeus, deep plantar creases, dysplastic toenails, and an occipital area of hypopigmented hair have been described in one patient each [22]. In the Chinese series, synophrys, transverse palmar creases, a tracheal bronchus, and hypomelanosis of Ito were observed [50].

Heart, kidney, genitalia, and other. Congenital heart disease occurs in 14% of the combined cohort (atrial septal defect, ventricular septal defect [VSD], and patent ductus arteriosus [PDA]) [22]. One patient in the 2026 cohort had coarctation of the aorta, a bicuspid aortic valve, and left ventricular asymmetry at birth [24]. Cryptorchidism affected 5 of 21 males in the founding cohort and 2 of 10 (unilateral) in the second [21, 22]. One patient with nephrolithiasis, one woman with polycystic ovary syndrome, hernias in 4–18%, and hearing loss in 13% have been described [21, 22]. Renal anomalies are rare [24]. Neonatal feeding difficulties occurred in 31% [21].

Two phenotypes. In a series of five Japanese patients from four families, phenotype 1 (helicase C-terminal domain variants; patients with IQs of 55 and 46) combined macrocephaly, hypertelorism, epicanthus, a broad-based nose, full cheeks, a thin upper lip, overgrowth, mild to moderate ID, and excessive friendliness; phenotype 2 (SNF2 domain or CHDCT2 variants; IQ below 20, 21–35, and 35) combined microcephaly, prominent ears, full lips, retrognathia, widely spaced teeth, prenatal growth restriction, poor weight gain, severe ID, and autistic traits, with epilepsy at age 5 in one patient [41]. The variants were p.(Gly961Arg) c.2881G>C, p.(His1161Leu) c.3482A>T, p.(Arg1169Trp) c.3505C>T, and p.(Glu1787Gly) c.5360A>G (published on NM_001005273.2; identical numbering on MANE NM_001005273.3) [41]. The hypothesis rests on five patients.

Overall severity and atypical cases. In the European cohort of 38 patients from 10 countries, a severity score (length and HC at birth, neonatal complications, stature and HC at last assessment, early delay, ID, speech, seizures, dysmorphism) classified 13 patients as severe, 10 as moderate, 13 as mild, and 2 as unclassifiable [24]. Four of the 28 patients with pathogenic or likely pathogenic variants were atypical: craniosynostosis with a severe phenotype in the daughter of a consanguineous couple; short stature and malabsorption with a synonymous splice variant; microcephaly and very short stature with the same variant; and coarctation of the aorta with ID and limited dysmorphism [24].

3.4 Snijders Blok-Campeau syndrome in adulthood

Adult data are scarce and cross-sectional: a 59-year-old man in the founding cohort, a 31-year-old patient in the second cohort, and four adults (up to 47.5 years) in the neurobehavioural study [1, 22, 39]. In heterozygous parents, who represent the mild end of the spectrum, mild or borderline ID occurred in 21%, macrocephaly in 50%, and facial features in 94% (16/17) [12]. There are no longitudinal studies of independence, employment, or physical health in middle age [39].

3.5 Sifrim-Hitz-Weiss syndrome

GeneReviews is based on 33 individuals with missense variants or in-frame indels (Sifrim 2016, Weiss 2016, Weiss 2020, and unpublished data) [4]. The frequencies are as follows.

Source: [4].

The 2020 cohort (32 individuals, 12 female and 20 male; age at last assessment from newborn to 30 years, mean and median 10 years) had de novo variants in all but one, with suspected inheritance from an affected, untested mother [13]. Gestation and birth weight and HC are usually normal [4, 13]. Speech delay affected 93% (29/31), with no cases of absent speech; most communicate in short sentences [4, 13]. ID was mild to moderate in 86% (19/22) of those assessed at school age, and 3 patients (13%) had a normal IQ, one of them with autism; GeneReviews reports 4 with low-normal or borderline IQ [4, 13]. Hypotonia affected 74% (17/23) and gross motor delay 83% (26/31), and independent walking was achieved at a mean of 30 months, with 3 children walking only after age 5 [13]. Some children have feeding difficulties, a few requiring a nasogastric tube or gastrostomy, and hydrocephalus requiring shunting is associated with mild to moderate ID [4]. Behavioural problems are uncommon: ASD in 3 individuals and attention deficit, impulsivity, and anxiety in another 3 [4].

HC is usually normal at birth; macrocephaly (above the 95th percentile) develops in 39% (11/28) and microcephaly in 2 [13]. Short stature affects about 30%, obesity appears in late childhood or adulthood in 7 of 30, and growth hormone (GH) deficiency was documented in 4 patients in GeneReviews (all treated, with a good overall response) and in 3 in the 2020 cohort (1 treated) [4, 13]. Brain imaging was abnormal in 69% of the total and in 96% of those who underwent MRI (ventriculomegaly, hydrocephalus requiring shunting, Chiari 1 malformation, thin corpus callosum); GeneReviews reports anomalies in 92% (22/24), with mild to moderate ventriculomegaly in 41%, Chiari 1 malformation in 29%, and hydrocephalus requiring shunting in 18%, as well as syringomyelia [4, 13]. Three children had a stroke in the postnatal period or in childhood, two of them with moyamoya disease (one congenital stroke and one in childhood) [4, 13].

Congenital heart disease affects 65% (19/29) in the 2020 cohort and 72% in GeneReviews: in order of frequency, atrial septal defect, VSD, pulmonary stenosis, PDA, tetralogy of Fallot, mitral valve anomalies, Ebstein anomaly, and truncus arteriosus [4, 13]. Hearing loss, mainly sensorineural and sometimes conductive or mixed, affects 55–58%, and recurrent otitis media is not common [4, 13]. Ophthalmological anomalies (73–74%) are, in order of frequency, strabismus, astigmatism, hypermetropia, glaucoma, small optic nerves, iris coloboma, and myopia, without significant visual impairment [4]. Skeletal anomalies include syndactyly of the hands and feet, polydactyly, acquired scoliosis, vertebral anomalies (mainly cervical fusions, which increase the risk of cervical instability), and tarsal and carpal coalitions; the true frequency of fusions may be higher, because not all patients underwent skeletal imaging [4]. Cryptorchidism or micropenis affected 65% (13/20) of males, with low gonadotropins documented in 5; hypogonadism has not been described in females, with normal hormone levels in the two postpubertal females studied [13]. Anteriorly placed anus and vesicoureteral reflux occurred in 3 patients [13]. There were two deaths: one after birth from complications of heart disease and one at age 21 from complications of cervical instability and long-term tracheostomy [13].

The face includes a high, broad forehead, a square face, periorbital fullness, widely spaced eyes, a short nose, and small or dysmorphic ears, more evident in childhood; computerised facial analysis distinguished 16 patients with ATPase/helicase variants from 32 controls (area under the curve 0.917; p = 0.009), but the dysmorphic features are mild and diagnosis is made by panel or exome [13]. The 2016 descriptions added coarse facies, epicanthus, ptosis, omphalocele, ambiguous genitalia, broad and short clavicles, flat acetabular roof, short femoral neck, fused carpal bones, postaxial polydactyly, wormian bones, Chiari 2 malformation, and chronic kidney disease in one patient [3]. The OMIM clinical synopsis (2025) also includes trigonocephaly, skull base anomalies, tapering fingers, intestinal malrotation, primary amenorrhoea, and hypothyroidism [3].

In the 2025 clinical cohort (27 patients), the frequencies were ID 19/23, speech delay 26/27, motor delay 27/27, macrocephaly 11/24, heart disease 12/24, abnormal brain imaging 14/23, hypogonadism 11/24, and hearing loss 9/23, with normal intelligence in 4; optic nerve anomalies, aortic root dilatation, and intestinal malrotation emerged as recurrent features, and autism was more frequent among carriers of truncating variants [25]. Moyamoya was estimated to be a low-frequency feature (3–5%) [25]. In the series of 39 trios with moyamoya angiopathy (USA, France, and Germany; mean onset at age 15), one patient with bilateral moyamoya at age 10, developmental delay, congenital heart disease, abnormal gait, and mild dysmorphism had the de novo variant p.(Ala1178Val); in 158 additional probands with moyamoya, six more rare CHD4 variants were found (five missense and one frameshifting splice variant subject to NMD), three of them in the C-terminal region that interacts with pericentrin, all with bilateral moyamoya and ID in only one additional case [46]. Moyamoya may present in adulthood without a history of developmental disorder [46]. Cancer has not been described, and life expectancy is unknown, with one patient alive at age 30 [4].

3.6 Pilarowski-Bjornsson syndrome

In the original description, the five subjects with a described phenotype, all female, had developmental delay and hypotonia, including a 10-year-old girl with regression; four had speech apraxia (between ages 4 and 10), three had autism with stereotypies, and five had seizures or an abnormal electroencephalogram (EEG); ID persisted in two, was absent in a 5-year-old, and was unknown in two, and two had postnatal growth delay [6, 23]. The series included two sisters conceived by in vitro fertilisation with oocytes from the same donor, who was not tested [23]. The dysmorphic features included a pointed chin, frontal bossing, arched eyebrows, almond-shaped eyes, downslanting palpebral fissures, periorbital fullness, and a depressed midface; the OMIM synopsis also lists macrocephaly, translucent skin, mild immunodeficiency, and allergic manifestations [6, 23]. For context, speech apraxia affects 1–2 per 1000 children in the general population [23].

In the 2026 cohort, among 40 individuals with 36 variants with substantial evidence, carriers of putatively disruptive missense variants (n = 21) had developmental delay in 86%, speech apraxia or delay in 81%, ASD in 67%, and ID in 68% of those assessable; carriers of truncating variants (n = 20) had speech apraxia or delay in 95%, ASD in 70%, developmental delay in 60%, and ID in 47% [14]. Growth deficiency is a common feature, and the phenotypic score tended to be more severe in females (p = 0.0533; Cohen's d 0.980), with the limitation of heterogeneous annotation and more recorded entries for females [14]. A Saudi girl, the daughter of third cousins, with the CHD1 nonsense variant c.966G>A p.(Trp322Ter) and an ASH1L missense variant, showed severe neonatal hypotonia, global delay with dissociation between domains, absent expressive language, autistic traits, and a febrile seizure with a normal EEG [54].

3.7 Parenti-Mignot neurodevelopmental syndrome

In the cohort of 16 patients (7 female and 9 male; aged 3 to 24 years, median 9 years and 6 months), the frequencies were speech delay 81% (13/16), behavioural abnormalities 69% (11/16; ASD and obsessive-compulsive tendencies, self-injury, temper tantrums, aggression), epilepsy 62% (10/16), subtle dysmorphism 11/16, motor delay 56% (9/16), ID 64% (9/14; moderate in 2 and severe in 4 of the 6 graded), hypotonia 7/14, and craniosynostosis 3/7 (sagittal, metopic, trigonocephaly; 2 operated) [28]. Four patients had a normal IQ (85–105) and one an above-average IQ [28]. Median ages of attainment were 13 months for sitting, 28 months for walking, and 24 months for first words, and three patients were nonverbal at ages 3, 9, and 24 years [28]. Epilepsy began at a median age of 10 months (from the first day of life to 16 years): three patients had 1–5 untreated seizures, three had medicated generalised epilepsy, and four had developmental and epileptic encephalopathy with burst-suppression or hypsarrhythmia (West syndrome in three) [28]. MRI was normal in 8 of 12, and there is no distinctive facial gestalt, nor one suggestive of 1p36 deletion [28]. A 54-year-old man with a nonsense variant had mild to moderate ID, with adaptive functioning, attention, executive function, and memory consistent with his intellectual level, a specific weakness in social cognition, internalising and externalising problems, impaired emotional regulation, and aggression [55].

3.8 Related entities

CHD2-related disorder is a developmental and epileptic encephalopathy with seizure onset between 6 months and 12 years (median 30 months), atonic-myoclonic-absence seizures, clinical photosensitivity in 80% (20/25), and drug resistance (only 13 of 33 seizure-free on medication for 2–5 years); among the 139 published individuals, seizures occurred in 96%, delay or ID in 95%, and ASD in 56% [31]. CHD2 features, together with MEF2C, SYNGAP1, and UBE3A, among the genes of childhood epilepsies with onset after the first year and among the causes of epilepsy with myoclonic-atonic seizures, whose peak onset lies between ages 2 and 4 [56]. CHD7-related disorder encompasses CHARGE syndrome and its subsets, including isolated hypogonadotropic hypogonadism with or without anosmia [33]. CHD8-related disorder (115 individuals in GeneReviews, sex ratio about 2:1) combines macrocephaly (80%), tall stature at puberty (80%), ASD and ID (75–80%), sleep disturbance (67%), gastrointestinal complaints, mainly constipation (63%), ADHD (50%), and seizures in 12% [35].

4. Clinical Recognition and Red Flags

4.1 When to suspect

None of the entities has formal clinical criteria, and the dysmorphism is too mild for a clinical diagnosis; suspicion serves to guide the interpretation of genomic results and to weigh a variant of uncertain significance [4, 13, 24]. Snijders Blok-Campeau syndrome should be considered in a child with global developmental delay and disproportionate impairment of expressive speech (apraxia, dysarthria, stuttering, or absent speech), especially if accompanied by macrocephaly or HC at the upper limit of normal, a prominent forehead or frontal bossing, widely spaced and deep-set eyes, a pointed chin, hypotonia, joint laxity, and visual abnormalities (strabismus, hypermetropia, cerebral visual impairment) [21, 22]. A friendly or hypersociable personality and a reduced response to pain are additional clues [40, 41]. Sifrim-Hitz-Weiss syndrome should be considered in the setting of developmental delay or mild to moderate ID with any of the following: generalised hypotonia in infancy; absolute or relative macrocephaly; septal, conotruncal, or valvular congenital heart disease; skeletal or limb anomalies (vertebral fusion, carpal or tarsal coalition, syndactyly, polydactyly); male hypogonadism (cryptorchidism, micropenis); ophthalmological anomalies (strabismus, hypermetropia, astigmatism); hearing impairment; or moyamoya with congenital or childhood stroke [4]. In a child with partial overlap with CHARGE syndrome but without coloboma or choanal atresia, CHD4 is a diagnosis to consider [3]. Pilarowski-Bjornsson syndrome should be considered in the setting of speech apraxia or absent expressive language with autism, hypotonia, and growth deficiency [14, 23]. Parenti-Mignot syndrome should be considered in the setting of language deficit, behavioural abnormalities, epilepsy (including infantile spasms), and craniosynostosis [28].

4.2 Signs by age

4.3 Red flags and emergency signs

Signs requiring urgent evaluation stem mainly from the structural complications of CHD4: rapid HC growth, a bulging fontanelle, vomiting, or altered level of consciousness (hydrocephalus, which required shunting in 18%); occipital headache with Valsalva manoeuvre, dysphagia, central apnoea, or pyramidal signs (Chiari 1 malformation, syringomyelia, spinal cord compression from cervical instability); and acute focal neurological deficit, hemiparesis, or new-onset seizure (moyamoya-related stroke) [4, 46]. Neck trauma in a patient with vertebral fusions warrants immediate evaluation [4]. In Snijders Blok-Campeau syndrome, absence of pain expression does not equal absence of pain perception: a change in behaviour, gait, or sleep may be the only manifestation of a fracture, sprain, constipation, or otitis, and injuries may go unnoticed [40]. In Parenti-Mignot syndrome, onset of infantile spasms is a neurological emergency [28]. There is no newborn screening for any of the entities.

5. Diagnosis: Criteria, Testing, and Algorithms

5.1 Diagnostic principle

The diagnosis of all entities in the class is molecular: it is established by identifying a heterozygous pathogenic or likely pathogenic variant in the corresponding gene in a patient with a compatible phenotype [4, 8]. A variant of uncertain significance (VUS) neither establishes nor excludes the diagnosis [4]. There is no biochemical, imaging, or electrophysiological marker; ATPase activity and remodelling assays and structural modelling are research tools [8]. Nonmolecular clinical assessment serves for phenotyping: developmental and cognitive assessment, formal speech and language assessment including oromotor function, serial HC, and examination for hypotonia and laxity [8].

5.2 Tests in order and what each detects

The usual pathway in a child with developmental delay or ID begins with chromosomal microarray analysis (CMA), followed by a broad ID panel, exome, or genome; single-gene testing is rarely useful because the phenotype is nonspecific, and some ID panels do not include CHD4 [4]. In the 2020 CHD4 cohort, all participants had a nondiagnostic microarray before sequencing, and the diagnosis was made mainly by clinical exome [13]. The following table summarises the proportion of variants or probands detected by each method according to the sources; these figures are proportions of the known variant spectrum, not clinical sensitivity.

5.3 Test limitations

A negative exome result does not exclude the diagnosis. Exome sequencing may not cover deep intronic regions, noncanonical splice variants may not be recognised by prediction programmes (such as the synonymous CHD3 variant c.5754G>A, whose pathogenicity was confirmed by RNA studies), and copy-number variants (CNVs) involving one or a few exons escape some pipelines [24]. Somatic mosaicism may reduce the allele fraction below the detection threshold in blood: in CHD5, two de novo variants were present in fewer than 25% of reads, and in CHD3, maternal mosaicism and parental gonosomal mosaicism of 3% have been documented [21, 28, 41]. Standard trio filtering strategies, which prioritise de novo or biallelic variants, are not optimised for variants inherited from clinically unaffected parents in dominant genes; in CHD3, a rare inherited variant with phenotypic concordance should be considered possibly pathogenic depending on its characteristics, and inheritance from a healthy parent is not in itself a criterion for benignity [8, 12]. In CHD1, 83% of missense variants in the 2026 cohort were VUS and one recurrent variant was present in unaffected parents, which calls for caution in classification [14]. Minigene splicing assays help classify intronic CHD3 variants [50]. Whole-gene deletion does not establish a diagnosis of Sifrim-Hitz-Weiss or Pilarowski-Bjornsson syndrome [10, 11].

5.4 Report content and nomenclature

The report should state the gene, the versioned transcript (MANE Select: CHD3 NM_001005273.3, CHD4 NM_001273.5, CHD1 NM_001270.4, CHD5 NM_015557.3, CHD2 NM_001271.4, CHD7 NM_017780.4, CHD8 NM_001170629.2), the Human Genome Variation Society (HGVS) nomenclature at the coding DNA and protein levels, the genomic coordinates with the reference genome build, zygosity, the ACMG/AMP class, and the criteria applied [9]. In CHD3, numbering on transcript NM_001005271.3 exceeds that of MANE by 59 amino acids, and comparison with the literature requires conversion with a validation tool [9, 19]. Designating a variant as de novo requires confirmation of parentage [31].

5.5 Variants of uncertain significance, segregation, and episignatures

Family segregation is of limited value in CHD3, because heterozygous parents may have only subtle features; before interpreting an inherited variant as benign, the parents should be examined for macrocephaly and facial features (prominent forehead, pointed chin) [12]. DNA methylation episignatures in leucocytes, obtained with methylation arrays and classifiers, are the main functional tool available for reclassifying VUS [24, 25].

In validation of the CHD3 episignature, three of ten VUS were reclassified as likely pathogenic with at least good confidence (NM_001005273.3:c.5198C>T p.(Ser1733Leu), c.1840G>A, and c.1072_1074del, all with a mild phenotype), and the remaining seven, including c.1903C>T, had intermediate profiles that preclude definitive classification; three VUS were inherited from probably asymptomatic parents [24]. The episignatures of CHD3, CHD8 (intellectual developmental disorder with autism and macrocephaly), and CHD7 share no strict DMPs; principal component analysis separates CHARGE syndrome from the other two and Snijders Blok-Campeau syndrome from CHD8, but some CHARGE or CHD8 patients reach up to 40% probability in the CHD3 classifier [24]. The CHARGE episignature splits into a strong-signal group (more severe patients, six of eight sampled before 1 month of age) and a weak-signal group, and an age effect cannot be excluded [24]. The CHD4 episignature partly overlaps those of class I truncating ADNP (21% of DMPs; ADNP forms the ChAHP complex with CHD4 and HP1, and the facial features are similar), the BAFopathies (27%), CHD8 (22%), and ATRX (20%) [25]. In routine diagnostic testing of 400 patients with neurodevelopmental disorders studied by exome and methylation, 67 (17%) had variants in chromatinopathy genes and 26 of 60 (43%) had a concordant episignature; CHD3 and CHD5 had no episignature on the commercial panel used, and one patient with clinically typical Snijders Blok-Campeau syndrome and the likely pathogenic missense variant c.3692G>A showed partial overlap with the KAT6A signature without a variant in that gene; one patient with a pathogenic truncating CHD4 variant was negative for the missense-defined episignature, and one patient with a CHD8 splice variant was positive for its signature [48]. Methylation complements, but does not replace, genomic analysis, and partial overlap with another signature should not be read as an alternative diagnosis in the absence of a variant [24, 48].

5.6 Algorithm

5.7 Prenatal and preimplantation diagnosis

Once the variant has been identified in a family member, prenatal diagnosis and preimplantation genetic testing (PGT) are possible [4, 35]. Fetal ultrasound may show nonspecific findings in CHD4 (heart disease, ventriculomegaly, omphalocele) and growth restriction in CHD3 phenotype 2, but none of the entities has ultrasound markers that allow diagnosis without molecular testing [3, 4, 41]. In CHD3, variable expressivity makes it impossible to predict the severity in a fetus carrying an inherited variant [12].

6. Differential Diagnosis

6.1 Principles

The phenotype of the entities in the class is not sufficient for a clinical diagnosis, and the differential strictly encompasses all forms of ID without distinctive findings, including the OMIM phenotypic series of autosomal dominant, autosomal recessive, and X-linked, syndromic and nonsyndromic ID [4]. In practice, the differential is useful in three situations: when there is a VUS in a CHD gene, to assess phenotypic concordance; when another recognisable syndrome is clinically suspected and the genomic result is negative; and in distinguishing among members of the class itself, whose phenotypes overlap [19, 24]. The discriminating test is almost always molecular, and episignatures can separate some of the entities [24, 25].

6.2 Differential diagnosis table

6.3 Main differentials

Within the class, CHD3 and CHD4 are distinguished by organ findings: congenital heart disease is much more frequent in CHD4 (65–72% versus 14%), as are hearing loss (55–58% versus 13%), male hypogonadotropic hypogonadism, cervical vertebral fusions, and hydrocephalus requiring shunting; macrocephaly, ventriculomegaly, cryptorchidism, and a similar face are shared [4, 21, 22]. Speech apraxia is common to CHD3 and CHD1 (four of the six patients in the original CHD1 series), but autism and growth deficiency are more characteristic of CHD1 [14, 21]. Frequent and severe epilepsy points towards CHD5 or CHD2, and craniosynostosis towards CHD5, although it has been described in CHD3 [24, 28]. CHARGE syndrome is distinguished by coloboma, choanal atresia, hypoplasia or aplasia of the semicircular canals, and the characteristic ear; the overlap with CHD4 (heart disease, hearing loss, hypogonadism, ear anomalies) has been noted since the initial description [3, 33]. In CHARGE syndrome and CHD7-related disorder, anal atresia is rare, unlike in the VACTERL association [33].

Among the overgrowth syndromes and neurodevelopmental disorders with macrocephaly, the differential of CHD3 includes Sotos, Malan, Tenorio, and Simpson-Golabi-Behmel syndromes, with which it shares the pointed chin and frontal bossing [19]. CHD8 combines macrocephaly with tall stature and sleep and gastrointestinal disturbances [35]. PTEN should be excluded in the presence of marked macrocephaly, because of its cancer implications [35]. The reference textbook on metabolic diseases includes the NuRD components among the genetic causes of macrocephaly, alongside the mTOR-AKT pathway syndromes, the leukodystrophies (Alexander disease, megalencephalic leukoencephalopathy with subcortical cysts), and the Rett-like syndromes, a reminder that MRI is needed in the presence of progressive macrocephaly [37]. The neurobehavioural profile of CHD3 resembles that of fragile X syndrome in adaptive functioning and communication, but socialisation is a relative strength and inappropriate speech is less marked [39].

Eponym confusion deserves attention: X-linked Snijders Blok syndrome, caused by DDX3X, is a different entity from Snijders Blok-Campeau syndrome, and the literature contains at least one report with the title mixed up [43]. OMIM numbers are also cited incorrectly in articles (gene numbers instead of phenotype numbers, and "518205" instead of 618205 in one review article), which requires confirmation in the primary databases [19]. In moyamoya with ID, heart disease, or developmental delay, CHD4 should be included in panels, alongside the classic syndromic causes and the Asian founder variant in RNF213, which is a susceptibility factor [46].

7. Treatment and Risk Management Strategies

7.1 Principles and sources of recommendations

There is no approved disease-modifying treatment for any of the entities, nor any scientific society guideline or international consensus specific to Snijders Blok-Campeau syndrome, Sifrim-Hitz-Weiss syndrome, or Pilarowski-Bjornsson syndrome: the ERN ITHACA guidelines page, accessed in October 2026, lists published guidelines for other syndromes (Beckwith-Wiedemann, fetal valproate syndrome, Cornelia de Lange, Phelan-McDermid, Rubinstein-Taybi, spinal dysraphisms, Kleefstra, Bardet-Biedl) and a general guideline on transition from paediatric to adult care (March 2026), but none for these entities [27]. The only structured source on evaluation, treatment, and surveillance for the class is the CHD4 GeneReviews chapter (2020, unrevised) [4]. The recommendations of Weiss 2020 for CHD4, of van der Spek 2022 for parental evaluation in CHD3, and of Ocay 2025 for pain assessment are expert opinion or conclusions from observational studies, not guidelines [12, 13, 40]. Management relies on multidisciplinary supportive care: early speech therapy, physical therapy, and occupational therapy, educational support, and treatment of associated medical problems [24, 40]. For Snijders Blok-Campeau syndrome and Pilarowski-Bjornsson syndrome, the measures below are derived from cohort frequencies and from general recommendations for ID and neurodevelopmental disorders, and represent extrapolation.

7.2 Evaluation after diagnosis

7.3 Speech, language, and communication

Speech and language disorder is universal in Snijders Blok-Campeau syndrome and Pilarowski-Bjornsson syndrome and almost universal in Sifrim-Hitz-Weiss syndrome, with a motor component (apraxia, dysarthria, oromotor problems) that requires formal assessment by a speech and language therapist with expertise in motor speech disorders [4, 14, 21]. Oromotor dysfunction should be assessed at each visit, with a swallowing study in the presence of choking, poor weight gain, or frequent respiratory infections [4]. Augmentative and alternative communication (AAC), including the sign language used by the index case, does not hinder speech acquisition and should be introduced early in children with very limited expression [4, 21]. Six of 33 patients in the founding cohort were nonverbal at last assessment [21].

7.4 Development, education, and adaptive functioning

The CHD4 GeneReviews chapter describes the US model: early intervention from 0 to 3 years (occupational therapy, physical therapy, speech and feeding therapy, infant mental health), developmental preschool with an individualised education plan from 3 to 5 years, follow-up by a developmental paediatrician at all ages, and transition planning in adolescence [4]. In Portugal and Brazil, the equivalents are early childhood intervention and the measures supporting learning and educational inclusion, which are not described in the sources consulted. Physical therapy targets mobility and the prevention of contractures, scoliosis, and hip dislocation, with assistive devices when needed; occupational therapy addresses fine motor skills and activities of daily living [4]. In Snijders Blok-Campeau syndrome, the developmental coordination disorder suggested in all those assessed, the proprioceptive deficit, and joint laxity justify physical and occupational therapy aimed at balance, coordination, and joint protection [39]. Formal neuropsychological assessment helps guide support, because adaptive functioning may be lower than expected from IQ, and socialisation is a relative strength that can be used in educational strategies [39, 55].

7.5 Behaviour and mental health

ADHD (inattention in 50% and hyperactivity-impulsivity in 61% of those assessed with the Conners-3), ASD, and anxiety are frequent in Snijders Blok-Campeau syndrome, and ASD affects about two thirds of patients with Pilarowski-Bjornsson syndrome [14, 39]. Management follows that of neurodevelopmental disorders: behavioural intervention (applied behaviour analysis), pharmacological treatment of ADHD guided by the developmental paediatrician, and child psychiatry for severe aggression [4]. In the neurobehavioural cohort, 4 patients were taking psychostimulants, 3 antipsychotics, 3 antidepressants, and 1 an anxiolytic, with no data on response or adverse effects [39]. There are no efficacy or safety data for any psychotropic drug in these entities; dosing of methylphenidate, risperidone, and melatonin according to the official labels for analogous indications is given in subsection 7.13. In CHD5, aggression, temper tantrums, and obsessive-compulsive tendencies are frequent and require psychiatric follow-up [28, 55].

7.6 Pain

In Snijders Blok-Campeau syndrome, a quarter of patients show insensitivity to deep pressure and impact and some do not express postoperative pain, whereas others are hypersensitive to innocuous stimuli [40]. Individualised observational pain scales, such as the Paediatric Pain Profile, the NCCPC-R (Non-communicating Children's Pain Checklist – Revised), or the INRS (Individualized Numeric Rating Scale), completed with caregivers, are recommended, together with an active search for causes of pain (constipation, otitis, sprains due to hypermobility, dental pain) when behaviour changes [40]. Absence of pain expression should not be interpreted as absence of pain perception, especially postoperatively [40].

7.7 Epilepsy

Seizures occur in 16% of patients with Snijders Blok-Campeau syndrome, have been described in some patients with CHD1 and CHD4 variants, and affect 62% of those with Parenti-Mignot syndrome, with developmental and epileptic encephalopathy in four of 16 [6, 22, 28]. No antiseizure medication has demonstrated efficacy specifically in these entities, and treatment follows the epilepsy syndrome and seizure type, under the guidance of a neurologist; in Snijders Blok-Campeau syndrome, one patient was treated with sodium valproate (Depakine®), with no published dose or response [24]. West syndrome in CHD5 is treated according to infantile spasms protocols [28]. Dosing of valproate and vigabatrin according to the official labels, with hepatic monitoring for the former and visual monitoring for the latter, is given in subsection 7.13 [57, 58]. CHD2-related epilepsy shows photosensitivity, with the recommendation to avoid intensely flashing lights [31].

7.8 Neurosurgery, spine, and moyamoya in CHD4

Hydrocephalus, Chiari 1 malformation, and syringomyelia are treated neurosurgically; 18% of patients in the GeneReviews series required ventricular shunting [4]. Cervical instability associated with vertebral fusions may require surgical fixation (C1–C2 or other) [4]. Patients should avoid activities involving rapid neck movements or risk of head and neck trauma (contact sports, high-speed amusement rides), and one patient died at age 21 from complications of cervical instability and prolonged tracheostomy [4, 13]. Moyamoya requires a high index of suspicion in the presence of focal neurological signs, headaches, or seizures; revascularisation follows the general criteria for moyamoya angiopathy, which are not addressed in the sources for the class [13, 46].

7.9 Heart, hearing, vision, and orthopaedics

Congenital heart disease is managed by paediatric cardiology and cardiac surgery according to the lesion; in CHD4, aortic root dilatation was described as a recurrent feature in 2025, which justifies including its measurement in the echocardiogram [4, 25]. Hearing loss is managed with hearing aids, otolaryngology, and support services for deaf children [4]. Refractive errors, strabismus, and cerebral visual impairment are followed by paediatric ophthalmology, with visual rehabilitation when there is cerebral visual impairment [4, 22]. Scoliosis, foot deformities, joint laxity, and dislocations are followed by orthopaedics and physical medicine and rehabilitation [1, 4].

7.10 Endocrinology: puberty and growth in CHD4

Male hypogonadotropic hypogonadism should be sought in the first year of life (FSH, LH, and testosterone in the presence of cryptorchidism or micropenis) and at the expected age of puberty, and females should be evaluated during or after puberty, because the female risk is unknown (primary amenorrhoea is listed in the OMIM synopsis) [3, 4, 13]. Cryptorchidism and micropenis are managed by paediatric endocrinology and urology [4]. In the presence of short stature, GH testing should be considered; four patients with documented deficiency were treated with GH and had a good overall response, but GeneReviews notes that there are no safety data for GH treatment in this disorder [4]. Somatropin dosing for paediatric GH deficiency, according to the official label, is given in subsection 7.13.

7.11 Sleep and gastrointestinal tract

Sleep disturbance should be screened for from 12 months of age and treated first with behavioural interventions [4]. When these are insufficient, prolonged-release melatonin has European marketing authorisation for insomnia in children and adolescents with ASD or neurogenetic disorders, which covers some of these patients; dosing is given in subsection 7.13 [59]. Constipation is a frequent unrecognised cause of nighttime pain in Snijders Blok-Campeau syndrome and is managed with fluids, fibre, and laxatives according to standard practice [35, 40]. Severe feeding difficulties may require thickened or chilled foods and, in severe cases, a nasogastric tube or gastrostomy [4]. The intestinal malrotation described in CHD4 should be considered in the presence of acute gastrointestinal symptoms [3, 25].

7.12 Drugs cited in the sources for the class

No drug has an approved indication for these entities, and the class-specific sources do not give doses. The table gathers the drugs mentioned in those sources, with their context and level of evidence; dosing for drugs used symptomatically, taken from the official labels for analogous indications, is given in subsection 7.13.

7.13 Paediatric dosing of symptomatic drugs (extrapolated from official labels)

The following doses come from the official labels for the analogous indication shown in each row (the FDA label on DailyMed and, in the comparison table that follows, the harmonised European product information from the EMA and the CMDh). None of these drugs is indicated for Snijders Blok-Campeau, Sifrim-Hitz-Weiss, Pilarowski-Bjornsson, or Parenti-Mignot syndrome: their use targets the symptom or the associated syndrome (epilepsy, infantile spasms, ADHD, irritability with aggression, insomnia, GH deficiency) and is an extrapolation, with no efficacy or safety data in these disorders. Formulations and brand names vary among the USA, the European Union, Portugal, and Brazil.

The differences between the US label and the European product information have practical consequences in these patients, summarised in the following table.

The sources consulted do not give the brand names in Brazil, and ANVISA information was not consulted.

Baclofen, tizanidine, and botulinum toxin type A are cited only in the generic ID management text of GeneReviews and are not part of the treatment described in the cohorts for these entities [4]. For constipation, the sources for the class specify the sequence of fluids, fibre, and laxatives, without specific dosing [35].

7.14 Regulatory table (USA and European Union)

There are no data on national access to specific medicinal products in Portugal or Brazil, because there is no authorised drug for these entities.

7.15 Clinical studies and preclinical evidence

The inclusion criteria for trial NCT06860672 required normal hepatic, cardiac, and immune function and normal coagulation and platelets, and excluded intracranial tumour or space-occupying lesion, contraindication to lumbar puncture, status epilepticus or unstable epilepsy, active systemic infection, an anti-AAV neutralising antibody titre above 1:50, immunomodulators in the preceding three months, and prior gene therapy [26]. The registry was last updated in July 2026 [26]. A 2026 review of developmental and epileptic encephalopathies in adults includes AAV-mediated in vivo base editing in Snijders Blok-Campeau syndrome among the molecular approaches with serious adverse events, including deaths, alongside AAV gene therapy for STXBP1 and intrathecal antisense oligonucleotides for KCNT1 [74]. Searches for "CHD4," "Sifrim-Hitz-Weiss," and "Pilarowski" on ClinicalTrials.gov in October 2026 returned no registered studies, which does not prove the absence of unregistered research [72].

The available therapeutic evidence is therefore only preclinical, and the only human study was halted by a product-related death. Thrombotic microangiopathy is a known adverse effect of high systemic or intrathecal AAV doses, and the variant targeted by the trial is the same as in the mouse model, which means that gene correction must be presented to families as early-stage research with documented serious risk, not as emerging therapy [26, 42, 74]. The approach would apply, by definition, only to a single variant (p.(Arg966Trp) on MANE), present in four patients in the 2020 cohort [9, 22]. Results in animal models do not permit indirect comparison with human studies, and there are no trials with efficacy endpoints in humans.

7.16 Developments in the past 12 months (October 2025 to October 2026)

7.17 Agents and circumstances to avoid

In CHD4, contact sports and activities involving sudden neck movements or risk of craniocervical trauma should be avoided, because of possible cervical instability and spinal cord compression [4]. In CHD2, intensely flashing lights should be avoided, as they can provoke eyelid myoclonia, absences, and generalised tonic-clonic seizures [31]. In CHD7-related disorder, it is recommended to minimise and cluster anaesthesia, because of postanaesthetic respiratory complications [33]. The sources consulted do not identify drugs that are contraindicated in Snijders Blok-Campeau, Pilarowski-Bjornsson, or Parenti-Mignot syndrome.

7.18 Pregnancy, lactation, perioperative period, and transition

There are no published data on pregnancy or lactation in women with these entities, although mothers with mild Snijders Blok-Campeau syndrome who transmitted the variant have been described, including three generations in one family [12]. In the CHD4 perioperative period, vertebral fusions and cervical instability require prior assessment of the cervical spine and care with neck mobilisation [4, 13]. In Snijders Blok-Campeau syndrome, postoperative analgesia should be guided by observational scales rather than spontaneous complaint, because four of seven operated patients showed no pain behaviours [40]. Transition to adult medicine follows the general ERN ITHACA guideline of March 2026 and, in Snijders Blok-Campeau syndrome, should include the plan for pain assessment and nonverbal communication [27, 40].

7.19 Recommendations from the sources and divergences

8. Reproductive Options and Family Counselling

8.1 Mode of inheritance and scenarios

All entities in the class have autosomal dominant inheritance, but the proportion of inherited variants and the penetrance differ among genes, and that difference determines counselling. In CHD4, almost all cases are de novo, with a presumed transmission from an untested mother with features of the disease; germline mosaicism had never been reported through 2020, and a parent with somatic mosaicism could be mildly affected [4, 13]. In the 2025 cohort, however, 7 of 26 variants were inherited, mostly truncating variants with incomplete penetrance [25]. In CHD3, most probands with a severe phenotype have de novo variants (20 of 22 with known inheritance in the second cohort; 80% in the 2023 series), but there are 21 families with transmission from mildly affected or clinically unaffected parents, with a maternal predominance (71%) [1, 12, 19]. In CHD1, 18% of missense variants and 20% of truncating variants with known segregation were inherited, and penetrance appears to be sex-dependent [14]. In CHD5, three families had unaffected carriers [28].

8.2 Testing the parents and cascade testing

Both parents should be tested for the variant, and the de novo designation requires confirmation of parentage [31]. In CHD3, before concluding that an inherited variant is benign, carrier parents should be examined for subtle features: macrocephaly occurs in 50% of heterozygous parents and facial features (prominent forehead, pointed chin, deep-set eyes) in 94% [12]. When blood testing is negative, gonadal mosaicism should be discussed, because recurrence in two siblings has been documented from a clinically unaffected parent with 3% gonosomal mosaicism [41]. In a family with transmission, cascade testing of at-risk relatives identifies carriers with mild expression and guides reproductive counselling [12].

8.3 Expressivity, penetrance, and sex

Reduced penetrance and variable expressivity in CHD3 have implications for recurrence risk, prenatal diagnosis, prognosis, and variant interpretation, as is also the case in ANKRD11-related KBG syndrome [12]. The spectrum of inherited cases ranges from moderate ID with multiple features to a single facial feature or macrocephaly, and the younger generations appear more affected, possibly through ascertainment bias [12]. The maternal predominance of transmissions and the excess of women among carriers of loss-of-function variants in gnomAD suggest a female protective effect in CHD3 [12]. In CHD1, the pattern is the reverse: an excess of males in the cohort and more severe expression in females, with an androgen-mediated male protective effect in the mouse model [14]. In CHD5 family 1, cousins with speech delay and craniosynostosis were the children of two sisters, one of them asymptomatic; in family 2, the nonsense variant was inherited from a father with dyslexia and borderline intellectual functioning; in family 3, the father had learning difficulties and severe psychiatric illness [28].

8.4 Prenatal and preimplantation diagnosis

Once the family variant is known, prenatal diagnosis and PGT are possible [4, 31]. Counselling should explain that, in CHD3, a positive result in a fetus from a family with an inherited variant does not allow the severity to be predicted, which may range from an isolated facial feature to moderate ID [12]. Ultrasound does not replace molecular testing [4]. Preconception counselling is recommended for affected or carrier adults of reproductive age [4].

8.5 Counselling of minors and special situations

Testing of asymptomatic minor siblings for exclusively reproductive purposes should be deferred until the age at which they can decide; by contrast, testing of a sibling with developmental delay or suggestive features has direct clinical benefit. Adults with mild ID who carry the variant should receive counselling adapted to their comprehension, because adults with Snijders Blok-Campeau syndrome who have had affected children have been described [12]. The 17p13.1 deletion encompassing TP53, detected by CMA in a patient with a deleted CHD3, has its own cancer implications (Li-Fraumeni syndrome) for the patient and for relatives who inherit it [22]. In consanguineous families, the possibility of biallelic CHD3 variants with a more severe phenotype and of other genetic factors should be considered [24, 49].

9. Special Considerations: Population Diversity and Epidemiology

9.1 Frequency

There is no validated population estimate of prevalence or incidence for any of the entities. Orphanet assigns them the prevalence class of less than 1/1,000,000, which is an assigned class and not a measured value, and records 60 cases or families for Snijders Blok-Campeau syndrome, 50 for Sifrim-Hitz-Weiss syndrome, and 5 for Pilarowski-Bjornsson syndrome [2, 5, 7]. The available figures are counts of published cases, which grow with access to exome sequencing: in Snijders Blok-Campeau syndrome, 35 patients in 2018, 59 in 2020, 83 in 2023, and 237 known cases in October 2026 according to the CHD3 Foundation (a figure self-reported by the organisation, without independent confirmation) [19–22]. In Sifrim-Hitz-Weiss syndrome, about 50 patients were known to the GeneReviews authors in 2020, and the 2025 cohort included 39 [4, 25]. In Pilarowski-Bjornsson syndrome, the 6 cases from 2018 grew to 56 in 2026 [14, 23]. In Parenti-Mignot syndrome, the delineation cohort had 16 patients [28].

In gnomAD, 15 carriers of CHD3 loss-of-function variants were found among 198,800 individuals, 12 of them women; this figure measures carrier frequency in the reference population, which includes people without a diagnosis, and not disease prevalence [12]. For the entities referred elsewhere, the yields in selected cohorts are likewise not equivalent to prevalence: in CHD2, about 1% of epileptic encephalopathy cohorts and 0.25% of about 10,000 patients tested with an epilepsy panel [31]; in CHD8, 9 of 2446 patients with ASD, 8 of 3730 with neurodevelopmental disorders, and 12 of 710 (1.7%) with overgrowth and ID [35]. In CHARGE syndrome, the birth prevalence of 1/8500 live births is the highest provincial figure from the three-year Canadian surveillance begun in September 2001 (77 confirmed cases), and the national estimate recorded by Orphanet from the same study is 3.5/100,000; the Netherlands estimated 1/15,000 newborns, and a 2006 review gave an incidence of 0.1–1.2/10,000, dependent on clinical recognition [33, 34, 75, 76].

9.2 Origin of the cohorts and populations

The cohorts come mainly from Europe and North America. The founding CHD3 cohort was assembled through the GeneMatcher platform, the second included patients from Australia, Canada, China, France, Germany, Italy, the Netherlands, and the USA, the 2023 series patients from Spain, India, Italy, and France, the families with transmission came from ERN ITHACA centres and the Solve-RD project, and the episignature cohort assembled 38 patients from 10 European countries through the French AnDDi-Rares network and ERN ITHACA, without breaking them down by country [12, 19, 21, 22, 24]. Outside Europe and North America, Japanese and Chinese patients have been described [41, 50]. The p.(Arg966Trp) variant is absent from East Asian populations in gnomAD [42]. The 2020 CHD4 cohort was recruited at the NIH, in the DDD study, at GeneDx, and in the Pediatric Cardiac Genomic Consortium, and the 2025 cohort at ERN ITHACA [13, 25]. The 2026 CHD1 cohort included 14 countries, and a Saudi girl, the daughter of third cousins, with the CHD1 variant and an additional ASH1L variant was described [14, 54]. No founder variants have been described in any gene of the class. The pain survey had 80% white participants, and the neurobehavioural studies do not analyse ethnic differences [39, 40]. None of the cohorts consulted identify patients from Portugal or Brazil, which does not demonstrate the absence of cases described in these countries.

9.3 Special populations

Adults are underrepresented: the series include patients up to the fifth and sixth decades, but without longitudinal follow-up, and in CHD4 adults are probably underdiagnosed [1, 4, 39]. Women with CHD3 are overrepresented among transmitting parents and population carriers, and women with CHD1 appear to have more severe expression, which makes sex a relevant variable in counselling [12, 14]. In nonverbal patients, assessment of pain and distress requires observational instruments [40].

9.4 Reference networks and patient organisations

The relevant European reference network is ERN ITHACA (rare malformation syndromes, intellectual disability, and other neurodevelopmental disorders), with a working group for neurodevelopmental disorders, a group for guidelines and consensus statements, the ILIAD registry, the SysNDD database, and the CPMS system for cross-border multidisciplinary consultation; the network served as the recruitment source for the CHD3 and CHD4 episignature cohorts [24, 25, 27]. The CHD3 Foundation (chd3.org) is the international organisation of families with Snijders Blok-Campeau syndrome, with a support group, research funding, and financial support for families, and collaborated on the pain study [20, 40]. The Simons Searchlight registry includes the genes CHD3, CHD2, and CHD8 [31, 73]. For CHD4, GeneReviews lists the American Association on Intellectual and Developmental Disabilities (AAIDD), the CDC resources on child development, MedlinePlus, and the patient registry of the Human Disease Genes Website Series [4]. For CHD2, there is the Coalition to Cure CHD2, and for CHD8 the British organisation Unique [31, 35].

10. Illustrative Clinical Case

10.1 Two siblings with a microcephalic phenotype and parental gonosomal mosaicism (CHD3)

Enomoto and colleagues described, in a Japanese series of five patients from four families with CHD3 variants confirmed by trio Sanger sequencing, two siblings with the variant NM_001005273.3:c.5360A>G p.(Glu1787Gly) (chr17:7908795, GRCh38), in C-terminal domain 2 [41]. Patient 4 was born at 38 weeks with a weight of 1742 g (−3.49 standard deviations [SD]), a length of 42 cm (−2.78 SD), and an HC of 28 cm (−3.55 SD), after intrauterine growth restriction [41]. The course was one of poor weight gain, microcephaly, hypotonia, global and speech delay, and autistic traits, with prominent ears, full lips, retrognathia, a broad nasal bridge, and midface hypoplasia; ophthalmological examination showed strabismus, astigmatism, hypermetropia, ptosis, and amblyopia; epilepsy began at age 5; and the IQ was between 21 and 35, without a friendly personality or overgrowth [41]. The sibling (patient 5), born at term at 2302 g (−2.63 SD), 46.5 cm (−1.44 SD), and an HC of 31.2 cm (−1.59 SD), also with intrauterine growth restriction, had global and speech delay, hypotonia, autistic traits, microcephaly, strabismus, astigmatism, hypermetropia, otitis media with effusion, and severe ID (IQ of 35), without seizures [41]. The variant arose from gonosomal mosaicism in a parent, with 3% mutant alleles, and the unaffected sibling did not carry it [41]. Figure 1 of the article, showing the patients and the location of the variants, is available at https://pmc.ncbi.nlm.nih.gov/articles/PMC12185722/figure/Fig1/ [41].

The case illustrates three points. First, Snijders Blok-Campeau syndrome is not synonymous with macrocephaly: these siblings belong to the phenotype 2 proposed by the authors (microcephaly, growth restriction, and severe ID), in contrast to patients with helicase C-terminal domain variants, who are macrocephalic and have a friendly personality [41]. Second, an apparently de novo variant may recur within a sibship through low-level parental mosaicism, which contradicts the idea of a negligible recurrence risk [41]. Third, the severity within the same family was similar for ID but different for epilepsy, which confirms the variable expressivity already described in other families [12, 41].

10.2 The index case: childhood apraxia of speech (CHD3)

The syndrome's index case was a boy (individual 22 of the founding cohort) included in a whole-genome sequencing study of children with rare speech disorders, with a formal diagnosis of severe apraxia of speech at age 3 [21]. He communicated effectively through sign language, had severe expressive speech problems with relatively normal scores on language comprehension tests, and a composite IQ of 72 on the KBIT scale [21]. Sequencing identified the de novo variant NM_001005273.3:c.3505C>T p.(Arg1169Trp) in the helicase domain, also published by Eising and colleagues with the numbering p.(Arg1228Trp) of transcript NM_001005271 [9, 21, 38]. It was this variant that prompted the collaborative search that assembled the 35 probands [21]. Figure 1 of the founding article, with photographs of 18 individuals (the 2019 erratum removed the images of one of them), is at https://pmc.ncbi.nlm.nih.gov/articles/PMC6218476/figure/Fig1/ [21]. The case shows that AAC, in this case sign language, can ensure effective communication when speech is very limited, and that mild or borderline ID can coexist with a severe motor speech disorder [21].

10.3 A girl with Pilarowski-Bjornsson syndrome and an additional variant (CHD1)

Al-Aamri and colleagues described a Saudi girl born at 36 weeks, at 2.08 kg (10th percentile), by emergency caesarean section for reduced fetal movements and meconium-stained fluid, with Apgar scores of 6 and 8 and cord blood gas without severe acidosis [54]. She was admitted to the neonatal intensive care unit for respiratory distress and severe central hypotonia, which required orogastric tube feeding; MRI on the sixth day of life showed only a small subdural haemorrhage over the convexity [54]. Exome sequencing at 22 days identified the nonsense variant c.966G>A p.(Trp322Ter) in CHD1 and the missense variant c.1723T>G p.(Ser575Ala) in ASH1L [54]. She was discharged on nasal cannula and a feeding tube, was readmitted for aspiration pneumonia, depended on oxygen for 30 days and on the tube for 45 days, and underwent physical therapy for 18 months [54]. At age 3 she weighed 16 kg (90th percentile), measured 87 cm (10th percentile), and had an HC of 48 cm (25th–50th percentiles), a depressed midface, almond-shaped eyes, arched eyebrows, a pointed chin, hypotonia, autistic traits, absent expressive language, and a febrile seizure with a normal EEG; developmental quotients were 33 in the socioemotional domain, 16.7 in language and communication, 16.7 in the cognitive domain, and 66 in the motor domain, which constitutes developmental dissociation [54]. The parents, third cousins, declined exome testing and the publication of photographs [54]. The authors attribute the dissociation possibly to the ASH1L variant and note that the weight at the 90th percentile contrasts with the growth deficiency described in the disease, perhaps because of hypercaloric feeding [54]. Without parental segregation, the origin of the variant and the contribution of each gene remain unclear, and the case shows the difficulty of phenotype attribution when variants in two chromatin genes coexist [14, 54].

11. Prognosis and Long-Term Follow-Up

11.1 Prognosis

In Snijders Blok-Campeau syndrome, survival does not appear reduced: the cohorts include adults up to the fifth and sixth decades and record no deaths attributed to the disease, although the absence of recorded deaths in small, cross-sectional series does not prove a zero risk [1, 22, 39]. Walking is acquired late, usually by age 5, first words appear at a mean of 2 years and 10 months, and 6 of 33 patients in the founding cohort were nonverbal at last assessment [1, 21]. Adaptive functioning is the main determinant of prognosis: two thirds of the patients assessed had profound global impairment on the Vineland-3, with socialisation as a relative strength [39]. The prognosis of inherited cases is, on average, more favourable (no inherited proband with severe ID; 79% of carrier parents with no history of ID) [12]. One patient developed parkinsonism later in life, an isolated finding that does not allow a risk to be estimated [21].

In Sifrim-Hitz-Weiss syndrome, life expectancy is unknown; the mortality described is 2 of 32 patients (complications of heart disease in the neonatal period; cervical instability with prolonged tracheostomy at age 21), and severe congenital anomalies may shorten life [4, 13]. Morbidity is dominated by congenital heart disease, cervical fusions, hydrocephalus, and hearing loss; the risk of moyamoya persists into adulthood [25, 46]. Cancer has not been described, and adults are probably underdiagnosed [4]. In Pilarowski-Bjornsson syndrome, there are no long-term prognostic data; expression appears more severe in females [14]. In Parenti-Mignot syndrome, most patients were seizure-free at the time of the study, with or without treatment, but 4 of 16 had developmental and epileptic encephalopathy, and the 54-year-old adult case shows persistence of the difficulties with social cognition and emotional regulation [28, 55].

11.2 Surveillance in Sifrim-Hitz-Weiss syndrome

The table reproduces the surveillance recommended by GeneReviews (2020), supplemented by the subsequent data indicated.

11.3 Surveillance in Snijders Blok-Campeau syndrome

There is no published surveillance table for Snijders Blok-Campeau syndrome. The following table is an extrapolation built from the cohort frequencies (visual abnormalities in 72%, seizures in 16%, heart disease in 14%, hearing loss in 13%, cryptorchidism, laxity, scoliosis, and missing teeth) and from the intervals in the CHD4 GeneReviews chapter, and should be read as such [4, 22].

11.4 Surveillance in Pilarowski-Bjornsson and Parenti-Mignot syndromes

There is likewise no published surveillance for CHD1 or CHD5. By extrapolation from the phenotypes, surveillance in CHD1 centres on growth (frequent growth deficiency) at every visit, development and speech with assessment for apraxia, ASD and behaviour every 1–3 years, and EEG when seizures are suspected [6, 14]. In CHD5, surveillance for seizures at every visit, with EEG when suspected and urgent evaluation of infantile spasms, surveillance of skull shape in infancy (craniosynostosis), and psychiatric and neuropsychological assessment are the priorities [28, 55].

11.5 Transition and adulthood

Transition to adult services should begin in adolescence, according to the general ERN ITHACA guideline of March 2026, with a plan that includes communication, pain assessment, independence and, in CHD4, continuity of cardiac, auditory, cervical, and endocrine surveillance and attention to adult-onset moyamoya [27, 40, 46]. Data on physical health and independence in middle age are nonexistent for all the entities [39].

12. Summary and Concluding Remarks

CHD gene–related disorders form a class of autosomal dominant neurodevelopmental diseases caused by variants in ATP-dependent chromatin remodellers. CHD3, CHD4, and CHD5, members of subfamily II, are alternative subunits of the NuRD complex; CHD1 and CHD2 belong to subfamily I; and CHD7 and CHD8 to subfamily III. The shared clinical core is global developmental delay with disproportionate impairment of expressive speech, ID of variable severity, hypotonia, and mild craniofacial dysmorphism with a prominent forehead and widely spaced eyes; the organ findings distinguish the entities.

Snijders Blok-Campeau syndrome, described in 2018 starting from a child with speech apraxia, combines universal speech and language disorder, ID ranging from borderline to severe that may be absent, macrocephaly in about half (with a minority microcephalic phenotype), visual abnormalities in about 70%, hypotonia, laxity, missing permanent teeth, a sociable personality, and a reduced or atypical response to pain. De novo variants are mainly missense clustered in the ATPase/helicase core, whereas inherited variants, often truncating and maternally transmitted, show reduced penetrance and variable expressivity. This architecture requires examining the parents before interpreting an inherited variant, not regarding the recurrence risk as negligible when the parents are negative, and using the methylation episignature published in 2026 as support for, and not a substitute for, genomic analysis.

Sifrim-Hitz-Weiss syndrome is distinguished by congenital heart disease, hearing loss, male hypogonadotropic hypogonadism, cervical vertebral fusions with a risk of instability, hydrocephalus, and moyamoya, and it has the only published surveillance table of the class, which dates from 2020 and should be supplemented with measurement of the aortic root and attention to intestinal malrotation. Pilarowski-Bjornsson syndrome, initially described only in girls, revealed in 2026 an excess of males in the cohorts, more severe expression in females, and an androgen-mediated protective effect in the mouse model, and its limited ClinGen classification predates those data. Parenti-Mignot syndrome adds severe epilepsy and craniosynostosis to the class.

There is no disease-modifying treatment or specific guideline, and management is supportive: speech therapy with early augmentative communication, physical and occupational therapy, educational support, treatment of ADHD, anxiety, and epilepsy according to general indications, structured pain assessment and, in CHD4, neurosurgical, cervical, cardiac, auditory, and endocrine surveillance. Gene correction by base editing has demonstrated results in a mouse model of a single CHD3 variant, but the only registered human trial ended with the participant's death from product-related thrombotic microangiopathy, and this information should be part of counselling for families. ERN ITHACA and the CHD3 Foundation are the reference resources for professionals and families, and the lack of longitudinal studies, of adult data, and of data from Portugal and Brazil defines the main gaps.

13. References

1. Online Mendelian Inheritance in Man. Snijders Blok-Campeau syndrome; SNIBCPS. MIM 618205. Baltimore: Johns Hopkins University. Available from: https://www.omim.org/entry/618205 [cited 2026 Oct 18]
2. Orphanet. CHD3-related developmental delay-speech delay-intellectual disability-abnormalities of vision-facial dysmorphism syndrome (ORPHA:599082). https://www.orpha.net/en/disease/detail/599082 [cited 2026 Oct 18]
3. Online Mendelian Inheritance in Man. Sifrim-Hitz-Weiss syndrome; SIHIWES. MIM 617159. Available from: https://www.omim.org/entry/617159 [cited 2026 Oct 18]
4. Weiss K, Lachlan K. CHD4 Neurodevelopmental Disorder. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews®. Seattle (WA): University of Washington; 1993–. Initial posting 2020 Sep 3. https://www.ncbi.nlm.nih.gov/books/NBK561516/. PMID: 32881470.
5. Orphanet. CHD4-related neurodevelopmental disorder (ORPHA:653712). https://www.orpha.net/en/disease/detail/653712 [cited 2026 Oct 18]
6. Online Mendelian Inheritance in Man. Pilarowski-Bjornsson syndrome; PILBOS. MIM 617682. Available from: https://www.omim.org/entry/617682 [cited 2026 Oct 18]
7. Orphanet. Intellectual disability-autism-speech apraxia-craniofacial dysmorphism syndrome (ORPHA:529965). https://www.orpha.net/en/disease/detail/529965 [cited 2026 Oct 18]
8. GeneGuidance. CHD3 — gene record (clinical curation). Version of 2026-09-26.
9. VariantValidator. HGVS nomenclature validation and MANE Select transcripts, GRCh38. Available from: https://variantvalidator.org [cited 2026 Oct 18]
10. GeneGuidance. CHD4 — gene record (clinical curation). Version of 2026-09-26.
11. GeneGuidance. CHD1 — gene record (clinical curation). Version of 2026-09-26.
12. van der Spek J, den Hoed J, Snijders Blok L, Dingemans AJM, Schijven D, Nellaker C, et al. Inherited variants in CHD3 show variable expressivity in Snijders Blok-Campeau syndrome. Genet Med. 2022;24(6):1283–96. doi:10.1016/j.gim.2022.02.014. PMID: 35346573.
13. Weiss K, Lazar HP, Kurolap A, Martinez AF, Paperna T, Cohen L, et al. The CHD4-related syndrome: a comprehensive investigation of the clinical spectrum, genotype-phenotype correlations, and molecular basis. Genet Med. 2020;22(2):389–97. doi:10.1038/s41436-019-0612-0. PMID: 31388190.
14. Anderson KJ, Thorolfsdottir ET, Nodelman IM, Halldorsdottir B, et al. Androgens mediate sexual dimorphism in Pilarowski-Bjornsson syndrome. Am J Hum Genet. 2026;113(9):1946–71. doi:10.1016/j.ajhg.2026.08.005. PMID: 42692004.
15. Online Mendelian Inheritance in Man. Chromodomain helicase DNA-binding protein 3; CHD3. MIM 602120. Available from: https://www.omim.org/entry/602120 [cited 2026 Oct 18]
16. Online Mendelian Inheritance in Man. Chromodomain helicase DNA-binding protein 4; CHD4. MIM 603277. Available from: https://www.omim.org/entry/603277 [cited 2026 Oct 18]
17. Online Mendelian Inheritance in Man. Chromodomain helicase DNA-binding protein 1; CHD1. MIM 602118. Available from: https://www.omim.org/entry/602118 [cited 2026 Oct 18]
18. ClinGen Gene-Disease Validity curations (file downloaded 2026-10-18). Clinical Genome Resource. Available from: https://search.clinicalgenome.org/kb/gene-validity [cited 2026 Oct 18]
19. Pascual P, Tenorio-Castano J, Mignot C, Afenjar A, Arias P, Gallego-Zazo N, et al. Snijders Blok-Campeau syndrome: description of 20 additional individuals with variants in CHD3 and literature review. Genes (Basel). 2023;14(9):1664. doi:10.3390/genes14091664. PMID: 37761804.
20. CHD3 Foundation. Snijders Blok-Campeau syndrome — community, research and family support. Available from: https://www.chd3.org/ [cited 2026 Oct 18]
21. Snijders Blok L, Rousseau J, Twist J, Ehresmann S, Takaku M, Venselaar H, et al. CHD3 helicase domain mutations cause a neurodevelopmental syndrome with macrocephaly and impaired speech and language. Nat Commun. 2018;9(1):4619. doi:10.1038/s41467-018-06014-6. PMID: 30397230.
22. Drivas TG, Li D, Nair D, Alaimo JT, Alders M, Altmüller J, et al. A second cohort of CHD3 patients expands the molecular mechanisms known to cause Snijders Blok-Campeau syndrome. Eur J Hum Genet. 2020;28(10):1422–31. doi:10.1038/s41431-020-0654-4. PMID: 32483341.
23. Pilarowski GO, Vernon HJ, Applegate CD, Boukas L, Cho MT, Gurnett CA, et al. Missense variants in the chromatin remodeler CHD1 are associated with neurodevelopmental disability. J Med Genet. 2018;55(8):561–6. doi:10.1136/jmedgenet-2017-104759. PMID: 28866611.
24. Santini A, Tognon A, Richard AC, Velasco G, Phan G, Marzin P, et al. Identification of an episignature for CHD3-related Snijders Blok-Campeau syndrome reveals heterogeneity in the CHARGE syndrome episignature: towards a better characterisation of chromatinopathies. Genome Med. 2026;18(1):39. doi:10.1186/s13073-026-01639-5. PMID: 41952182.
25. Karimi K, Lichtenstein Y, Reilly J, McConkey H, Relator R, Levy MA, et al. Discovery of a DNA methylation profile in individuals with Sifrim-Hitz-Weiss syndrome. Am J Hum Genet. 2025;112(2):414–27. doi:10.1016/j.ajhg.2024.12.020. PMID: 39824190.
26. ClinicalTrials.gov. Safety, tolerability and preliminary efficacy study of a single intrathecal injection of the dual vector AAV-CHD3-R1025W base editor for the treatment of developmental disorders caused by the R1025W mutation in the CHD3 gene. NCT06860672. Terminated. Available from: https://clinicaltrials.gov/study/NCT06860672 [cited 2026 Oct 18]
27. European Reference Network ITHACA (Rare Malformation Syndromes, Intellectual and Other Neurodevelopmental Disorders). Available from: https://ern-ithaca.eu/ [cited 2026 Oct 18]
28. Parenti I, Lehalle D, Nava C, Torti E, Leitão E, Person R, et al. Missense and truncating variants in CHD5 in a dominant neurodevelopmental disorder with intellectual disability, behavioral disturbances, and epilepsy. Hum Genet. 2021;140(7):1109–20. doi:10.1007/s00439-021-02283-2. PMID: 33944996.
29. Online Mendelian Inheritance in Man. Chromodomain helicase DNA-binding protein 5; CHD5. MIM 610771. Available from: https://www.omim.org/entry/610771 [cited 2026 Oct 18]
30. Orphanet. Gene–disease links (Orphadata, local copy in books.db: tables orpha and orpha_gene) [cited 2026 Oct 18]
31. Carvill GL, Mefford HC. CHD2-Related Neurodevelopmental Disorders. In: Adam MP, et al., editors. GeneReviews®. Seattle (WA): University of Washington, Seattle; 1993–. Initial posting 2015 Dec 10; updated 2021 Jan 21. Available from: https://www.ncbi.nlm.nih.gov/books/NBK333201/
32. Orphanet. Lennox-Gastaut syndrome (ORPHA:2382). https://www.orpha.net/en/disease/detail/2382 [cited 2026 Oct 18]
33. van Ravenswaaij-Arts CM, Hefner M, Blake K, Martin DM. CHD7 Disorder. In: Adam MP, et al., editors. GeneReviews®. Seattle (WA): University of Washington, Seattle; 1993–. Initial posting 2006 Oct 2; last update 2020 Sep 17; revised 2025 Aug 14. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1117/
34. Orphanet. CHARGE syndrome (ORPHA:138). Available from: https://www.orpha.net/en/disease/detail/138 [cited 2026 Oct 18]
35. Mitchel MW, Myers SM, Heidlebaugh AR, Taylor CM, Rea H, Neuhaus E, et al. CHD8-Related Neurodevelopmental Disorder with Overgrowth. In: Adam MP, et al., editors. GeneReviews®. Seattle (WA): University of Washington, Seattle; 1993–. Initial posting 2022 Oct 27. Available from: https://www.ncbi.nlm.nih.gov/books/NBK585456/
36. Orphanet. CHD8 overgrowth syndrome (ORPHA:642675). https://www.orpha.net/en/disease/detail/642675 [cited 2026 Oct 18]
37. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 1, Clinical Approach to Inborn Errors of Metabolism in Paediatrics, section 1.5.
38. Eising E, Carrion-Castillo A, Vino A, Strand EA, Jakielski KJ, Scerri TS, et al. A set of regulatory genes co-expressed in embryonic human brain is implicated in disrupted speech development. Mol Psychiatry. 2019;24(7):1065–78. PMID: 29463886.
39. Ionescu A, Mazur-Lainé E, Protsenko M, et al. Neurobehavioral profile of individuals with pathogenic variants in CHD3. Eur J Hum Genet. 2025;33(10):1350–8. doi:10.1038/s41431-025-01926-6. PMID: 40830229.
40. Ocay DD, Campeau PM, Berde CB, et al. Pain experience and perception in individuals with Snijders Blok-Campeau syndrome. Front Pain Res (Lausanne). 2025;6:1540422. doi:10.3389/fpain.2025.1540422. PMID: 40881826.
41. Enomoto Y, Shiromizu T, Yasojima T, et al. Two distinct phenotypes in Snijders Blok-Campeau syndrome and characterization of the behavioral phenotype in a zebrafish model. Eur J Hum Genet. 2025;33(6):747–57. doi:10.1038/s41431-025-01815-y. PMID: 39988727.
42. Yang K, Li WK, Geng YX, Zhang SQ, Wu SH, Cheng YB, et al. In vivo base editing of Chd3 rescues behavioural abnormalities in mice. Nature. 2026;651(8106):785–95. doi:10.1038/s41586-026-10113-6. PMID: 41708849.
43. Paul L, Ziesenitz VC, Gorenflo M, et al. Snijders Blok-Campeau syndrome associated with pulmonary arterial hypertension: a case report. Reports (MDPI). 2025;8(2). PMID: 40710838.
44. Sifrim A, Hitz MP, Wilsdon A, Breckpot J, Al Turki SH, Thienpont B, et al. Distinct genetic architectures for syndromic and nonsyndromic congenital heart defects identified by exome sequencing. Nat Genet. 2016;48(9):1060–5. doi:10.1038/ng.3627. PMID: 27479907.
45. Weiss K, Terhal PA, Cohen L, Bruccoleri M, Irving M, Martinez AF, et al. De novo mutations in CHD4, an ATP-dependent chromatin remodeler gene, cause an intellectual disability syndrome with distinctive dysmorphisms. Am J Hum Genet. 2016;99(4):934–41. doi:10.1016/j.ajhg.2016.08.001. PMID: 27616479.
46. Pinard A, Guey S, Guo D, Cecchi AC, Kharas N, Wallace S, et al. The pleiotropy associated with de novo variants in CHD4, CNOT3, and SETD5 extends to moyamoya angiopathy. Genet Med. 2020;22(2):427–31. doi:10.1038/s41436-019-0639-2. PMID: 31474762.
47. GeneGuidance. CHD8 — gene record (clinical curation). Version of 2026-09-26.
48. Koparir A, Kerkhof J, Rzasa J, et al. Chromatinopathies: clinically overlapping disorders, revealing novel variants and their DNA methylation signatures. Clin Epigenetics. 2026;18(1). doi:10.1186/s13148-026-02120-1. PMID: 41957673.
49. Mitchell ZH, den Hoed J, Claassen W, Demurtas M, Deelen L, Campeau PM, et al. The NuRD component CHD3 promotes BMP signalling during cranial neural crest cell specification. EMBO Rep. 2025;26(19):4723–41. doi:10.1038/s44319-025-00555-w. PMID: 40835974.
50. Gao Y, Wang P, Chen M, Pang K, et al. Novel genotypes and phenotypes in Snijders Blok-Campeau syndrome caused by CHD3 mutations. Front Genet. 2024;15:1347933. doi:10.3389/fgene.2024.1347933. PMID: 39050258.
51. GeneGuidance. CHD5 — gene record (clinical curation). Version of 2026-09-26.
52. GeneGuidance. CHD2 — gene record (clinical curation). Version of 2026-09-26.
53. GeneGuidance. CHD7 — gene record (clinical curation). Version of 2026-09-26.
54. Al-Aamri M, Alshaqaq M, Al-Abdulwahed H, et al. A Saudi girl with co-occurring CHD1 (Pilarowski-Bjornsson syndrome) and ASH1L gene variants. Cureus. 2023;15(12):e49905. doi:10.7759/cureus.49905. PMID: 38174187.
55. Melgers C, Roelofs R, van Bon B, Wingbermühle E, Pfundt R, Egger J. New insights into the relation between cognition, behavior, and the CHD5 gene: a case-report of an adult male with Parenti-Mignot neurodevelopmental syndrome. Mol Syndromol. 2025. doi:10.1159/000550089. PMID: 41584030.
56. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 412–413.
57. AbbVie Inc. DEPAKOTE (divalproex sodium) delayed-release tablets: prescribing information [Internet]. DailyMed, U.S. National Library of Medicine; revision published 14 Apr 2026 [cited 2026 Oct 18]. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=08a65cf4-7749-4ceb-6895-8f4805e2b01f
58. Lundbeck Pharmaceuticals LLC. SABRIL (vigabatrin) powder for oral solution: prescribing information [Internet]. DailyMed, U.S. National Library of Medicine; revision published 17 Dec 2025 [cited 2026 Oct 18]. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=a88ac1b4-e2c9-45c0-b321-4785902172e3
59. European Medicines Agency. Slenyto (melatonin) prolonged-release tablets: EPAR — product information (summary of product characteristics) [Internet]. Amsterdam: EMA [cited 2026 Oct 18]. Available from: https://www.ema.europa.eu/en/documents/product-information/slenyto-epar-product-information_en.pdf
60. Sandoz Inc. RITALIN (methylphenidate hydrochloride) tablets: prescribing information [Internet]. DailyMed, U.S. National Library of Medicine; revision published 21 Feb 2025 [cited 2026 Oct 18]. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=d6fb2750-cdab-4749-ba0d-7534840a5892
61. Janssen Pharmaceuticals Inc. RISPERDAL (risperidone) tablets, orally disintegrating tablets and oral solution: prescribing information [Internet]. DailyMed, U.S. National Library of Medicine; revision published 20 Aug 2026 [cited 2026 Oct 18]. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=7e117c7e-02fc-4343-92a1-230061dfc5e0
62. Pfizer Inc. GENOTROPIN (somatropin) for injection: prescribing information [Internet]. DailyMed, U.S. National Library of Medicine; revision published 6 Aug 2026 [cited 2026 Oct 18]. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=ffebf88b-d257-4542-9808-74d9b7167765
63. INFARMED. Depakine 200 mg/ml oral solution (valproic acid): summary of product characteristics [Internet]. Infomed (med_id 2413). Lisbon: INFARMED [cited 2026 Oct 18]. Available from: https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=2413&tipo_doc=rcm
64. European Medicines Agency, CMDh. Valproate and related substances: Article 31 referral — Annex III (amendments to the summary of product characteristics) and new measures to avoid valproate exposure during pregnancy (EMA/375438/2018, 31 May 2018) [Internet]. Amsterdam: EMA; 2018 [cited 2026 Oct 18]. Available from: https://www.ema.europa.eu/en/medicines/human/referrals/valproate-related-substances-0
65. European Medicines Agency. Potential risk of neurodevelopmental disorders in children born to men treated with valproate medicines: PRAC recommends precautionary measures [Internet]. Amsterdam: EMA; 12 Jan 2024 [cited 2026 Oct 18]. Available from: https://www.ema.europa.eu/en/news/potential-risk-neurodevelopmental-disorders-children-born-men-treated-valproate-medicines-prac-recommends-precautionary-measures
66. INFARMED. Sabril 500 mg film-coated tablets (vigabatrin): summary of product characteristics [Internet]. Infomed, database of medicinal products for human use. Lisbon: INFARMED [cited 2026 Oct 18]. Available from: https://extranet.infarmed.pt/INFOMED-fo/
67. European Medicines Agency. Methylphenidate: Article 31 referral — Annexes I–IV [Internet]. London: EMA; 2009 [cited 2026 Oct 18]. Available from: https://www.ema.europa.eu/en/medicines/human/referrals/methylphenidate
68. INFARMED. Rubifen 10 mg tablets (methylphenidate): summary of product characteristics [Internet]. Infomed (med_id 36077). Lisbon: INFARMED [cited 2026 Oct 18]. Available from: https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=36077&tipo_doc=rcm
69. European Medicines Agency. Risperdal and associated names (risperidone): Article 30 referral — Annexes I–IV [Internet]. Amsterdam: EMA [cited 2026 Oct 18]. Available from: https://www.ema.europa.eu/en/medicines/human/referrals/risperdal
70. INFARMED. Genotropin 5.3 mg/1 ml and 12 mg/1 ml powder and solvent for solution for injection (somatropin): summary of product characteristics [Internet]. Infomed (med_id 3849). Lisbon: INFARMED [cited 2026 Oct 18]. Available from: https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=3849&tipo_doc=rcm
71. European Medicines Agency. Omnitrope (somatropin): EPAR — product information [Internet]. Amsterdam: EMA [cited 2026 Oct 18]. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/omnitrope
72. ClinicalTrials.gov. Searches "CHD3," "CHD4," "Snijders Blok," "Sifrim-Hitz-Weiss," and "Pilarowski" (API v2). Available from: https://clinicaltrials.gov [cited 2026 Oct 18]
73. ClinicalTrials.gov. Online study of people who have genetic changes and features of autism: Simons Searchlight. NCT01238250. Recruiting. Available from: https://clinicaltrials.gov/study/NCT01238250 [cited 2026 Oct 18]
74. Dudley A, Widdess-Walsh P, Linehan C, Andrade D, Delanty N, Moloney P. Genetic developmental and epileptic encephalopathies in adulthood: gaps in evidence and care. Epilepsia Open. 2026. Available from: https://pmc.ncbi.nlm.nih.gov/articles/PMC13666571/
75. Issekutz KA, Graham JM Jr, Prasad C, Smith IM, Blake KD. An epidemiological analysis of CHARGE syndrome: preliminary results from a Canadian study. Am J Med Genet A. 2005;133A(3):309–17. doi:10.1002/ajmg.a.30560. PMID: 15637722.
76. Blake KD, Prasad C. CHARGE syndrome. Orphanet J Rare Dis. 2006;1:34. doi:10.1186/1750-1172-1-34. PMID: 16959034.