Multiple Congenital Anomalies — General Approach

Classification of errors of morphogenesis, epidemiology, chromosomal, genomic, monogenic, teratogenic, mechanical and multifactorial aetiologies, dysmorphological assessment, postnatal and prenatal genomic testing strategy, diagnostic yield, counselling with and without a diagnosis, and follow-up of the undiagnosed child.

Literature and regulatory information updated: 17 October 2026.

0. Initial Reference Table

1. Introduction

1.1. From dysmorphology to genomics

The term dysmorphology refers to the recognition and study of birth defects and syndromes; it was first used by David Smith, in the USA, in the 1960s, for the study of human congenital malformations and patterns of birth defects [1]. In 1964, Marden, Smith and McDonald performed a surface examination of 4412 newborns, recording major anomalies and minor variants, and took buccal smears for sex chromatin; that study founded the quantitative relationship between the number of minor anomalies and the probability of a major anomaly, which still guides the examination of the newborn [25, 26]. In 1982, an international working group (Spranger, Benirschke, Hall, Lenz, Lowry, Opitz, Pinsky, Schwarzacher and Smith) published the recommendations "Errors of morphogenesis: concepts and terms", which established the concepts of malformation, disruption, deformation, dysplasia, sequence, syndrome, association and developmental field [2]. The original text was not accessed for this article; the definitions given in 1.2 follow the wording of the Oxford Desk Reference: Clinical Genetics, which derives from that terminological tradition [1].

G-banded cytogenetics, with a resolution of about 5–10 Mb, identified aneuploidies and visible rearrangements; fluorescence in situ hybridisation (FISH) and multiplex ligation-dependent probe amplification (MLPA) later made it possible to detect known microdeletions and subtelomeric rearrangements [12, 20]. In 2010, the International Standard Cytogenomic Array (ISCA) consensus, based on 33 studies with 21,698 patients, recommended CMA instead of the karyotype as the first-tier cytogenetic test in unexplained developmental delay (DD) or ID, ASD or MCA [22]. Next-generation sequencing (NGS) made multigene panels, the exome and the genome possible; the Deciphering Developmental Disorders (DDD) study, in the United Kingdom and Ireland, analysed more than 13,000 children with severe developmental disorders by exome and array and added 60 new genes [10, 27]. In 2021, the American College of Medical Genetics and Genomics (ACMG) strongly recommended the exome or genome as a first- or second-tier test in patients with one or more congenital anomalies with onset before 1 year [21], and in 2025 the American Academy of Pediatrics (AAP) placed the exome or genome, together with CMA, in the first tier of the evaluation of ID and global developmental delay [9].

Standardisation of phenotypic description accompanied this evolution: the Elements of Morphology series, begun in 2009 by 34 dysmorphologists, defined the terms for the head, face, hands, feet, mouth, ear, periorbital structures, nose and philtrum, and the Human Phenotype Ontology (HPO), described in 2008, encodes phenotypes for computational analysis [1, 28]. Hennekam and Biesecker summarised the practical consequence in 2012: NGS requires next-generation phenotyping [1].

1.2. Morphogenetic definitions

The categories are distinguished by mechanism, and the mechanism has implications for prognosis and recurrence risk [1]. The table summarises the definitions used in this article.

The US National Birth Defects Prevention Network (NBDPN) defines a syndrome as a pattern of anomalies forming a specific diagnosis whose natural history and recurrence risk are usually known, which implies a specific common aetiology [32]. The WHO defines congenital disorders (congenital anomalies, malformations or birth defects) as structural or functional anomalies, for example metabolic, that occur during intrauterine life and can be identified prenatally, at birth or later in childhood, such as deafness [3].

1.3. Major anomalies, minor anomalies and common variants

Aase's classification (1990) divides phenotypic abnormalities into anomalies proper (malformations, deformations, disruptions and dysplasias) and minor variants; the latter are subdivided into minor anomalies, with a prevalence of 4% or less in the general population, and common variants, above 4% [33]. For Smith, minor anomalies are unusual morphological features without serious medical or cosmetic consequence, valuable as indicators of altered morphogenesis or as clues to a specific pattern; many are deformations due to mechanical forces acting on normal tissue, and a minor deformation may result from a more primary malformation, which explains the association between minor anomalies and major malformations [26]. EUROCAT defines a major congenital anomaly as a structural change with significant medical, social or cosmetic consequences [8]. The term "dysmorphic", strictly speaking, means abnormality of structure, and internal organs can be affected by the same mechanism as visible malformations [1].

1.4. Operational definition of multiple congenital anomalies

EUROCAT defined MCA as two or more major anomalies of different organ systems, excluding sequences and chromosomal and monogenic syndromes, identified by a computer algorithm and reviewed manually by geneticists [16]. In other clinical series, MCA corresponds to two or more birth defects in any combination, including cases with a recognised syndrome [17]. The difference is material: the first definition measures the residual group without a syndromic diagnosis, the second the clinical starting point before testing. Orphanet groups genetic multiple congenital anomalies/dysmorphic syndromes (ORPHA:183533) and lethal ones (ORPHA:459787) as classification categories, without a definition of their own [5, 6].

1.5. Classifications by system

For limb reduction defects, the EUROCAT classification (Stoll, Mastroiacovo and colleagues, 1988) distinguishes five categories: terminal transverse (absence of distal structures with near-normal proximal structures, from one nail or phalanx to the whole limb; rudimentary digital nubbins do not reclassify the defect, unless there are significant skeletal elements); proximal intercalary (absence or severe hypoplasia of proximal segments with digits present, for example radial aplasia with an intact thumb or femoral hypoplasia); longitudinal (absence or severe hypoplasia of the lateral part, such as radial aplasia with absent thumb or ulnar aplasia with absent 5th digit); split hand or foot (absence of the central digits, as in EEC syndrome (ectrodactyly, ectodermal dysplasia and clefting)); and multiple types [34]. Classification by mechanism should precede classification by organ: a transverse defect suggests vascular disruption or amniotic bands, whereas a preaxial longitudinal defect raises the possibility of a genetic syndrome, and in EUROCAT 29% of preaxial longitudinal defects had a genetic diagnosis [8].

1.6. Scope of the article

This article covers the reasoning common to all congenital anomalies; specific entities enter the differential diagnosis: ID and developmental delay, congenital heart disease, congenital anomalies of the kidney and urinary tract, VACTERL association, severe and lethal skeletal dysplasias diagnosed prenatally, teratogens (alcohol, hydantoin, valproate), non-immune hydrops fetalis, posterior fossa malformations, congenital diaphragmatic hernia and anterior segment dysgeneses. MCAHS1 syndrome due to PIGN variants appears only as an illustrative case of the testing strategy.

2. Molecular Genetics and Pathophysiology

2.1. Genetic control of morphogenesis

Morphogenesis is genetically controlled, but the genetic potential is modulated by prenatal and postnatal environmental factors; among malformations and syndromes of known aetiology, most appear to be genetically determined [12]. The haploid genome has just over 20,000 protein-coding genes and about 1.5% of the genome encodes proteins; the three main mechanisms are altered gene dosage (chromosomal and genomic anomalies), single-gene alterations and polygenic susceptibility modulated by other genes and by the environment [12]. The contribution of each mechanism depends on the developmental stage: about half of 1st-trimester miscarriages have a chromosomal anomaly, compared with 6/1000 live births [12].

Germline alterations affect developmental programming through four main routes, according to the synthesis by Donnai and Read cited in the Oxford Desk Reference: failure of structural integrity, through qualitative or quantitative defects of structural molecules (COL2A1 in Stickler syndrome); failure of regulation of cell number (ASPM, MCPH5, in autosomal recessive primary microcephaly); failure of cell migration (MID1 in Opitz syndrome); and failure of a developmental "switch", through deficiency of transcription factors or of cell–cell signalling [1]. Many genes (aneuploidy), several contiguous genes (microdeletions such as that of Williams syndrome) or a single gene may be involved, and some single variants are devastating [1].

2.2. Chromosomal anomalies

Cytogenetically visible chromosomal anomalies are present in at least 4% of recognised pregnancies; about 50% of spontaneous miscarriages and 0.5% of live births have a chromosomal anomaly, and among live births sex chromosome aneuploidies and trisomy 21 predominate [12]. Survival to term is estimated at about 1/500 45,X conceptuses, 4% of trisomies 18 and 13 and 20% of trisomies 21, and mosaicism appears to influence that survival; chromosomes 21, 18 and 13 are relatively gene-poor (chromosome 19 is the richest, with 1621 genes, and the Y the poorest, with 251), which relates to the viability of the respective trisomies [12]. Advanced maternal age is the recognised factor for autosomal trisomies, 47,XXX and 47,XXY; in 45,X the lost chromosome is usually paternal, in trisomy 21 the extra chromosome is maternal in 95% (mainly non-disjunction at meiosis I), 47,XXX is usually maternal and 47,XXY is of maternal or paternal origin in similar proportions; 1–5% of spermatozoa from chromosomally normal men are aneuploid, and oocyte aneuploidy is estimated at almost 25% [12].

Errors may result from gonadal mosaicism, non-disjunction at meiosis I or II, errors in the first zygotic divisions, somatic mosaicism or trisomic rescue; mosaic individuals show every gradation of the phenotype, the proportion of abnormal cells in blood does not by itself predict prognosis, and study of more than one tissue may be necessary [12]. Among structural rearrangements, a balanced translocation has no phenotype, unless there is a cryptic deletion at the breakpoints or disruption of an exonic or regulatory sequence, but it carries a risk of unbalanced gametes; about 4% of people with Down syndrome have 46 chromosomes with a Robertsonian translocation, and isochromosome Xq accounts for about 10% of Turner syndrome in live births [12]. The trisomies and Turner syndrome enter the differential diagnosis.

2.3. Copy number variants, microdeletions and microduplications

Chromosomal deletions and duplications are a major cause of malformation and ID; submicroscopic ones were not detectable on the G-banded karyotype [20]. The old concept of a "chromosomal phenotype" (facial dysmorphism, unusual growth parameters, developmental delay and characteristic behaviour) is used less today, because many microdeletions are due to the dosage of a single gene, such as KANSL1 in the 17q21.31 microdeletion; that pattern should rather prompt investigation of a genetic cause by array, panel, exome or genome [20]. Microdeletions and microduplications, taken together, are more frequent than aneuploidies in children with psychomotor delay presenting after birth [12].

Recurrent deletions and duplications result from non-allelic homologous recombination (NAHR) between low-copy repeats (LCR, segmental duplications), are reciprocal and have standard breakpoints; duplications were delineated less often because partial trisomy is better tolerated than monosomy, as shown by the 17p11.2 duplication that causes Charcot-Marie-Tooth disease type 1A [12, 20]. In the gene-rich subtelomeric regions, small imbalances may have a severe phenotype; terminal deletions are stabilised by telomere "healing", telomere retention (interstitial deletion) or telomere capture, and may result from a reciprocal translocation with duplication of the reciprocal segment [20]. CNV of 100 kb to 2 Mb are widely distributed in the genome of healthy people and many include genes, which makes interpretation dependent on parental studies and population databases [20].

Before arrays, a CNV inherited from a parent without a phenotype was considered benign and a de novo one pathogenic; today recurrent CNV with incomplete penetrance and variable expressivity are recognised, more frequent in children with neurodevelopmental disorders but often inherited from an apparently normal parent, acting as susceptibility loci for ASD, schizophrenia and epilepsy [20]. Smith places these CNV (1q21.1, 15q11.2, 15q13.3, 16p11.2, 16p13.3) within the multifactorial model, associated with ID, ASD, epilepsy, subtle dysmorphic features and specific heart defects such as tetralogy of Fallot [35]. The table reproduces the estimates of Rosenfeld and colleagues (2013), with cases studied by postnatal array and 22,246 controls; penetrance refers to a clinical phenotype sufficient to request an array, not to a specific disease [20].

Proximal 16p11.2 deletions showed a maternal transmission bias (14/68 carrier mothers versus 0/38 fathers; p=0.0018) [20]. The original coordinates are in hg18.

2.4. Monogenic disorders

Biallelic expression is the rule; monoallelic expression occurs in most X-chromosome genes and in about 100 autosomal genes [36]. The dominant mechanisms are haploinsufficiency, gain of function and dominant-negative effect; in osteogenesis imperfecta, an abnormal procollagen incorporated into the triple helix gives a severe phenotype, whereas a quantitative reduction causes mild fractures, and the same gene may have different phenotypes depending on the mechanism [36]. In recessive disorders, a reduction to half of the activity is tolerated, as with the 50% alpha-L-iduronidase of Hurler carriers [36]. Expression is variable (severe, usual, mild or absent, that is, non-penetrance), especially in structural defects [36].

De novo variants are the most prevalent cause of severe developmental disorders in the children of non-consanguineous parents; the DDD identified 94 genes enriched in damaging de novo variants at genome-wide significance (p <7×10⁻⁷), and the probability of a dominant de novo variant increases with female sex, advanced paternal age (and less so maternal age) and absence of consanguinity [37]. In the final DDD analysis (13,449 probands), 76% of clinically diagnosed trios had a pathogenic de novo variant; of the pathogenic or likely pathogenic variants, 79% were de novo (or inherited from a mosaic parent) in dominant genes, 32% in autosomal recessive genes, 23% X-linked inherited from the mother and 11% autosomal dominant inherited from an affected parent; 71% altered the protein, 19% were truncating, 3% non-coding, 6% CNV and 1% other structural variants [10]. Even in consanguineous families, about 5% of children with a developmental disorder have a dominant de novo variant [37, 38].

Few variants are unequivocally pathogenic on their own; the classic example is achondroplasia [39]. Normalisation in VariantValidator (GRCh38, MANE Select transcript) gives, for FGFR3, NM_000142.5:c.1138G>A, NP_000133.1:p.(Gly380Arg), NC_000004.12:g.1804392G>A, in achondroplasia, and NM_000142.5:c.1948A>G, NP_000133.1:p.(Lys650Glu), NC_000004.12:g.1806162A>G, in thanatophoric dysplasia type 2, lethal in the perinatal period [1, 40]. The FGFR3 spectrum includes achondroplasia, hypochondroplasia and thanatophoric dysplasia.

2.5. Imprinting, uniparental disomy, triploidy, repeats and mitochondrial DNA

Parent-of-origin effects are seen in triploidy: diandric triploidy (two paternal complements and one maternal) gives a large hydatidiform placenta with a small, malformed but proportionate fetus, and digynic triploidy gives disproportionate growth restriction with a very small placenta [36]. Fewer than 100 human imprinted genes are known, regulated by cytosine methylation and histone acetylation, erased in the germline and re-established early in embryogenesis and clustered with imprinting centres; Prader-Willi, Angelman and Beckwith-Wiedemann syndromes are the main examples [36]. Uniparental disomy frequently results from trisomic rescue, because in one-third of cases the two remaining chromosomes come from the same parent; it causes Beckwith-Wiedemann syndrome (paternal disomy 11p15), phenotypes in carriers of apparently balanced Robertsonian translocations involving chromosomes 14 or 15, Prader-Willi syndrome and autosomal recessive disease with only one carrier parent [36].

Unstable mutations due to repeat expansion entail premutation, anticipation, normal transmitting males and parent-of-origin effects (fragile X, myotonic dystrophy); polyalanine expansions in transcription factors cause synpolydactyly type II, cleidocranial dysplasia in one family and holoprosencephaly 5 [36]. Mitochondrial deoxyribonucleic acid (DNA) has 37 genes (22 transfer RNAs, 2 ribosomal RNAs and 13 respiratory chain proteins), but most genes for mitochondrial energy production are nuclear [36, 41].

2.6. Multifactorial inheritance

Common isolated malformations (cleft lip and palate, isolated cleft palate, neural tube defects, clubfoot, pyloric stenosis, hip dysplasia, congenital heart disease) are discontinuous traits that cluster in families; the parameter λ, the ratio between the incidence in relatives and the population incidence, measures the relative risk [42]. In Falconer's polygenic threshold model, liability has a continuous, normal distribution, those who exceed the threshold manifest the condition and close relatives have a shifted distribution; for 1st-degree relatives, the incidence approximates the square root of the population incidence, so that a condition of 1/1000 gives about 3% in siblings, parents and children, about 1% in 2nd-degree relatives and values close to the population risk in 3rd-degree relatives [42]. Smith lists the predictions of the model: familial clustering and ethnic variation; 1st-degree risk of 2–5% for most defects, 20 to 40 times the population frequency; a sharp fall in 2nd-degree relatives; an increase with the number of affected individuals; an increase with consanguinity; greater recurrence with greater severity (severe bilateral cleft lip and palate 5.7% versus 2.5% for the less severe defect); greater risk for relatives of a proband of the less often affected sex (in pyloric stenosis, transmission of 24% by affected mothers versus 6% by affected fathers); and concordance in dizygotic twins similar to that of siblings [29].

The sex ratios (male:female) of isolated malformations are 5:1 in pyloric stenosis, 2:1 in clubfoot and in cleft lip with or without cleft palate, 1:1.3 in isolated cleft palate, 1:1.5 in myelomeningocele, 1:3 in anencephaly and 1:5.5 in congenital dislocation of the hip; the explanations include structural genital differences, the prostatic urethra in urethral obstruction, the effect of testosterone and the faster growth of the male fetus in the 3rd trimester, which favours constraint deformations, except those linked to joint laxity [29]. In cleft lip with or without cleft palate, susceptibility appears to depend on 2 to 8 genes with a multiplicative effect and environmental interaction [35]. Genome-wide association studies have identified susceptibility factors with typical relative risks of 1.05–1.3 that, even combined, almost never give clinically useful prediction; part of the "missing heritability" is due to rare factors of strong effect, including microdeletions and microduplications [42]. Environmental influences include social class in neural tube defects and birth order (hip dislocation and pyloric stenosis more frequent in the firstborn) [35].

2.7. Teratogens, maternal disease and gene–environment interaction

Before considering genetic causes, the geneticist looks for evidence of abnormal physical forces in utero (deformation), vascular interruption (disruption) and teratogenic exposure, such as fetal alcohol spectrum disorder, fetal anticonvulsant syndrome and diabetic embryopathy [1]. The WHO lists as environmental factors maternal infections (syphilis, rubella, Zika), radiation, pollutants, nutritional deficiencies (iodine, folate), maternal disease (diabetes) and drugs and substances (alcohol, phenytoin); advanced maternal age increases chromosomal anomalies, and consanguinity increases the prevalence of rare genetic disorders and almost doubles the risk of neonatal and infant death, ID and other anomalies [3].

Pre-gestational diabetes is the model of teratogenic maternal disease. In the population-based cohort from the North of England (1996–2008; 401,149 singleton pregnancies, 1677 in women with type 1 or type 2 diabetes; live births, fetal deaths ≥20 weeks and terminations for fetal anomaly), the prevalence of non-chromosomal major congenital anomalies was 71.6/1000 pregnancies (95% CI 59.6–84.9) in diabetic women, with a relative risk of 3.8 (3.2–4.5) and a 3- to 6-fold increase in all common anomaly groups [43]. Periconceptional glycated haemoglobin (HbA1c) was associated with risk with an adjusted odds ratio of 1.3 (1.2–1.4) for each 1% (11 mmol/mol) above 6.3% (45 mmol/mol), and pre-existing nephropathy with an adjusted odds ratio of 2.5 (1.1–5.3); type and duration of diabetes, ethnicity, age, body mass index, preconception care and smoking were not associated [43]. According to the Oxford Desk Reference, the risk is 1/9 with an HbA1c of 10% versus 1/33 with 6.1%, does not differ significantly between type 1 and type 2 diabetes and is confined to diabetes present at conception, not gestational diabetes [44].

Even caudal dysplasia occurs, in absolute numbers, just as probably in the children of non-diabetic women, and VACTERL and sirenomelia have an increased relative risk but are also more frequent in the children of non-diabetic women; the mechanism is poorly understood (inositol, prostaglandins, reactive oxygen species) [44]. Susceptibility to teratogens is itself genetic: Fraser produced cleft palate with high-dose cortisone in practically all A/Jax mice versus 17% in another strain, and in humans susceptibility to hydantoin embryopathy correlates with epoxide hydrolase activity [35]. The most relevant specific teratogens are alcohol, hydantoin, valproate and retinoids.

Among drugs, warfarin embryopathy (chondrodysplasia punctata with nasal hypoplasia and stippled epiphyses) occurs in 6–30% of fetuses exposed between 6 and 12 weeks, and the risk of adverse outcome is 80% with a mean daily dose >5 mg versus 3–10% with <5 mg; fetopathy due to angiotensin-converting enzyme (ACE) inhibitors and angiotensin II receptor antagonists in the 2nd and 3rd trimesters causes renal tubular dysgenesis, anuria, hypocalvaria, pulmonary hypoplasia and growth restriction; mycophenolate mofetil, a purine synthesis inhibitor, causes cleft lip and/or palate, micrognathia, external auditory canal atresia with or without microtia, oesophageal atresia, hydronephrosis, heart defects and myelomeningocele, in a pattern similar to that of CHARGE syndrome; carbimazole or thiamazole in the 1st trimester are associated with embryopathy in 0.8–4.1% of those exposed (11.1% in a Hong Kong cohort), with aplasia cutis, choanal atresia, persistent omphalomesenteric duct and abdominal wall defects; continuous high-dose fluconazole (>400 mg/day) in the 1st trimester produces a phenocopy of Antley-Bixler syndrome, which is not seen with the single 150 mg dose; and systemic retinoids (isotretinoin, acitretin) cause central nervous system defects, microtia or anotia, conotruncal heart defects and thymic defects, with neurodevelopmental deficit in up to 60% of those exposed without structural malformation [45]. The diagnosis of a teratogenic syndrome requires exclusion of a genetic cause, especially when the proposed phenotype rests on few cases with limited genetic investigation (for example, allopurinol); the recurrence risk is high if the woman continues the same drug at the same dose, although severity varies between siblings, and teratology information services (UKTIS in the United Kingdom, ENTIS in Europe, MotherToBaby/OTIS in the USA) carry out individual risk assessment [45]. Exposure to ionising radiation below 100 mGy at any stage of pregnancy does not increase the risk of loss, microcephaly, microphthalmia, cataracts or growth restriction; UK recommendations set the limit at 50 mGy, the estimated fetal dose is 0.001–0.01 mGy for a chest radiograph and 10–50 mGy for abdominal or thoracic computed tomography (CT), and the additional absolute risk of childhood cancer is about 1 in 13,000 per mGy [46]. In the Cancer Survivor Study (4699 offspring of cancer survivors), the prevalence of congenital anomalies in the children of mothers exposed to radiation or alkylating agents was 3.0%, versus 3.5% in the children of unexposed mothers [46].

2.8. Mechanical and vascular mechanisms

Deformations result from fetal constraint or immobility, and many minor anomalies of the auricle at birth are constraint deformations [26]. Redundant skin due to tangential traction from external constraint is distinguished from tight, thin skin over joints immobilised early by a neurological cause, which makes it possible to separate constraint clubfoot from neurogenic clubfoot [26]. Disruptions result from destruction of normal tissue: amniotic bands, vascular interruption, death of a co-twin [30, 47]. In limb reduction defects, 1st-trimester bleeding was associated with a 15-fold increase (Matsunaga and Shiota, 1979), and different aetiologies can produce indistinguishable pictures, such as TBX5 in Holt-Oram syndrome and thalidomide, both with limb and heart defects [34].

2.9. Twinning

Monozygotic (MZ) twinning is considered by Smith to be the most common aberration of morphogenesis, at about 1/200 births; the Oxford Desk Reference gives 3–4/1000 births in the United Kingdom, with the same rate worldwide, and the difference probably reflects different denominators and periods [47, 48]. Twins account for about 1/80 live births, with an increase linked to maternal age and in vitro fertilisation (IVF); MZ twins are about 4 times more frequent after IVF, with 5.6% MZ twins after blastocyst transfer versus 2% at day 3 (Milki 2003), and only 29% of twin pregnancies seen before 10 weeks end in a twin delivery [48]. Dizygotic twinning varies with the population: 4/1000 in Asians, 10–14/1000 in Europeans and North Americans and 26–40/1000 in Africans, peaking at 37 years [48].

Congenital anomalies occur in at least 10% of twin pregnancies, and structural defects are 2 to 3 times more frequent in liveborn MZ twins than in dizygotic twins or singletons; among spontaneous miscarriages, the MZ:DZ ratio is 17:1 versus 0.8:1 in live births (Livingston and Poland, 1980) [47, 48]. The table presents the origin of defects in MZ twins and placentation according to the timing of division.

Source of the placentation table: Oxford Desk Reference, Table 6.18, with Smith's divergent estimate for monoamniotic twins [47, 48].

In the artery-to-artery disruption sequence (TRAP), the arterial pressure of one twin overrides that of the other, who receives reversed flow of blood already used, through the iliac vessels, and preferentially perfuses the lower half, with variable absence of head, heart, upper limbs, lungs, pancreas and upper intestine; the donor may have cardiomegaly, decompensation and hydrops [47]. In artery-to-vein transfusion, the recipient is larger, polycythaemic and hypervolaemic, with polyhydramnios and enlarged kidneys and heart, and the donor is smaller and anaemic, with oligohydramnios and transient renal failure; 18% of MZ twins were discordant in size and haematocrit at birth (Tan, 1979) [47]. In utero death of a co-twin causes hypoperfusion and hypoxia in the survivor, with disseminated intravascular coagulation, aplasia cutis, porencephalic cyst to hydranencephaly, limb amputation, intestinal atresia and gastroschisis; in the Collaborative Perinatal Project (50,000 deliveries), about 3% of MZ twins near term had a dead co-twin and about one-third of survivors had severe brain injury [47]. Discordant MZ twins may be due to a new dominant variant, aneuploidy, CNV, differential X inactivation, imprinting (KCNQ1OT1 in MZ twins discordant for Beckwith-Wiedemann syndrome), mitotic recombination or mitochondrial load, which is relevant to prenatal diagnosis; at least 8% of dizygotic twins have lymphocyte chimerism [48].

2.10. Relative contribution of aetiologies

On Smith's scale, chromosomal anomalies affect fewer than 1% of births, major mutant genes more than 6000 individually rare disorders, and multifactorial inheritance about 10% (reading of Figure 2.1, on a logarithmic axis) [12]. For congenital anomalies in general, the cause is unknown in about 80% of cases, which are considered multifactorial [8]. The probability of a genetic cause depends on the type of anomaly: rare in isolated ventricular septal defect (VSD) and much higher in anophthalmia, omphalocele or MCA [8]. The distribution by subgroup in EUROCAT is given in section 3.2.

3. Clinical Spectrum and Manifestations by System

3.1. Patterns of association between systems

In the EUROCAT study of 19 registries (959,446 births from 2004 and 2010), MCA without a recognised syndrome had a prevalence of 15.8/10,000 births; respiratory anomalies and those of the ear, face and neck were the most often associated with other anomalies (34% and 32% of the cases in each group), and heart defects and limb anomalies the least often (13%), although heart defects were present in half of the MCA cases because of their high prevalence; genital anomalies were more frequent in boys with MCA [16]. Compared with isolated anomalies, MCA had more fetal deaths and terminations of pregnancy (p<0.001) and more prenatal diagnoses, and liveborn infants with MCA were more often preterm and weighed less than 2500 g [16]. Among infants with Down syndrome in the USA, 17% have an atrioventricular septal defect (AVSD) and 3% tetralogy of Fallot; 6% of infants with trisomy 18 and 11% of those with trisomy 13 have an omphalocele [32].

3.2. Probability of a genetic aetiology according to the index anomaly

The EUROCAT study by Bergman and colleagues included 20 registries from 14 countries, births from 2013 to 2022 and all pregnancy outcomes (100,099 cases), and measured the proportion with a recorded genetic diagnosis (monogenic disorders, including genetic syndromes, genodermatoses and skeletal dysplasias, and chromosomal anomalies) [8]. The table presents the subgroups with 50 or more cases; the subgroups are not mutually exclusive and the percentage measures the recorded diagnosis, not the fraction of cases tested [8].

The presence of additional anomalies increases the probability of a genetic diagnosis: in VSD, 10% of all cases had a genetic diagnosis versus 5% in isolated VSD (including trisomies), and 4% versus 1% excluding trisomies [8]. The overall proportion was 20% (95% CI 20–21%) and, excluding trisomies, 10% (95% CI 9–10%), with an annual increase of 1.4% (0.8–1.9%) and 1.2% (0.4–2.0%), respectively; the trend was upward for non-trisomic diagnoses in hydrocephalus, severe heart defects (10.3% in 2022), hypoplastic left heart syndrome (13.4%), cleft lip and oesophageal atresia [8]. In the Netherlands, among euploid fetuses with a severe heart defect born in 2012–2016, there was a definitive genetic diagnosis in 13% (94/708), most often the 22q11.2 deletion [8].

3.3. Prenatal and perinatal natural history

About 3% of pregnancies show a fetal structural anomaly on ultrasound [49]. In the British PAGE study, of the 474 fetuses with structural anomalies and known outcome, 30% ended in termination, 3% in miscarriage, 5% in stillbirth, 3% in neonatal death and 59% in live birth; diagnosis by exome was more frequent in fetuses that did not survive to delivery (27/192, 14.1%) than in live births (20/282, 7.1%; p=0.018) [49]. In Portugal, among the cases notified to RENAC in 2020–2021, the pregnancy ended in a live birth in 66.1%, in termination of pregnancy for medical reasons in 31.8%, in fetal death in 1.3% and in spontaneous miscarriage in 0.7% [18]. These figures describe cohorts with different selection and legal frameworks and are not directly comparable.

3.4. Exemplary sequences and complexes

In the oligohydramnios (Potter) sequence, renal agenesis (before 31 days of development) or another urinary anomaly (polycystic kidneys, obstruction) causes oligohydramnios, fetal compression and pulmonary hypoplasia, the cause of death through lack of late alveolar development, with Potter facies, abnormal limb positions, large flattened ears (in more than 97%) and redundant skin [31]. In chronic fluid leakage with arthrogryposis, pulmonary hypoplasia occurs in only 50%, the contractures respond to physiotherapy and, in the absence of a central nervous system deficit, the prognosis is good [31].

In the amnion rupture sequence, described by Portal in 1685 and delineated by Torpin in 1965, amniotic bands encircle structures, mainly limbs, with annular constrictions, pseudosyndactyly, intrauterine amputations and cord constriction; secondary deformations arise from tethering of the limb or from fluid loss (scoliosis, foot deformities), oedema, haemorrhage and necrosis [30]. Internal organ defects almost never occur, no two fetuses are alike and there is no constant feature, and examination of the placenta and membranes is diagnostic, with bands or remnants of amnion wrapped around the base of the cord; the most common form is a normal child with constrictions or amputations, Z-plasty treats deep constrictive grooves with vascular or lymphatic compromise, and chronic fluid leakage may lead to an oligohydramnios sequence with pulmonary hypoplasia, with a prognosis that can be excellent under ventilatory support [30]. Rupture probably occurs before 12 weeks, the aetiology is idiopathic except for rare trauma and recurrence is usually negligible [30].

The limb–body wall complex (LBWC) is distinguished from amnion rupture because the amnion is intact in some cases: thoraco- or abdominoschisis with evisceration into a persistent extra-embryonic coelom, short cord, limb defects (single bone of the forearm or leg, ectrodactyly, radio-ulnar synostosis, polydactyly and, occasionally, amputations), anterior encephaloceles, often multiple, and atypical facial clefts, with a high incidence of internal cardiac, pulmonary, diaphragmatic, intestinal, renal and genitourinary anomalies [50]. Most are spontaneously aborted and the rest are stillborn; the pathogenesis is controversial (embryonic vascular alteration at 4–6 weeks or defect of the embryonic disc), and the Spanish epidemiological studies of the ECEMC (Spanish Collaborative Study of Congenital Malformations) support their being entities distinct from amnion rupture [50].

Robin sequence and holoprosencephaly illustrate sequences that may be isolated or syndromic; in EUROCAT, 14% of Pierre Robin sequences and 45% of holoprosencephalies had a genetic diagnosis [1, 8].

3.5. Recognisable subtelomeric and interstitial deletion syndromes

Chromosomal deletions with a recognisable clinical phenotype illustrate the correlation between gene content and dysmorphic picture [20].

Source: Oxford Desk Reference, pp. 661–662 [20].

3.6. Composite phenotype due to two diagnoses

About 5% of children with a developmental disorder that is difficult to diagnose have a composite phenotype due to two distinct genetic disorders (DDD 2015) [13]. In the final DDD analysis, 3% (121) of probands had two or more diagnoses (7% counting computational predictions) and 13% were partially diagnosed [10]; in Clark's meta-analysis, about 4% of children had more than one genetic diagnosis [11]. A diagnosis that explains only part of the picture should not close the investigation.

3.7. Mortality and morbidity

The WHO estimates 240,000 deaths of newborns each year in the first 28 days due to congenital disorders and a further 170,000 deaths between 1 month and 5 years; the most common severe congenital disorders are heart defects, neural tube defects and Down syndrome, and as global mortality falls they become a growing proportion of neonatal and under-5 deaths [3]. In Brazil, congenital malformations are the second cause of infant mortality, with 11.2% of those deaths [51]. In Portugal, 6.9% (2020) and 5.7% (2021) of perinatal deaths (stillbirths at 28 or more weeks and live births dying by day 7) had at least one congenital anomaly notified to RENAC [18].

4. Clinical Recognition and Red Flags

4.1. When to suspect multiple anomalies or a syndrome

"Dysmorphic" describes the child whose physical features, especially facial ones, are not usually found in a child of the same age or ethnic background [13]. Some findings are abnormal in any circumstances, such as premature fusion of the cranial sutures, and act as good diagnostic "handles", present in a small number of conditions rather than as familial variants; others, such as 2/3 toe syndactyly, may be useful or may correspond to a familial trait without significance [13]. The further the growth parameters are from normal, the greater the probability of a genetic diagnosis; recording standard deviations (for example, head circumference at −5.5 SD) is more informative than "below the 0.4th centile" [13].

The number of minor anomalies predicts the presence of a major anomaly. The table summarises the data of Marden, Smith and McDonald (1964), according to Smith [25, 26].

Later studies with a different design found a major malformation in 26% (Méhes) and 19.6% (Leppig and colleagues) of newborns with three or more minor anomalies; any newborn with three or more minor anomalies should be assessed for a major malformation, which is often occult [26]. Before attributing significance to a minor anomaly, it is checked whether it is present in other relatives (for example, a chin dimple in father and daughter) [26].

In ID and global developmental delay, history and examination establish the aetiology in 17.2–34.2% of cases; the genetic yield is about 80% in severe ID versus about 20% in mild ID, and is higher when there are dysmorphic features, congenital anomalies or neurological comorbidity [9]. In the DDD, each additional organ system affected increased the probability of diagnosis (odds ratio 1.08; 95% CI 1.06–1.11), as did features suggestive of a syndrome (1.23), severe ID or delay (2.41) and being the only affected person in the family (1.74); prematurity (0.73), in utero exposure to antiepileptic drugs (0.44) and maternal diabetes (0.52) decreased it [10].

4.2. Frequent minor anomalies and what they mean

Other regional signs have specific value: prominent lateral palatine ridges reflect prolonged lack of tongue pressure due to hypotonia or a neurological deficit of sucking; absence of the lingual frenulum with a single central incisor suggests holoprosencephaly; low-set ears (helix below the plane passing through the inner canthi) and posteriorly rotated ears (more than 15° from the vertical) reflect delayed morphogenesis, but cranial deformation due to constraint distorts the landmarks; a shawl scrotum reflects incomplete migration of the labioscrotal folds (Aarskog syndrome); and a median raphe is never seen in 46,XX fetuses without abnormal androgen exposure [26].

4.3. Dysmorphological history

The history includes a three-generation pedigree, with specific questions about consanguinity, affected relatives, pregnancy losses, stillbirths, neonatal deaths and relatives with learning difficulties or congenital anomalies; the pregnancy (conception, bleeding, fever, medication, investigations, alcohol and drugs asked about tactfully, fetal movements, amniotic fluid volume, gestational age and mode of delivery); the neonatal period (weight, length and head circumference at birth, resuscitation, feeding difficulties, ventilation, malformations, surgery, seizures); developmental milestones and schooling; vision and hearing; past medical and surgical history; and the behavioural phenotype [1, 13, 20]. Paternal age is always recorded, because new dominant variants are more likely with advanced paternal age, which has been demonstrated in at least 12 autosomal dominant multiple malformation syndromes [36].

The consultation is prepared with the notes, radiographs and contact with the paediatrician; who accompanies the child is confirmed without presuming parentage; and, even when the diagnosis seems obvious at first sight, it is confirmed by history and examination [1]. The steps follow the order presentation and concerns, observation, history, examination with clinical photographs (face, profile, hands, unusual features and photograph with the parents), investigations, conclusions with risk and counselling, correspondence and follow-up [1]. Old photographs and serial measurements allow the natural history to be seen, because syndromic features change with age [1, 13].

4.4. Systematic dysmorphological examination

Source: checklist from the Oxford Desk Reference, Table A10, which recommends free text, drawings and clinical photographs [52]. Video of movement is useful because many syndromes, especially microdeletion syndromes, have a peculiar facial expression in movement [13].

The description uses the standardised terminology of the Elements of Morphology project: each feature has a preferred term, a subjective and an objective definition (the objective one is always preferable), synonyms and terms to be abandoned; the terms are descriptive, carry no pathogenetic implication, avoid lumped terms (for example, "large nose" is broken down into prominent nose, wide nasal bridge, prominent nasal tip and wide nasal base) and do not judge normal or abnormal, because many isolated findings exist in the normal population and frequency varies between populations [28]. Anthropometric norms include Farkas (1981) and the Handbook of Physical Measurements (2nd edition, 2007), and normative data are lacking for non-Caucasian groups [28].

4.5. Facial analysis tools

In an Indian cohort of 159 photographs of 72 molecularly diagnosed syndromes, top-3 accuracy was 77% for Face2Gene and 58% for GestaltMatcher, and top-10 accuracy 85% and 71%; GestaltMatcher identified in the top 10 two of seven syndromes without a model in Face2Gene, and performance was better in rare than in ultra-rare syndromes [53]. The AAP considers that the clinical utility of facial recognition software is not yet established [9].

4.6. Emergency signs

Upper airway obstruction in Robin sequence, pulmonary hypoplasia in the oligohydramnios sequence, hydrops fetalis, severe hypotonia with feeding difficulty and neonatal seizures require stabilisation before aetiological investigation [1, 31, 54]. In infants admitted to intensive care with suspected genetic disease, rapid (6–15 days) or ultra-rapid (1–3 days) exome or genome sequencing should be available when acute management, eligibility for targeted therapy or decisions about goals of care justify it [21]. In Seattle SeqFirst-neo, the classic selection criteria would have missed 42% (26/62) of the diagnoses obtained with rapid genome sequencing [55].

4.7. Neonatal and family context

In a child of consanguineous parents, the result of newborn screening (for example phenylketonuria and medium-chain acyl-CoA dehydrogenase (MCAD) deficiency) is confirmed [38]. In a microcephalic child without normal siblings, maternal phenylketonuria is investigated [13]. There is no population newborn screening for multiple congenital anomalies as a group; examination of the newborn, with the count of minor anomalies, is the detection tool [26].

5. Diagnosis: Criteria, Tests and Algorithms

5.1. Diagnostic framework

There are no formal criteria for "multiple congenital anomalies"; diagnosis is a process of classification. The Oxford Desk Reference proposes starting with basic questions (single or multiple malformations, probable multiple anomaly syndrome, deformations related to the history of the pregnancy, contribution of the family history), considering the mechanisms (chromosomal, monogenic including alterations of expression and imprinting, effect of multiple variants as in Hirschsprung disease, multifactorial, environmental with possible genetic predisposition, and mosaicism), recognising a "gestalt", searching the literature and databases for distinctive features, such as a median cleft lip, rather than for common or subjective features, such as "low-set ears", and sharing photographs with colleagues [13]. Smith's index of differential diagnosis by anomaly (26 categories, from the central nervous system to growth patterns) and databases such as the London Dysmorphology Database and POSSUM serve the search for distinctive handles [13, 28, 56].

5.2. Available tests

The exome may be complete (about 20,000 genes), clinical (about 4000–5000 OMIM genes) or targeted to fetal and neonatal phenotypes (about 1300–1600 genes, with fewer VUS); panels analyse from a few to hundreds of genes and do not look for secondary findings [23]. Long-read sequencing is beginning to be evaluated in the acute setting: in Thailand, rapid singleton long-read genome sequencing gave a diagnosis in 61% (11/18) of children in intensive care, with a median time of 9 days, three pathogenic structural variants, phasing in six recessive patients and reclassification of three VUS as likely pathogenic, in a proof of concept [65]. Trio exome sequencing is best practice, because it places rare variants in context (de novo, phase), and singleton or duo testing has a slightly lower yield [21].

5.3. Guidelines and positions

The main divergence is one of order rather than content: the ACMG accepts exome or genome as second tier after CMA, the AAP places them in tier 1 alongside CMA, and the 2019 consensus proposes the exome before CMA; all reserve targeted testing for a specific clinical suspicion [9, 21, 59]. The AAP recommends that fragile X testing (yield of about 1% in both sexes), which exome, genome and CMA do not detect, move to tier 2 in non-syndromic cases and be brought forward when there is clinical suspicion, whereas the 2017 Oxford Desk Reference considered it in all children with delay [9, 37]. Metabolic screening without specific signs has a very low yield (Curry consensus, 1997); the AAP attributes 1–5% to treatable inborn errors and brings it forward in the presence of regression, neurological signs, multisystem involvement, abnormal odour, acidosis or hypoglycaemia, abnormal magnetic resonance imaging or consanguinity [9, 13].

5.4. Diagnostic yield

In Clark's meta-analysis, the odds of diagnosis by exome or genome increased by 16% per year between 2013 and 2017, while that of CMA decreased by 14% per year, and in-hospital interpretation with a detailed phenotype gave 0.42 versus 0.29 in a reference laboratory [11]. In the DDD, the probability of diagnosis ranged from 52% in the top decile to 16% in the bottom one, and the yield was lower in African ancestry (OR 0.51), partly because of the lack of matched allele frequency controls [10]. These values derive from cohorts with different selection and are not directly comparable.

5.5. Limitations of the tests

A negative result does not exclude a genetic cause. CMA does not detect truly balanced rearrangements or low-level mosaicism, although these causes are infrequent (less than 1%) in the population with DD/ID or MCA [22]; arrays do not detect mosaicism below about 20%, which may require karyotyping or FISH in several tissues with counting of many cells [36]. The array may be normal in a parent carrying a balanced rearrangement, who is studied by karyotype, FISH or genome [20]. True uniparental disomy is detected by SNP array in fewer than 0.5% of cases, and only isodisomy appears as homozygosity [37]. The exome does not typically detect intronic variants beyond those adjacent to exons, repeat expansions or methylation abnormalities, has limited CNV detection and reduced sensitivity for intragenic deletions of fewer than 3 exons, and CNV analysis from the exome is not standard in all laboratories [9, 21, 59]. In three studies with CNV calling from the exome, two pathogenic CNV were not detected by CMA because of insufficient coverage, which shows that the two methods complement each other [59]. In the DDD, the combination of exome and array did not analyse most non-coding variants, complex structural variants or tissue mosaicism [10]. In 56 patients with a previous non-diagnostic exome studied by genome, 29 could have been resolved by reanalysis and 27 were missed because of a technical limitation of the exome [61]. Mosaicism may require fibroblasts or another affected tissue; skin biopsy for karyotyping is indicated in the presence of streaky pigmentation along Blaschko's lines (hypomelanosis of Ito), asymmetry, or 3/4 finger syndactyly with 2/3 toe syndactyly and bulbous fingertips (diploid/triploid mosaicism) [9, 13]. Skewed X inactivation in healthy mothers of affected boys is seen in only about 50% of X-linked ID and may be due to chance [37]. The proportion of variants detectable by a method is not equivalent to the clinical sensitivity of the test, which depends on the genetic architecture of each phenotype.

5.6. Interpretation, reporting and variants of uncertain significance

Each genome has 4–5 million variants, including rare functional variants without phenotypic consequence; NGS increases the yield but also the opportunity to attribute causality falsely, because the prior probability per gene is much lower than in targeted testing [39]. Of 406 published "severe mutations" observed in 104 sequenced individuals, 122 (27%) were common polymorphisms or had no direct evidence of pathogenicity (Bell 2011) [39]. The MacArthur guidelines (2014) require presenting positive and negative evidence in a quantitative framework, first assessing genes already implicated in similar phenotypes, declaring a new gene only with variants in the same gene and a similar picture in several unrelated individuals, not taking in silico predictors as definitive or independent evidence, not presuming complete penetrance, and communicating uncertain results clearly [39].

Classification follows the five ACMG/AMP classes (pathogenic, likely pathogenic, uncertain significance, likely benign, benign) for sequence variants and the ACMG/ClinGen scoring system for CNV, which separates the classification of the variant from the implications for an individual [58, 67]. A likely pathogenic variant has about a 90% probability of being causal, and pathogenic variants may have incomplete penetrance or explain only part of the phenotype [59]. The report should state the gene, the reference transcript with version (preferably MANE Select), the HGVS nomenclature at DNA and protein level, zygosity, inheritance (de novo, maternal, paternal), the ACMG/AMP class and the relationship to the phenotype; phenotyping coded in HPO improves interpretation [23, 54].

For CNV, the algorithm adapted from Miller and colleagues distinguishes three results [20]. A normal array or one with known benign CNV leads to further assessment (fragile X, single genes, panels, exome). A VUS leads to consultation of DECIPHER (gene content, phenotype of other patients, typical inheritance) and to parental studies: de novo is potentially causal; inherited from an unaffected parent is probably familial, with attention to incomplete penetrance, unmasking of a recessive allele, imprinting, amplification of a parental duplication and complex inheritance. An abnormal result leads to confirmation in the proband if necessary (FISH, G-banding, qPCR, MLPA), definition of the mechanism and study of the parents: de novo implies a low recurrence risk, and inherited unbalanced or from a balanced carrier implies an increased risk [20]. Parental studies are essential, especially in subtelomeric rearrangements, to detect a cryptic balanced translocation in a parent [20]. Visible imbalances with hundreds of genes are almost always causal, whereas smaller imbalances are often normal variation [12].

The PIGN case in section 10 illustrates a VUS with compatible segregation: a biallelic exon duplication, absent from population databases, was classified as a VUS because of the lack of functional validation, despite the phenotype being concordant [54]. In the DDD, about 22% of probands had VUS in genes robustly associated with developmental disorders [10]. A VUS should not guide clinical or reproductive decisions without clinical reassessment and, when possible, segregation, functional studies or a methylation episignature [9, 59].

5.7. Complementary non-genetic investigations

Neuroimaging is not routine with a normal head circumference and no neurological signs; it is considered when there is no known diagnosis, especially with neurological symptoms, abnormal skull contour, micro- or macrocephaly, and magnetic resonance imaging is the modality of choice (computed tomography if craniosynostosis or congenital infection) [13, 37]. An electroencephalogram is performed when seizures or regression are suspected, creatine kinase in small boys with delay (about one-third of patients with Duchenne dystrophy present with developmental delay), and system investigations assess specific findings, such as an echocardiogram for a murmur [13, 37]. In ID with dysmorphism, targeted metabolic investigation includes maternal phenylalanine, sterols (Smith-Lemli-Opitz), peroxisomal studies and transferrin isoelectric focusing [37]. Diabetic embryopathy is investigated with periconceptional HbA1c, other exposures, echocardiography and genomic array if there are multiple anomalies [44]. In twins, zygosity is determined by DNA with at least 7 microsatellites, and the placenta is always examined (membranes, anastomoses, dead twin) [47, 48].

5.8. Prenatal diagnosis

The dating scan is performed at 8–12 weeks, nuchal translucency (NT) at 11–13+6 weeks (3 mm corresponds to the 95th centile, and many fetuses with slightly increased NT are normal) and the routine anomaly scan at 18–22 weeks; specialist fetal echocardiography is indicated in the presence of suspicion or increased risk [19]. Ultrasound detects many but not all major anomalies: isolated cleft palate and atrial septal defect are difficult, and hydrocephalus, microcephaly, achondroplasia and some renal diseases have late signs [19]. In the Eurofetus study (unselected population with routine ultrasound; 3685 fetuses with structural anomalies), sensitivity was 61.4%, 73.7% for major anomalies and 45.7% for minor ones, with 88.3% in the central nervous system, 84.8% in the urinary tract and 38.8% in the heart and great vessels [73]. In maternal diabetes, ultrasound detection of anomalies was 30% versus 73% in non-diabetic women [44]. Fetal magnetic resonance imaging helps to delineate intracranial anomalies, and the MERIDIAN study showed better diagnostic accuracy and confidence in fetal brain anomalies [19].

Source: Oxford Desk Reference, Table 6.7, after Pathak and Lees (2009) [19].

A few isolated anomalies on a detailed scan do not justify chromosomal study (gastroschisis, unilateral multicystic dysplastic kidney, clubfoot); in the others, the decision is individual according to other anomalies, the detail of the scan and risk factors [19]. In a systematic review, prenatal CMA with a normal karyotype detected clinically significant CNV in 2.4% (295/12,362) overall, 6.5% (201/3090) with an abnormal scan, 1.0% for advanced maternal age and 1.1% for other indications [74]. QF-PCR plus karyotype or CMA diagnose up to 40% of fetal anomalies [60].

The incremental yield of prenatal exome sequencing after a normal karyotype or CMA was 31% (95% CI 26–36%) in the Mellis meta-analysis (66 studies, 4350 fetuses), 42% in cases preselected for a probable monogenic aetiology versus 15% in unselected cases (p<0.0001) [60]. The two large unselected prospective studies gave 8.5% (PAGE, 52/610; 95% CI 6.4–11.0) and 10% (Columbia, 24/234 trios) [49, 75]. In PAGE, the yield was 15.4% (22/143) in multisystem anomalies, 15.4% in skeletal anomalies, 11.1% in cardiac anomalies and 9.0% in hydrops; 61.5% of diagnoses were de novo, 36.5% inherited (14 autosomal recessive, 5 dominant) and 1 was a disomy 15; no diagnosis would have led to in utero treatment, and postnatal follow-up or autopsy were concordant in 90.4% [49].

Source: ISPD 2022, Table 2 [23].

The 2022 position statement of the International Society for Prenatal Diagnosis (ISPD), which replaces the 2018 joint statement, considers prenatal sequencing beneficial in a fetus with a single major anomaly or anomalies of several systems without a diagnosis after CMA and with a phenotype suggestive of a genetic aetiology according to a clinical geneticist, or whose pattern strongly suggests a monogenic disorder, with CMA beforehand or in parallel because prenatal exome sequencing is not validated for all CNV; and in recurrence of anomalies similar to those of a previous undiagnosed fetus or child [23]. There is no evidence for routine sequencing of fetal tissue without anomalies, nor at the parents' request [23]. The 2024 Danish guideline indicates prenatal genome sequencing for NT ≥6.0 mm, multiple anomalies, skeletal dysplasia, neuromuscular disorders, non-immune hydrops, central nervous system malformations, congenital diaphragmatic hernia and severe fetal growth restriction; in Danish practice, CMA precedes the genome for reasons of timing (legal termination up to 23+0 weeks, CMA within 2 weeks, genome about 2 weeks), with a transition to direct genome sequencing [64]. The Society for Maternal-Fetal Medicine (SMFM, Consult Series #75, March 2026) recommends diagnostic fetal testing with CMA ± karyotype in all pregnancies with at least one fetal effusion and exome or genome sequencing in non-immune hydrops after a non-diagnostic CMA, or simultaneously when the risk of aneuploidy is low or a monogenic disorder is strongly suspected (GRADE 1C, Grading of Recommendations, Assessment, Development and Evaluation) [76]. "Isolated" omphalocele is associated with Beckwith-Wiedemann syndrome in 5–25%, which requires methylation testing, not detected by the exome [64].

5.9. Practical algorithm

1. Stabilise and document: major and minor anomalies, anthropometry in standard deviations, photographs, placenta in newborns and twins [13, 47].
2. Classify each anomaly by mechanism (malformation, deformation, disruption, dysplasia) and look for sequences; exclude deformations explainable by the pregnancy and disruptions with an evident cause [1, 13].
3. Review exposures (pre-gestational diabetes, alcohol, antiepileptic drugs, retinoids, infections) without closing the genetic investigation, because exposure does not exclude a genetic cause [1, 10].
4. Clinical suspicion of a specific diagnosis: targeted test (karyotype if trisomy is suspected, FISH, methylation, fragile X) [21, 36].
5. No evident diagnosis: trio exome or genome with CMA (or genome with CNV calling), on a rapid basis in severely ill infants [9, 21].
6. Positive result: confirm, study the parents, counsel and define targeted surveillance; VUS result: segregation, phenotypic reassessment, functional tests or episignature; negative result: targeted 3rd-tier tests, genome if only exome was done, skin biopsy if mosaicism is suspected, periodic reanalysis and clinical follow-up [9, 13, 59].

6. Differential Diagnosis

6.1. Comparative table by aetiological category

6.2. Coexisting genetic and non-genetic causes

The lower probability of a genetic diagnosis in the DDD with prematurity, maternal diabetes and antiepileptic drug exposure suggests a greater role for environmental factors in those patients, but does not exclude a genetic cause [10]. Valproate embryopathy has its own ERN-ITHACA guideline (2018); the most relevant teratogens are alcohol, hydantoin, valproate and retinoids [24]. Advanced maternal age increases chromosomal anomalies, and in the USA mothers aged 40 or more had the highest prevalence for almost all defects, especially chromosomal ones [3, 15].

6.3. Child of consanguineous parents

The incidence of autosomal recessive disorders is increased, but non-genetic, chromosomal, dominant and X-linked conditions also occur, and about 5% have a dominant de novo variant [38]. In communities with a long tradition of consanguinity, the risk may be higher than the three-generation pedigree suggests [38]. Recessive inheritance is suggested by the same phenotype in two siblings of different sexes or in two sisters, regression, similarity to a known recessive condition and severe ID without environmental factors; with malformations and a non-progressive course, chromosomal causes and non-recessive syndromes remain in the differential [38]. The SNP array identifies regions of homozygosity that guide the choice of candidate genes, and chromosomal instability studies detect Fanconi anaemia, ataxia-telangiectasia, Bloom and Roberts syndromes; the exome or genome covers the metabolic and recessive hypotheses simultaneously [38]. In Clark's meta-analysis, each 10% increase in consanguinity reduced the odds of a de novo diagnosis by 21%, and in one study the exome gave 35% in 453 consanguineous families versus 27% in 443 non-consanguineous families [11].

6.4. Isolated anomalies that do not justify extensive genetic investigation

Prenatally, isolated gastroschisis, unilateral multicystic dysplastic kidney and clubfoot on a detailed scan do not justify chromosomal study [19]. In EUROCAT, the genetic proportion was 3% in gastroschisis, 2% in hypospadias and hip dislocation and 1% in isolated VSD excluding trisomies [8]. The decision still depends on the complete examination, because an "isolated" anomaly ceases to be so when additional anomalies are found.

6.5. Subtelomeric deletions and overlapping presentations

In the neonatal period, Kleefstra syndrome overlaps with Down syndrome, visible del(15q) without IGF1R gives a phenotype similar to that of Silver-Russell syndrome, and the SHH deletion and del(18p) should be considered in holoprosencephaly [20]. The list in section 3.5 summarises the recognisable deletions.

7. Treatments and Risk Management Strategies

7.1. Principles of management

There is no treatment for multiple congenital anomalies as a group; management is that of each defect and of the identified aetiology. The WHO stresses that surgery with appropriate follow-up reduces the lethality of defects such as heart defects and the morbidity of others, such as clubfoot and clefts, and that children need long-term support through physiotherapy, speech therapy and occupational therapy [3]. The prognosis of an isolated multifactorial defect depends on the possibility of surgical correction, or postural correction in constraint deformations: it is poor in certain neural tube defects and good in many heart defects [78]. The contractures of the oligohydramnios sequence respond to physiotherapy, Z-plasty treats deep constrictive grooves of amnion rupture, and ventilatory support can give a good prognosis in pulmonary hypoplasia due to chronic fluid leakage [30, 31]. Surgical, medical and rehabilitation management of each system follows the specific recommendations for congenital heart disease, congenital anomalies of the kidney and urinary tract, VACTERL association, orofacial clefts, Robin sequence, neural tube defects and Hirschsprung disease.

Prognostic assessment and therapeutic decisions are made by a multidisciplinary team that includes neonatology or paediatrics, clinical and laboratory genetics, paediatric surgery, cardiology, neurology, rehabilitation and, in the prenatal period, maternal-fetal medicine [23, 51, 75]. Close interaction between clinical and laboratory geneticists improves the yield: in-hospital interpretation with a detailed phenotype from the clinical record had a diagnostic utility of 0.42 versus 0.29 in a reference laboratory, and in a dual interpretation of 115 children the hospital interpretation gave 0.43 with no false positives versus 0.33 with 0.03 false positives [11, 37].

7.2. Impact of the aetiological diagnosis on management

With a diagnosis, counselling is usually accurate, prognosis and natural history can be discussed, surveillance is targeted, prenatal diagnosis may be possible and the family gains access to support groups [1]. The AAP lists the benefits of a genetic diagnosis: prognosis, surveillance, recurrence risk, prenatal diagnosis and preimplantation genetic testing (PGT), specific treatment, access to trials and an end to unnecessary testing [9]. In the ACMG meta-analysis (all patients tested, with or without a diagnosis), exome or genome sequencing changed short-term active management in 8% (95% CI 6–11%), long-term management or surveillance in 10% (7–15%), reproductive outcomes in 9% (4–21%) and family outcomes, such as cascade testing, in 4% (2–9%); only 2 studies reported a reduction in mortality or morbidity or an improvement in quality of life [21, 69].

In Clark's meta-analysis, clinical utility, defined as a change in medical or surgical management without counting counselling alone, was 0.27 (95% CI 0.17–0.40) for genome, 0.17 (0.12–0.24) for exome and 0.06 (0.05–0.07) for CMA [11]. Among those diagnosed by exome in neurodevelopmental disorders, medical management or medication changed in 30% (2–46%) and the diagnosis was informative for reproductive planning in 80% [59]. Concrete examples from the Korean cohort include anticoagulation in antithrombin III deficiency and intraocular pressure surveillance in Axenfeld-Rieger syndrome type 3 [17].

These studies have limitations that prevent direct comparisons: only NSIGHT1 and NSIGHT2 are randomised trials, NSIGHT1 was stopped because of loss of equipoise (73% of controls underwent sequencing as a standard test), the certainty of evidence for clinical utility was rated by the ACMG as very low, and change in management includes palliative care decisions that are not a survival benefit [21, 69, 79]. Populations (infants in intensive care, mixed paediatric cohorts, MCA after conventional testing) and denominators (all those tested or only those diagnosed) differ between studies. The sources consulted did not identify negative trials of sequencing in MCA; a negative search does not prove that none exist.

7.3. Palliative care and decisions about goals of care

The diagnosis of a lethal condition or one with a severe prognosis may lead to a transition to palliative care, as in 36% of the changes in management in the Meng cohort and in 4% of the British French cohort [69]. The ACMG considers that rapid exome or genome sequencing should be available when the result may alter decisions about goals of care [21]. Assessment of the parents' readiness to receive the diagnosis is part of the consultation [1].

7.4. Primary prevention

The WHO recommends a healthy diet and appropriate weight, folic acid (fortification or supplementation) and iodine, avoidance of alcohol and tobacco, avoidance of travel to areas with outbreaks of teratogenic infections, reduced exposure to heavy metals and pesticides, control of diabetes before and during pregnancy (including insulin), justification of any exposure to drugs or radiation, rubella vaccination and screening for rubella, varicella and syphilis [3]. In pre-gestational diabetes, glycaemic control is optimised before conception and the pregnancy is followed in maternal-fetal medicine, because the risk increases linearly above an HbA1c of 6.3% (45 mmol/mol) [43, 44]. The periconceptional folic acid dose depends on the risk of neural tube defect. Information on exposures in pregnancy and breastfeeding is available, in the USA and Canada, from the MotherToBaby service of the Organization of Teratology Information Specialists (OTIS), with personalised risk assessment, more than 300 fact sheets and observational pregnancy exposure registries [82]. In 2010 the World Health Assembly adopted resolution WHA63.17 on birth defects, which includes registration and surveillance [3].

7.5. Medicinal products and regulatory status

No medicinal product is indicated for multiple congenital anomalies; the drugs used are those for each component or for each diagnosed syndrome, and doses follow the recommendations for each entity and the summaries of product characteristics. Genomic tests are the main management tool, and their regulation and reimbursement vary between countries: in EUROCAT, the proportion with a genetic diagnosis was lower in Poland (Wielkopolska), where advanced tests were not reimbursed, and in Ukraine and Malta, where termination for fetal anomaly is illegal or restricted [8].

The sources consulted do not include data on national access to genomic testing in Portugal or Brazil.

7.6. Developments in the last 12 months (October 2025 to October 2026)

The search found no update of the 2021 ACMG guideline in 2025–2026 nor an ACOG/SMFM statement specific to prenatal exome sequencing beyond SMFM #75; absence from the search does not prove non-existence [21, 76]. The ACMG secondary findings list was updated to version 3.3 in June 2025, before the window (section 8.7) [88].

7.7. Pregnancy, transition and special situations

In pregnancies following a diabetic embryopathy, glycaemic control is optimised before conception and the pregnancy is followed in maternal-fetal medicine [44]. In monochorionic twins, specialist ultrasound follow-up identifies twin–twin transfusion and the TRAP sequence [47, 48]. In March 2026 ERN-ITHACA published a guideline on transition from paediatric to adult care, applicable to young people with malformation syndromes and neurodevelopmental disorders [24].

8. Reproductive Options and Family Counselling

8.1. Confirmation of the diagnosis

Definitive counselling requires confirmation of the diagnosis, because of the medico-legal implications: clinical assessment, a result from an accredited laboratory (see the critical result), images or radiological report, death certificate (less reliable) and family photographs; some diagnoses reported by the family are notoriously unreliable, such as "achondroplasia" [89]. A known diagnosis confirmed in the laboratory makes counselling on recurrence and offspring risk relatively straightforward [13].

8.2. Recurrence risks by mechanism

Variable expression alters the meaning of the risk: in Waardenburg syndrome type I (PAX3), about 20% of affected individuals are deaf, so the risk of deafness in the offspring is 50% × 20% = 10% [36]. Unstable mutations may show anticipation dependent on the sex of the transmitting parent, such as congenital myotonic dystrophy, which arises only through maternal transmission [41]. In PAGE, fetal diagnoses by exome implied a low recurrence in 33 cases (5% of the total) and a high recurrence in 19 (3%) [49].

8.3. Features suggesting a high recurrence risk

In the undiagnosed child, a high risk, up to 25%, is suggested by a previously affected child, previous stillbirths or late miscarriages due to fetal anomalies, a relative with a similar picture (X-linked, chromosomal, dominant with variable expression), consanguinity, neurodevelopmental regression, facial features that become coarse, symmetrical structural defects described in recessive conditions and maternal influences, such as myasthenia with transplacental antibodies, with a very high risk in subsequent pregnancies [13]. Without these features, the empirical risk of about 5% may be given in counselling, explaining that it derives from the probability of a genetic cause [13].

8.4. Isolated multifactorial defects

Multifactorial counselling applies only after a careful examination has confirmed that the defect is isolated; the risk is 3–5% or less for most common defects (except scoliosis), slightly greater with a severe defect and lower with a mild defect, sex-specific where applicable, about 10–15% (2 to 3 times greater) with two affected children and 3–5% for the offspring of an affected person [35]. Explaining that the defect resulted from a single, localised problem of early development, which occurred before a given gestational age, dispels concerns about later events in the pregnancy [78].

Source: Smith, Table 2.1 [29]. For neural tube defects, a Brazilian chapter gives 4–5%, 25 to 50 times the population risk [51]. Empirical risks derive from surveys and not from models, and increase with two or more affected close relatives, a situation in which dominant inheritance with incomplete penetrance is considered [42].

8.5. Special cases

In bilateral renal agenesis or dysgenesis, renal ultrasound of the parents and siblings is recommended because 9% of 1st-degree relatives have asymptomatic renal malformations (Roodhooft 1984); autosomal recessive renal tubular dysgenesis has a recurrence risk of 25% [31]. In disruptions (amnion rupture, limb–body wall complex, twin defects), recurrence is usually negligible, although there is one report of two siblings with LBWC born to unaffected parents [30, 47, 50]. In MZ twins, the risk of both being affected approaches the risk of a singleton pregnancy, and in dizygotic twins each twin is assessed independently [57].

For susceptibility loci, there is no consensus on cascade testing or prenatal diagnosis, because of prognostic uncertainty, and some laboratories do not routinely report syndromes with substantial incomplete penetrance on prenatal arrays [20]. For a de novo microdeletion, prenatal diagnosis may be offered for reassurance, after first confirming that the rearrangement is pathogenic; for an inherited one, when the CNV is the cause of the anomaly, it is offered with adequate time and consent, even if the parent has learning difficulties; without an apparent phenotype, the CNV may be a polymorphism and prenatal diagnosis is not indicated [20]. Amniocentesis or chorionic villus sampling (CVS) is the only way to exclude recurrence due to gonadal mosaicism of a de novo anomaly, if the parents so wish [78].

8.6. Invasive procedures

Green-top Guideline No. 8 of the Royal College of Obstetricians and Gynaecologists (RCOG; 5th edition, October 2021; reviewed in December 2024 and extended for 2 years) recommends informing women that the additional risk of miscarriage after amniocentesis or CVS performed by a trained operator is probably below 0.5% (grade B); amniocentesis is not performed before 15+0 weeks (grade A) and CVS not before 10+0 weeks (grade D), with CVS preferred from 11+0 weeks (good practice point); in multiple pregnancies, the additional risk is about 1% (grade B); and blood-borne virus screening results, viral load and antigen should be reviewed before the procedure (grade C) [84]. CVS is usually performed at 11+0–13+6 weeks and may be performed at 14+0–14+6 weeks with individual counselling; amniocentesis is offered from 15+0 weeks, with written consent, and non-sensitised RhD-negative women receive anti-D prophylaxis [84].

In the meta-analysis by Akolekar and colleagues (studies from 2000 to 2014 with more than 1000 procedures), loss before 24 weeks was 0.81% (95% CI 0.58–1.08; 324/42,716) after amniocentesis and 2.18% (1.61–2.82; 207/8899) after CVS, versus 0.67% and 1.79% in controls, with a procedure-related risk of 0.11% (95% CI −0.04 to 0.26) for amniocentesis and 0.22% (−0.71 to 1.16) for CVS [90]. The running text of the Oxford Desk Reference swaps these two values, but its Table 6.8 agrees with the original [57]. In the update with studies up to September 2017 (10 studies with 64,901 amniocenteses and 6 with 19,000 CVS), cited by the RCOG, the weighted procedure-related risk was 0.35% for both procedures [84]. The risk increases with less experienced operators, and the RCOG advises that each operator should ideally perform or supervise at least 20 procedures per year and review practice if their annual rate of loss of normal babies exceeds 3% [84].

Sources: RCOG, Table 1; Oxford Desk Reference, Table 6.8 [57, 84, 90]. CVS before 10 weeks was associated with oromandibular and transverse limb defects (clusters in 1991–1992; risk 10–20 times baseline up to 9 weeks, close to baseline from 11 weeks), and early amniocentesis with more losses and clubfoot (1.3% versus 0.1% in the Canadian CEMAT trial) [57, 84]. Third-trimester amniocentesis is associated with preterm birth before 34 weeks in 3–4% and with culture failure in about 10% [84].

8.7. Prenatal and postnatal sequencing: counselling, consent and findings

Pre-test counselling is the standard of care: expectations, options regarding secondary findings and carrier status, the possibility of misattributed paternity or consanguinity, implications for life, disability and long-term care insurance (outside the protection of GINA in the USA) and a follow-up plan; post-test counselling includes a pedigree, discussion of pathogenic, benign and VUS results, risk in an unaffected parent and the opportunity for reanalysis [9, 21]. The harms described include misattributed paternity (2/155 cases in one series, with ethics consultation), actionable secondary findings (6/95 in one series, in FBN1, CACNA1S and SDHB), financial burden and loss of privacy, with no clinically significant psychological harm in the available studies, none of which with 20 or more patients assessed psychosocial outcomes [21, 69].

The ACMG SF v3.3 list (June 2025) is the minimum set of gene–disease pairs for opportunistic reporting of secondary findings in clinical exome and genome sequencing, updated annually; it added ABCD1, CYP27A1 and PLN and considered GCK, RUNX1 and SLC4A1 without adding them; it does not replace diagnostic testing and is not population screening, and in children disclosure prioritises the child's best interests [88].

Prenatally, the ISPD recommends trio testing, individualised pre-test counselling for both parents, separate consent from each parent, discussion of the policy on VUS, secondary findings (fetal: moderate to severe childhood conditions) and incidental findings, including carrier status, non-paternity and consanguinity, post-test counselling on all results, including uninformative ones, and reanalysis if a new phenotype emerges after birth or in a new pregnancy [23]. The laboratory should be accredited, with validated pipelines of stated sensitivity, a mandatory phenotype (biometry and imaging, preferably in HPO), interpretation by a multidisciplinary team and reporting focused on pathogenic or likely pathogenic variants relevant to the phenotype [23]. Danish practice reports only class 4/5 variants that explain the phenotype and does not usually report VUS in pregnancy [64]. The French Federation of Human Genetics proposes non-directive information, consent that respects autonomy regardless of educational level, a prior decision on reporting VUS and unsolicited results, national harmonisation of access and multidisciplinary meetings in complex cases [83]. In a qualitative study of 22 parents who declined prenatal genome sequencing in continued pregnancies with structural anomalies, the reasons were the risk of the invasive procedure (86%), the absence of actionable findings prenatally (77%), the low expected yield (59%), intolerance of uncertainty (59%), satisfaction with previous tests (50%), emotional burden (50%), unclear benefit (41%) and privacy (23%), with no postnatal regret [85].

Without a specific diagnosis, prenatal diagnosis is generally not possible, unless there are malformations attributed to the condition (for example a heart defect or polydactyly), which are looked for by 2nd-trimester ultrasound with a fetal medicine specialist, explaining the limitations [13]. In the absence of a sample from the deceased proband, the ISPD accepts sequencing both parents for shared carrier status of recessive variants, although it is preferable to obtain tissue from the previous fetus or child [23]. With a known variant, preimplantation genetic testing is one of the reproductive options listed by the AAP [9]. Counselling of minors about carrier status follows general principles and is not detailed in the sources consulted.

9. Special Considerations: Population Diversity and Epidemiology

9.1. What each figure measures

Prevalence estimates differ by numerator (major anomalies or all, isolated or multiple, with or without syndromes), by the outcomes included (live births only or also fetal deaths and terminations), by the period of ascertainment (at birth, up to 1 year or longer) and by the surveillance method (active or passive). An evident major congenital anomaly occurs in 2–3% of singleton newborns and doubles with follow-up [1], and the NBDPN report bases its estimate of 3–5% on a study by Bower and colleagues (2010) on age at diagnosis of birth defects [15]. The figures below describe specific programmes and are not extrapolated to national prevalences when the programme is not national.

9.2. Global burden

The WHO estimates that 9 in 10 children with a severe congenital disorder are born in low- or middle-income countries (about 94% of severe congenital disorders), with 240,000 neonatal deaths and 170,000 deaths between 1 month and 5 years each year [3].

9.3. United States

The USA does not have a national surveillance system; national estimates derive from NBDPN state programmes [32]. The CDC estimate that about 1 in 33 babies is born with a birth defect, and the NBDPN 2018–2022 report indicates that major structural defects affect 3–5% of births [15]. That report pooled 11 programmes with ascertainment of all pregnancy outcomes and review of medical records, with 4,231,701 live births (23% of US births), including infants with multiple defects in each applicable category; one additional programme was excluded because of systematic under-ascertainment [15].

Source: CDC/NBDPN, Data Brief of April 2026; unadjusted pooled prevalences, with exact Poisson CI [15]. Prevalence was stable for most defects over the 5 years; for 20 of the 33 defects, the highest prevalence occurred in the children of non-Hispanic American Indian or Alaska Native mothers and the lowest in the children of Asian or Pacific Islander mothers; mothers aged 40 or more had the highest prevalence for almost all defects (trisomy 21 117.9/10,000 versus 8.0 in mothers under 20 years), and mothers under 20 years the highest for anencephaly, encephalocele and gastroschisis (15.9/10,000) [15].

The previous NBDPN estimate (2016–2020; 13 programmes, 4,141,961 live births, 22% of births) adjusted for maternal race and ethnicity and, for the trisomies and gastroschisis, also for maternal age; age-adjusted trisomy 21 rose from 13.65/10,000 (1999–2001) to 17.19/10,000 (2016–2020; 1 in 582), transposition of the great arteries (TGA) fell from 4.73 to 2.99, AVSD rose from 4.36 to 5.84, and gastroschisis rose in the first three periods and fell in 2016–2020 [32]. The two series use different programmes and methods, so the differences between them are not read as a trend.

9.4. Europe

The EUROCAT network brings together population-based registries covering about 1.5 million births per year, more than 25% of births in Europe, and includes live births, fetal deaths from 20 weeks and terminations of pregnancy for fetal anomaly at any gestational age [8]. The JRC-EUROCAT platform publishes prevalence tables updated to birth year 2024 (August 2026 update) and, because of the General Data Protection Regulation, no longer shows prevalences by individual registry [91]. MCA without a syndrome had a prevalence of 15.8/10,000 births [16]. In 2000–2006, 16 registries from 11 countries notified 10,323 cases of unbalanced chromosomal anomalies (43.8/10,000 births): trisomies 21, 18 and 13 with 23.0, 5.9 and 2.3/10,000; 45,X with 3.3/10,000; sex chromosome trisomies with 2.0/10,000; and rare chromosomal anomalies (triploidy, other trisomies, markers, unbalanced translocations, deletions, duplications) with 7.4/10,000 (2.4–12.9 between registries), 65% of them diagnosed prenatally and 49% live births [92]. The table shows the proportion with a genetic diagnosis by registry in 2013–2022 [8].

The estimate is probably an underestimate: access to clinical genetics and follow-up vary, registries with more diagnoses after 1 year have a higher genetic proportion, there may be over-reporting of microarray findings without clinical consequence, and EUROCAT does not know what proportion of cases underwent genetic testing [8].

9.5. Portugal

The National Registry of Congenital Anomalies (RENAC), of the Instituto Nacional de Saúde Doutor Ricardo Jorge and a member of EUROCAT since 1990, notified 1213 cases in 79,873 births in 2020 (151.9/10,000) and 1058 in 84,832 in 2021 (124.7/10,000), a total of 2271 cases in 164,705 births (137.9/10,000), counting live births, stillbirths and terminations of pregnancy for medical reasons notified up to September 2024 [18]. The prevalence is always lower than expected by international standards and than that of EUROCAT, because of recognised under-reporting: 19.1% of obstetric departments and 8.5% of paediatric or neonatology departments did not report [18]. By group, heart defects had 57.0/10,000, chromosomal anomalies 30.8, musculoskeletal anomalies 30.7, urinary tract anomalies 28.7 and central nervous system anomalies 18.2; there was a prenatal diagnosis or suspicion in 66.0% of cases (obstetric ultrasound was the first investigation in 91.7%), and ultrasound surveillance did not follow Portuguese guidelines in 5.1% [18]. For chromosomal anomalies, terminations for medical reasons (25.1/10,000) exceeded live births (4.8/10,000), versus 31.3 and 12.8 in EUROCAT, and for heart defects live births were 30.0/10,000 in Portugal versus 74.8 in EUROCAT [18]. The differences reflect under-reporting and legal differences between countries, not necessarily real differences in occurrence.

9.6. Brazil

The Ministry of Health monitoring dashboard (SINASC, source of September 2026) gives absolute numbers of live births with a congenital anomaly recorded on the Live Birth Declaration (D18.0 and Q00–Q99) [7]. The prevalence was calculated for this article by dividing the dashboard numerator by the total number of live births from the same system and the same extraction date; it is not a value published by the Ministry and excludes stillbirths and terminations, so it is not directly comparable with RENAC or EUROCAT [7].

Between 2010 and 2021 the calculated prevalence ranged between 74.8 and 91.2/10,000; the rise in 2023–2025 is seen in the dashboard figures, but its cause (better recording or a real change) is not explained by the source [7]. Congenital malformations are the second cause of infant mortality in Brazil (11.2% of deaths) [51].

9.7. Special populations and equity

Consanguinity increases the prevalence of rare genetic disorders and almost doubles the risk of neonatal and infant death, ID and other anomalies [3]; in communities with a tradition of consanguinity, investigation includes recessive disorders with a high local carrier frequency [38]. In the DDD, the yield was lower in African ancestry (OR 0.51), partly because of the lack of allele frequency data from matched ancestry [10]. In SeqFirst-neo, the difference in yield in favour of first-tier genome sequencing was greater in non-white (45.9% versus 20.7%) and black newborns (80% versus 0/4), and neonatologists were less accurate in predicting a diagnosis in these groups [55]. In EUROCAT, the genetic proportion was lower in countries where termination is illegal or restricted and advanced tests are not reimbursed [8]. The lack of anthropometric norms for non-Caucasian populations limits dysmorphological assessment [28]. Twins are at increased risk (section 2.9), and mothers with pre-gestational diabetes constitute a defined risk group (section 2.7).

9.8. European Reference Network and patient organisations

ERN-ITHACA (European Reference Network for Rare Malformation Syndromes, Intellectual and Other Neurodevelopmental Disorders) maintains working groups on fetal medicine and pathology, registries (ILIAD registry) and guidelines, the CPMS (Clinical Patient Management System) for case discussion, the online "ITHACA Clinic", EuroDysmorpho 2026 and a winter school on fetal diagnosis; it has published guidelines on Beckwith-Wiedemann (2018), valproate embryopathy (2018), Cornelia de Lange (2019), Phelan-McDermid (2023), Rubinstein-Taybi (2024), spinal dysraphisms (2025), Kleefstra (2026) and transition to adult care (2026), and co-produced that on Bardet-Biedl (2018) [24].

Organisations and resources include Unique, for rare chromosomal and gene disorders, with leaflets for each anomaly [20, 93]; EURORDIS, the European rare disease federation [94]; SWAN UK (Syndromes Without A Name), of Genetic Alliance UK, a free community for families of children and young people up to 25 years with an undiagnosed genetic condition, including those with negative tests or VUS [95]; Contact a Family in the United Kingdom, which puts families in touch with others with similar features [13]; ARC (Antenatal Results and Choices) for decisions after prenatal diagnosis [19]; and, in Brazil, the Instituto Vidas Raras [96]. The Raríssimas website (Portugal) did not respond when accessed in October 2026, so its current activity was not confirmed. The DECIPHER and ECARUCA databases collect genomic imbalances with phenotypes, and POSSUM and the London Dysmorphology Database support dysmorphological diagnosis [13, 20, 78]. The Treatable-ID resource lists treatable causes of ID by clinical signs [97].

10. Illustrative Clinical Case

In 2026 Li and colleagues described a boy, the first child of non-consanguineous Chinese parents, born at term by spontaneous vaginal delivery, with a weight of 3.15 kg, length of 50 cm and head circumference of 34 cm (25th–50th centiles) and an Apgar score of 9/10 [54]. At birth he had a low-set right ear with a malformed helix, submucous cleft palate with bifid uvula, an accessory nipple on the left, bilateral cryptorchidism, micropenis, bilateral clubfoot and distal hypoplasia of the 5th fingers and of the toes, with profound generalised hypotonia and weak sucking that required nasogastric tube feeding in the first week [54]. The findings combine major anomalies (cleft palate, bilateral clubfoot, genital anomalies) and minor ones (accessory nipple, malformed ear), in a pattern of MCA with hypotonia that, according to section 4.1, requires a search for a syndrome.

Full blood count, biochemistry, serum alkaline phosphatase, tandem mass spectrometry and urinary organic acids were normal; brain magnetic resonance imaging, echocardiography and abdominal ultrasound showed no major abnormalities, except mild left pyelectasis, and the electroencephalogram showed multifocal irritative discharges without clinical seizures [54]. Trio exome sequencing, with CNV analysis by read depth validated by a second algorithm, identified a duplication of exons 4–22 of PIGN (MANE Select transcript NM_176787.5), with a minimum interval chr18:59,768,307–59,828,586 (hg19) and about 60 kb in the maximum interval; the proband had four copies and each parent three, which was confirmed by quantitative PCR (relative quantification of about 2.3 in the proband and about 1.5 in the parents) [54]. The variant was absent from the Database of Genomic Variants (DGV), ClinVar and gnomAD and was classified as a VUS by ACMG criteria; the exome did not allow the structure or orientation of the duplication to be determined [54]. The authors interpreted it as the most plausible explanation for MCAHS1, an autosomal recessive disorder of glycosylphosphatidylinositol (GPI) anchor biosynthesis [54]. Orphanet classifies MCAHS1 as ORPHA:280633 (synonyms PIGN deficiency CDG, PIGN-CDG; OMIM 614080; prevalence below 1/1,000,000; prenatal, neonatal or childhood onset); the GeneGuidance mapping to ORPHA:2059 (Fryns syndrome) does not correspond to this entity [14].

Figure 1 of the article shows the bilateral clubfoot with overlapping toes and hypoplasia of the 5th digits, the accessory nipple, the low-set, malformed right ear, the micropenis with undescended testes, the submucous cleft palate with bifid uvula and the renal ultrasound with mild pyelectasis; Figure 2 presents the pedigree with segregation, the CNV plot of chromosome 18 with the duplication at 18q21, the detail of the PIGN locus and the validation by quantitative PCR (https://pmc.ncbi.nlm.nih.gov/articles/PMC13628969/) [54]. At 12 months the boy was alive, with severe global delay (he did not roll over or smile responsively), persistent hypotonia, dependence on a nasogastric tube, no clinical seizures, a normal follow-up MRI and normal growth (head circumference 42 cm, weight 7.2 kg, length 68 cm), receiving physiotherapy and early intervention [54]. The phenotype was milder than usual in MCAHS1, in which most of the more than 30 described PIGN variants are associated with intractable seizures and early death [54].

The case illustrates three points of this article. First, CNV calling from the exome detected an intragenic duplication that analysis of single nucleotide variants alone would have missed, and whose detection at exon level requires dedicated CNV analysis or high-resolution exonic CMA [9, 54]. Second, the trio made it possible to demonstrate segregation compatible with recessive inheritance [21, 54]. Third, classification as a VUS, owing to the lack of functional validation at transcript level, limits counselling: the 25% recurrence risk and prenatal or preimplantation diagnosis are only securely supported by a variant classified as pathogenic or likely pathogenic, and the family's decisions should be discussed with that caveat [54, 67]. The management described (tube feeding, physiotherapy and early intervention) is that of the case and does not constitute a general recommendation.

11. Prognosis and Long-term Follow-up

11.1. Prognosis

The prognosis of MCA depends on the aetiology and on the individual defects, and there is no validated overall estimate. MCA have more fetal deaths and terminations than isolated anomalies, and liveborn infants are more often preterm and of low birth weight [16]. In the PAGE fetal cohort, 59% of fetuses with structural anomalies and known outcome were born alive, 3% died in the neonatal period and 65.4% of molecular diagnoses were associated with learning difficulties, 16 of them in fetuses without a brain or multisystem anomaly, which shows that ultrasound does not predict cognition [19, 49]. In subtelomeric deletions, mortality varies with the size of the deletion, as in Wolf-Hirschhorn syndrome, with 21% mortality in the first 2 years [20]. In MZ twins with a co-twin who died near term, about one-third of survivors have severe brain injury [47]. In the oligohydramnios sequence without a central nervous system deficit and in amnion rupture with isolated constrictions, the prognosis is good [30, 31]. Early prenatal ultrasound diagnosis influences neonatal prognosis, mode of delivery and any in utero treatment, and epidemiological surveillance allows prevention measures that reduce frequency and mortality [51].

11.2. Follow-up of the undiagnosed child

The child with an undiagnosed dysmorphic syndrome has a second consultation to complete investigations and exchange information, then a review at 1 year and, thereafter, reviews according to the family's needs, for example when there are reproductive plans [13]. Reviews serve to offer new tests, exclude additional medical problems, confirm developmental progress, increase the probability of recognition (some syndromes only become recognisable after a few years) and rediscuss the recurrence risk; departmental clinical photograph sessions and sharing with colleagues and databases contribute to late diagnosis [13]. SWAN UK supports these families in the United Kingdom [95].

11.3. Reanalysis of genomic data

In the DDD, each round of reanalysis added about one diagnostic variant for every six trios, and each year since recruitment increased the probability of diagnosis (OR 1.25) [10]. The 2019 consensus recommends exome reanalysis every 1–3 years (10–16% increase in yield) and the AAP every 1–2 years with an updated phenotype [9, 59]. In a cohort with six-monthly reanalysis for 3 years, the overall diagnostic rate rose from 18.3% to 20.4%, with a yield of 2.6% of non-diagnostic reports and a maximum about 2 years after the initial report, higher in paediatric groups [86]. In postnatal reanalysis of 87 negative or inconclusive prenatal exomes, new postnatal findings emerged in 64% and the incremental yield was 2% overall and 7% in those with new findings [98]. In the French AnDDI-Prénatome cohort, reanalysis of 58 unresolved prenatal cases with a postnatal or autopsy phenotype gave one diagnosis and three additional VUS, and the authors suggest postnatal genome sequencing [99]. The ISPD recommends reanalysis when a new phenotype emerges after birth or a new pregnancy is planned [23].

11.4. Surveillance

11.5. Transition to adulthood

In March 2026 ERN-ITHACA published a guideline on transition from paediatric to adult care in malformation syndromes and neurodevelopmental disorders [24]. SWAN UK supports undiagnosed young people up to 25 years [95].

12. Summary and Final Considerations

Congenital anomalies affect 2–3% of newborns at birth and about twice that with follow-up, and multiple anomalies are the subgroup with the highest probability of a genetic cause, the highest fetal mortality and the greatest need for coordinated care [1, 8, 16]. Assessment begins with the morphogenetic classification of each anomaly as a malformation, deformation, disruption or dysplasia and with the recognition of sequences, because the mechanism defines the prognosis and the recurrence risk: deformations and disruptions usually have a low recurrence, sequences inherit the risk of the initiating anomaly, and syndromes that of their aetiology [1, 30]. Systematic examination, with a count of minor anomalies (three or more are associated with a major malformation in 19.6–90% depending on the study), anthropometry in standard deviations, serial photographs and standardised terminology, is the foundation of the phenotyping that genomics requires [26, 28].

The testing strategy has moved from the karyotype to CMA in 2010 and to trio exome or genome sequencing from 2021, with CMA retained or replaced by CNV calling from the genome; targeted testing keeps its place when the suspicion is specific or the cause is not detectable by sequencing, such as methylation abnormalities and fragile X [9, 21, 22]. The yield varies with cohort selection, from about 26% in infants with MCA after negative karyotype and CMA to about 40% in severe paediatric cohorts, and in the fetus from 8.5–10% without selection to 31–42% in multisystem or selected phenotypes [10, 17, 49, 60]. A negative result does not exclude a genetic cause, a VUS does not support decisions without reassessment, and periodic reanalysis with an updated phenotype adds diagnoses [59, 86].

Counselling with a diagnosis follows the mechanism; without a diagnosis, it rests on an empirical risk of about 5%, composite and changeable, and on identifying the features that suggest a 25% risk [13]. Invasive procedures, performed by a trained operator, have an additional risk of loss probably below 0.5% [84]. There is no pharmacological treatment for multiple anomalies; management is surgical, rehabilitative and guided by the aetiology, and a genetic diagnosis changes management in a measurable minority of the patients tested, including decisions about palliative care [3, 21]. Primary prevention, with folate, preconception glycaemic control, rubella vaccination and avoidance of teratogens, remains the intervention with the widest population reach [3, 43]. Epidemiological surveillance (EUROCAT, NBDPN, RENAC, SINASC) measures what each system records and is comparable only when numerators, outcomes and methods coincide [7, 15, 18].

13. References

1. Approach to the consultation with a child with dysmorphism, congenital malformation, or developmental delay. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 4–7.
2. Spranger J, Benirschke K, Hall JG, Lenz W, Lowry RB, Opitz JM, et al. Errors of morphogenesis: concepts and terms. Recommendations of an international working group. J Pediatr. 1982;100(1):160-165. doi:10.1016/s0022-3476(82)80261-8. PMID: 7057306.
3. World Health Organization. Congenital disorders [fact sheet] [Internet]. Geneva: WHO; 2023 Feb 27 [accessed 17 Oct 2026]. Available from: https://www.who.int/news-room/fact-sheets/detail/birth-defects
4. Orphanet. Rare developmental defect during embryogenesis (ORPHA:93890) [Internet]. Paris: INSERM; [accessed 17 Oct 2026]. Available from: https://www.orpha.net/en/disease/detail/93890
5. Orphanet. Genetic multiple congenital anomalies/dysmorphic syndrome (ORPHA:183533) [Internet]. Paris: INSERM; [accessed 17 Oct 2026]. Available from: https://www.orpha.net/en/disease/detail/183533
6. Orphanet. Lethal multiple congenital anomalies/dysmorphic syndrome (ORPHA:459787) [Internet]. Paris: INSERM; [accessed 17 Oct 2026]. Available from: https://www.orpha.net/en/disease/detail/459787
7. Brasil. Ministério da Saúde, Secretaria de Vigilância em Saúde e Ambiente, DAENT. Painel de Monitoramento de Malformações Congênitas, Deformidades e Anomalias Cromossômicas (SINASC) [Internet]. Brasília: Ministério da Saúde; September 2026 [accessed 17 Oct 2026]. Available from: https://svs.aids.gov.br/daent/centrais-de-conteudos/paineis-de-monitoramento/natalidade/anomalias-congenitas/
8. Bergman JEH, Perraud A, Garne E, Barisic I, Tucker D, Ballardini E, et al. Genetic Diagnoses Among Congenital Anomaly Cases in Europe: Data From the EUROCAT Network. Paediatr Perinat Epidemiol. 2026;40(3):414-425. doi:10.1111/ppe.70099. PMID: 41277385. PMC13124677.
9. Rodan LH, Stoler J, Chen E, Geleske T; Council on Genetics. Genetic Evaluation of the Child With Intellectual Disability or Global Developmental Delay: Clinical Report. Pediatrics. 2025;156(1):e2025072219. doi:10.1542/peds.2025-072219. PMID: 40545261.
10. Wright CF, Campbell P, Eberhardt RY, Aitken S, Perrett D, Brent S, et al. Genomic Diagnosis of Rare Pediatric Disease in the United Kingdom and Ireland. N Engl J Med. 2023;388(17):1559-1571. doi:10.1056/nejmoa2209046. PMID: 37043637. PMC7614484.
11. Clark MM, Stark Z, Farnaes L, Tan TY, White SM, Dimmock D, et al. Meta-analysis of the diagnostic and clinical utility of genome and exome sequencing and chromosomal microarray in children with suspected genetic diseases. NPJ Genom Med. 2018;3:16. doi:10.1038/s41525-018-0053-8. PMID: 30002876. PMC6037748.
12. Genetics, genetic counseling, and prevention: genetic imbalance caused by gross chromosomal abnormalities and submicroscopic genomic imbalance. In: Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 929–937.
13. Dysmorphic child. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 134–137.
14. Orphanet. Multiple congenital anomalies-hypotonia-seizures syndrome (ORPHA:280633) [Internet]. Paris: INSERM; [accessed 17 Oct 2026]. Available from: https://www.orpha.net/en/disease/detail/280633
15. Lindsey NP, Markwalter VK, Michalski AM, Pabst LJ; National Birth Defects Prevention Network. Prevalence of selected birth defects from population-based birth defects surveillance programs in the United States, 2018-2022 [Data Brief]. Atlanta: Centers for Disease Control and Prevention; April 2026 [accessed 17 Oct 2026]. Available from: https://www.cdc.gov/birth-defects/media/pdfs/2026/04/Data-Brief-Birth-Defects-National-Estimates-2018-2022.pdf ; page Data and Statistics on Birth Defects (updated 21 Apr 2026): https://www.cdc.gov/birth-defects/data-research/facts-stats/index.html
16. Calzolari E, Barisic I, Loane M, Morris J, Wellesley D, Dolk H, et al. Epidemiology of multiple congenital anomalies in Europe: a EUROCAT population-based registry study. Birth Defects Res A Clin Mol Teratol. 2014;100(4):270-276. doi:10.1002/bdra.23240. PMID: 24723551.
17. Yang M, Kim JA, Jo HS, Park JH, Ahn SY, Sung SI, et al. Diagnostic Utility of Whole Genome Sequencing After Negative Karyotyping/Chromosomal Microarray in Infants Born With Multiple Congenital Anomalies. J Korean Med Sci. 2024;39(36):e250. doi:10.3346/jkms.2024.39.e250. PMID: 39315442. PMC11419962.
18. Braz P, Machado A, Aniceto C, Dias CM. Registo Nacional de Anomalias Congénitas: relatório 2020-2021. Lisboa: Instituto Nacional de Saúde Doutor Ricardo Jorge; November 2025. ISBN 978-989-9236-08-0. Available from: http://hdl.handle.net/10400.18/10770
19. Imaging in prenatal diagnosis. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 742–743.
20. Deletions and duplications (including microdeletions and microduplications). In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 660–665.
21. Manickam K, McClain MR, Demmer LA, Biswas S, Kearney HM, Malinowski J, et al. Exome and genome sequencing for pediatric patients with congenital anomalies or intellectual disability: an evidence-based clinical guideline of the American College of Medical Genetics and Genomics (ACMG). Genet Med. 2021;23(11):2029-2037. doi:10.1038/s41436-021-01242-6. PMID: 34211152.
22. Miller DT, Adam MP, Aradhya S, Biesecker LG, Brothman AR, Carter NP, et al. Consensus statement: chromosomal microarray is a first-tier clinical diagnostic test for individuals with developmental disabilities or congenital anomalies. Am J Hum Genet. 2010;86(5):749-764. doi:10.1016/j.ajhg.2010.04.006. PMID: 20466091. PMC2869000.
23. Van den Veyver IB, Chandler N, Wilkins-Haug LE, Wapner RJ, Chitty LS, ISPD Board of Directors. International Society for Prenatal Diagnosis Updated Position Statement on the use of genome-wide sequencing for prenatal diagnosis. Prenat Diagn. 2022;42(6):796-803. doi:10.1002/pd.6157. PMID: 35583085. PMC11220784.
24. ERN-ITHACA. European Reference Network for Rare Malformation Syndromes, Intellectual and Other Neurodevelopmental Disorders [Internet]. Paris: ERN-ITHACA; [accessed 17 Oct 2026]. Available from: https://ern-ithaca.eu/ (diretrizes: https://ern-ithaca.eu/guidelines/ern-ithaca-guidelines/)
25. MARDEN PM, SMITH DW, MCDONALD MJ. CONGENITAL ANOMALIES IN THE NEWBORN INFANT, INCLUDING MINOR VARIATIONS. A STUDY OF 4,412 BABIES BY SURFACE EXAMINATION FOR ANOMALIES AND BUCCAL SMEAR FOR SEX CHROMATIN. J Pediatr. 1964;64:357-371. doi:10.1016/s0022-3476(64)80188-8. PMID: 14130709.
26. Minor anomalies: clues to more serious problems and to the recognition of malformation syndromes. In: Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 955–972.
27. Deciphering Developmental Disorders Study. Large-scale discovery of novel genetic causes of developmental disorders. Nature. 2015;519(7542):223-228. doi:10.1038/nature14135. PMID: 25533962. PMC5955210.
28. Allanson JE, Biesecker LG, Carey JC, Hennekam RC. Elements of morphology: introduction. Am J Med Genet A. 2009;149A(1):2-5. doi:10.1002/ajmg.a.32601. PMID: 19127575. PMC2774524.
29. Genetics, genetic counseling, and prevention: multifactorial inheritance. In: Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 947–948.
30. Amnion rupture sequence. In: Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 898–901.
31. Oligohydramnios sequence (Potter syndrome). In: Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 892–893.
32. Stallings EB, Isenburg JL, Rutkowski RE, Kirby RS, Nembhard WN, Sandidge T, et al. National population-based estimates for major birth defects, 2016-2020. Birth Defects Res. 2024;116(1):e2301. doi:10.1002/bdr2.2301. PMID: 38277408. PMC10898112.
33. Minor congenital anomalies. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 240–241.
34. Limb reduction defects. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 212–215.
35. Genetics, genetic counseling, and prevention: genetic counseling for defects that are a result of multifactorial inheritance. In: Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 949–950.
36. Genetics, genetic counseling, and prevention: indications and sequence of chromosomal and genomic studies; genetic counseling for chromosomal and genomic abnormalities; genetic imbalance caused by single-gene disorders. In: Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 938–944.
37. Intellectual disability. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 188–193.
38. Developmental delay in the child with consanguineous parents. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 126–127.
39. Genomic sequencing and interpretation of data from WES or WGS analyses. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 32–33.
40. VariantValidator [Internet]. Manchester: University of Manchester; [accessed 17 Oct 2026]. Available from: https://variantvalidator.org/
41. Genetics, genetic counseling, and prevention: genetic counseling for single-gene disorders. In: Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 945–947.
42. Complex inheritance. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 14–15.
43. Bell R, Glinianaia SV, Tennant PW, Bilous RW, Rankin J. Peri-conception hyperglycaemia and nephropathy are associated with risk of congenital anomaly in women with pre-existing diabetes: a population-based cohort study. Diabetologia. 2012. doi:10.1007/s00125-012-2455-y. PMID: 22314812.
44. Maternal diabetes mellitus and diabetic embryopathy. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 756–757.
45. Drugs in pregnancy. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 718–721.
46. Radiation exposure, chemotherapy, and landfill sites. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 780–781.
47. Monozygotic twinning and structural defects—general. In: Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 914–921.
48. Twins and twinning. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 792–795.
49. Lord J, McMullan DJ, Eberhardt RY, Rinck G, Hamilton SJ, Quinlan-Jones E, et al. Prenatal exome sequencing analysis in fetal structural anomalies detected by ultrasonography (PAGE): a cohort study. Lancet. 2019;393(10173):747-757. doi:10.1016/s0140-6736(18)31940-8. PMID: 30712880. PMC6386638.
50. Limb–body wall complex. In: Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 902–903.
51. Viana LBOP, Ribeiro EM. Malformações do SNC. In: Ribeiro EM, coord. Síndromes Genéticas: Reconhecer para Tratar. Fortaleza: EdUnichristus; 2021. p. 289–290.
52. Dysmorphology examination checklist (Table A10). In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 830.
53. Gupta S, Bawa P, Kumari A, Panigrahi I, Srivastava P, Kaur A. Real-world performance of Face2Gene and GestaltMatcher for facial image analysis in a large Indian ethnic cohort. Eur J Med Genet. 2025;78:105063. doi:10.1016/j.ejmg.2025.105063. PMID: 41318057.
54. Li X, Wang X, Zhu J, Li H. Novel biallelic multi-exon duplication in the PIGN gene associated with multiple congenital anomalies-hypotonia-seizures syndrome 1. BMC Pediatr. 2026;26(1):926. doi:10.1186/s12887-026-06923-2. PMID: 42420894. PMC13628969.
55. Wenger TL, Scott A, Kruidenier L, Sikes M, Keefe A, Buckingham KJ, et al. SeqFirst: Building equity access to a precise genetic diagnosis in critically ill newborns. Am J Hum Genet. 2025;112(3):508-522. doi:10.1016/j.ajhg.2025.02.003. PMID: 39999847. PMC11947171.
56. Appendix: pattern of malformation differential diagnosis by anomalies. In: Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 999–1065.
57. Invasive techniques and genetic tests in prenatal diagnosis. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 744–747.
58. Riggs ER, Andersen EF, Cherry AM, Kantarci S, Kearney H, Patel A, et al. Technical standards for the interpretation and reporting of constitutional copy-number variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics (ACMG) and the Clinical Genome Resource (ClinGen). Genet Med. 2020;22(2):245-257. doi:10.1038/s41436-019-0686-8. PMID: 31690835. PMC7313390.
59. Srivastava S, Love-Nichols JA, Dies KA, Ledbetter DH, Martin CL, Chung WK, et al. Meta-analysis and multidisciplinary consensus statement: exome sequencing is a first-tier clinical diagnostic test for individuals with neurodevelopmental disorders. Genet Med. 2019;21(11):2413-2421. doi:10.1038/s41436-019-0554-6. PMID: 31182824. PMC6831729. Erratum: Genet Med. 2020;22(10):1731-1732 (doi:10.1038/s41436-020-0913-3).
60. Mellis R, Oprych K, Scotchman E, Hill M, Chitty LS. Diagnostic yield of exome sequencing for prenatal diagnosis of fetal structural anomalies: A systematic review and meta-analysis. Prenat Diagn. 2022;42(6):662-685. doi:10.1002/pd.6115. PMID: 35170059. PMC9325531.
61. Guo F, Liu R, Pan Y, Collins C, Bean L, Ma Z, et al. Evidence from 2100 index cases supports genome sequencing as a first-tier genetic test. Genet Med. 2024;26(1):100995. doi:10.1016/j.gim.2023.100995. PMID: 37838930.
62. Lesurf R, Jain A, Hanafi N, Mitina A, Yin Y, Kolla VA, et al. Genomic Yield of Long-Read Sequencing in Congenital Heart Disease. Circ Genom Precis Med. 2026:e005305. doi:10.1161/circgen.125.005305. PMID: 42813287.
63. Fu Y, English AC, Paulin LF, Jhangiani SN, Gogate N, Weissenberger G, et al. Cost-efficient long-read trio-barcoded adaptive sequencing improves rare disease diagnosis. Nat Commun. 2026;17(1):10546. doi:10.1038/s41467-026-77535-8. PMID: 42838976.
64. Vogel I, Andreasen L, Balslev-Harder M, Becher N, Ernst A, Gadsbøll K, et al. Whole Genome Sequencing in Prenatal Diagnostics: The Danish Approach to Guideline Formation and Implementation Within Public Healthcare. Prenat Diagn. 2026;46(5-6):669-686. doi:10.1002/pd.6780. PMID: 40122702. PMC13170064.
65. Kamolvisit W, Cheawsamoot C, Chetruengchai W, Kor-Anantakul P, Thangpong R, Srichomthong C, et al. Singleton rapid long-read genome sequencing as first tier genetic test for critically Ill children with suspected genetic diseases. Eur J Hum Genet. 2025;33(6):768-774. doi:10.1038/s41431-025-01818-9. PMID: 40011756. PMC12185673.
66. Mithyantha R, Kneen R, McCann E, Gladstone M. Current evidence-based recommendations on investigating children with global developmental delay. Arch Dis Child. 2017;102(11):1071-1076. doi:10.1136/archdischild-2016-311271. PMID: 29054862. PMC5738593.
67. Richards S, Aziz N, Bale S, Bick D, Das S, Gastier-Foster J, et al. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genet Med. 2015;17(5):405-424. doi:10.1038/gim.2015.30. PMID: 25741868. PMC4544753.
68. Matthijs G, Souche E, Alders M, Corveleyn A, Eck S, Feenstra I, et al. Guidelines for diagnostic next-generation sequencing. Eur J Hum Genet. 2016;24(1):2-5. doi:10.1038/ejhg.2015.226. PMID: 26508566. PMC4795226.
69. Malinowski J, Miller DT, Demmer L, Gannon J, Pereira EM, Schroeder MC, et al. Systematic evidence-based review: outcomes from exome and genome sequencing for pediatric patients with congenital anomalies or intellectual disability. Genet Med. 2020;22(6):986-1004. doi:10.1038/s41436-020-0771-z. PMID: 32203227. PMC7222126.
70. Kim HH, Chang YS, Park WY, Park J. Trio Whole-Genome Sequencing for Diagnosis of Neonates with Multiple Congenital Anomalies. Neonatology. 2026:1-8. doi:10.1159/000552497. PMID: 42139173.
71. Keçeci R, Keçeci HN, Başdemirci M. Microarray Application in Newborns With Multiple Congenital Anomalies: Genotype-Phenotype Correlation. Birth Defects Res. 2025;117(7):e2509. doi:10.1002/bdr2.2509. PMID: 40650423.
72. Lucca C, Rosina E, Pezzani L, Piazzolla D, Spaccini L, Scatigno A, et al. First-Tier Versus Last-Tier Trio Whole-Genome Sequencing for the Diagnosis of Pediatric-Onset Rare Diseases. Clin Genet. 2025;108(4):412-421. doi:10.1111/cge.14760. PMID: 40274276. PMC12405173.
73. Grandjean H, Larroque D, Levi S. Sensitivity of routine ultrasound screening of pregnancies in the Eurofetus database. The Eurofetus Team. Ann N Y Acad Sci. 1998;847:118-124. doi:10.1111/j.1749-6632.1998.tb08932.x. PMID: 9668704.
74. Callaway JL, Shaffer LG, Chitty LS, Rosenfeld JA, Crolla JA. The clinical utility of microarray technologies applied to prenatal cytogenetics in the presence of a normal conventional karyotype: a review of the literature. Prenat Diagn. 2013;33(12):1119-1123. doi:10.1002/pd.4209. PMID: 23983223. PMC4285999.
75. Petrovski S, Aggarwal V, Giordano JL, Stosic M, Wou K, Bier L, et al. Whole-exome sequencing in the evaluation of fetal structural anomalies: a prospective cohort study. Lancet. 2019;393(10173):758-767. doi:10.1016/s0140-6736(18)32042-7. PMID: 30712878.
76. Society for Maternal‐Fetal Medicine (SMFM), Sparks TN, Norton ME, SMFM Publications Committee. Society for Maternal-Fetal Medicine Consult Series #75: Evaluation and management of non-immune hydrops fetalis. Pregnancy (Hoboken). 2026;2(2):e70208. doi:10.1002/pmf2.70208. PMID: 42596935. PMC13344273.
77. Marie S, Dewulf JP, Nassogne MC. Disorders of Purine and Pyrimidine Metabolism. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 32, section 32.1. doi:10.1007/978-3-662-63123-2_32
78. Genetics, genetic counseling, and prevention: prenatal diagnosis. In: Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 950–953.
79. Petrikin JE, Cakici JA, Clark MM, Willig LK, Sweeney NM, Farrow EG, et al. The NSIGHT1-randomized controlled trial: rapid whole-genome sequencing for accelerated etiologic diagnosis in critically ill infants. NPJ Genom Med. 2018;3:6. doi:10.1038/s41525-018-0045-8. PMID: 29449963. PMC5807510.
80. Dimmock D, Caylor S, Waldman B, Benson W, Ashburner C, Carmichael JL, et al. Project Baby Bear: Rapid precision care incorporating rWGS in 5 California children's hospitals demonstrates improved clinical outcomes and reduced costs of care. Am J Hum Genet. 2021;108(7):1231-1238. doi:10.1016/j.ajhg.2021.05.008. PMID: 34089648. PMC8322922.
81. Wenger TL, Scott AA, Kruidenier L, Hayek JC, Beck AE, Chow P, et al. Hospital-wide implementation of inpatient first-tier rapid genome sequencing. Genet Med. 2026:102685. doi:10.1016/j.gim.2026.102685. PMID: 42615209.
82. MotherToBaby – Organization of Teratology Information Specialists (OTIS) [Internet]. [accessed 17 Oct 2026]. Available from: https://mothertobaby.org/
83. Cogan G, Troadec MB, Devillard F, Saint-Frison MH, Geneviève D, Vialard F, et al. Use of Prenatal Exome Sequencing: Opinion Statement of the French Federation of Human Genetics Working Group. Prenat Diagn. 2025;45(3):299-309. doi:10.1002/pd.6692. PMID: 39532683. PMC11893516.
84. Navaratnam K, Alfirevic Z; Royal College of Obstetricians and Gynaecologists. Amniocentesis and chorionic villus sampling: Green-top Guideline No. 8 July 2021. BJOG. 2022;129(1):e1-e15. doi:10.1111/1471-0528.16821. PMID: 34693616.
85. Weingarten LS, Rosenbaum A, de Voest J, Galloway S, Giordano JL, Stover S, et al. The Perspectives and Experiences of Prospective Parents Declining Diagnostic Prenatal Genome Sequencing in Continuing Pregnancies With Fetal Structural Anomalies. Prenat Diagn. 2026;46(5-6):727-736. doi:10.1002/pd.70034. PMID: 41339121.
86. Ting YL, Williams TJ, Metz H, Li M, Stetler M, Knight Johnson AE, et al. Diagnostic yield of exome reanalysis over time: Contribution of reevaluation type, timing, and patient phenotype. Genet Med. 2026;28(5):102549. doi:10.1016/j.gim.2026.102549. PMID: 41795598.
87. Zhang M, Li J, Ren C, Liu Y, Bian W, Han X, et al. Comparison of Medium-Coverage Whole-Genome Sequencing and Chromosomal Microarray in Prenatal Testing of Absence of Heterozygosity. Prenat Diagn. 2026;46(10):1565-1581. doi:10.1002/pd.70218. PMID: 42559979. PMC13571803.
88. Lee K, Abul-Husn NS, Amendola LM, Brothers KB, Chung WK, Gollob MH, et al. ACMG SF v3.3 list for reporting of secondary findings in clinical exome and genome sequencing: A policy statement of the American College of Medical Genetics and Genomics (ACMG). Genet Med. 2025;27(8):101454. doi:10.1016/j.gim.2025.101454. PMID: 40568962. PMC12318660.
89. Confirmation of diagnosis. In: Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 18–19.
90. Akolekar R, Beta J, Picciarelli G, Ogilvie C, D'Antonio F. Procedure-related risk of miscarriage following amniocentesis and chorionic villus sampling: a systematic review and meta-analysis. Ultrasound Obstet Gynecol. 2015;45(1):16-26. doi:10.1002/uog.14636. PMID: 25042845.
91. European Commission Joint Research Centre. EUROCAT – European network of population-based registries for the epidemiological surveillance of congenital anomalies [Internet]. Ispra: JRC; [accessed 17 Oct 2026]. Available from: https://eu-rd-platform.jrc.ec.europa.eu/eurocat_en
92. Wellesley D, Dolk H, Boyd PA, Greenlees R, Haeusler M, Nelen V, et al. Rare chromosome abnormalities, prevalence and prenatal diagnosis rates from population-based congenital anomaly registers in Europe. Eur J Hum Genet. 2012;20(5):521-526. doi:10.1038/ejhg.2011.246. PMID: 22234154. PMC3330224.
93. Unique – Understanding Rare Chromosome and Gene Disorders [Internet]. [accessed 17 Oct 2026]. Available from: https://www.rarechromo.org/
94. EURORDIS – Rare Diseases Europe [Internet]. Paris: EURORDIS; [accessed 17 Oct 2026]. Available from: https://www.eurordis.org/
95. Genetic Alliance UK. SWAN UK (Syndromes Without A Name) [Internet]. London: Genetic Alliance UK; [accessed 17 Oct 2026]. Available from: https://geneticalliance.org.uk/support-and-information/swan-uk-syndromes-without-a-name/
96. Instituto Vidas Raras [Internet]. [accessed 17 Oct 2026]. Available from: https://www.vidasraras.org.br/
97. GeneGuidance. Developmental delay and intellectual disability — diagnosis library. 28 Jul 2022 [accessed 17 Oct 2026]. Local GeneGuidance database.
98. Swanson K, Hodoglugil U, Sparks TN, Lianoglou BR, Slavotinek AM, Norton ME. Diagnostic Yield After Postnatal Reanalysis of Prenatal Exome Sequencing Results. Prenat Diagn. 2026;46(5-6):924-932. doi:10.1002/pd.6886. PMID: 40935603. PMC12722113.
99. Thauvin-Robinet C, Garde A, Favier M, Delanne J, Racine C, Rousseau T, et al. Reanalysis of unsolved prenatal exome sequencing for structural defects: diagnostic yield and contribution of postnatal/postmortem features. Eur J Hum Genet. 2025;33(5):675-682. doi:10.1038/s41431-025-01823-y. PMID: 40186013. PMC12048470.