Maternal Diabetic Embryofetopathy

Teratogenic spectrum of maternal metabolic origin: congenital malformations induced by periconceptional hyperglycaemia in pregestational diabetes (embryopathy) and fetal and neonatal effects of hyperglycaemia in the second half of pregnancy (fetopathy), with caudal regression syndrome as the most characteristic form.

Literature and regulatory information updated: 9 October 2026.

0. Initial Reference Table

1. Introduction

Maternal diabetes mellitus is the metabolic teratogen with the greatest population impact in high-income countries, because it combines a common exposure with a malformative effect exerted within a very short window, often before pregnancy is recognised. Orphanet defines diabetic embryopathy as a rare teratological disorder characterised by a range of congenital anomalies associated with poorly controlled maternal diabetes, frequently cardiovascular or central nervous system (CNS) anomalies, but also skeletal, craniofacial and other system anomalies; uncontrolled maternal diabetes also causes fetal and neonatal metabolic, respiratory and delivery complications [1]. The term embryofetopathy encompasses these two components: embryopathy, resulting from periconceptional and first-trimester hyperglycaemia, and fetopathy, resulting from hyperglycaemia and fetal hyperinsulinism in the second half of gestation, which the Brazilian textbook of genetic syndromes describes as infant of a diabetic mother syndrome [2].

1.1. Historical overview

The most characteristic caudal anomaly was described long before its relationship with diabetes was known. Caudal regression was described by Geoffroy Saint-Hilaire and Hohl in 1852, and in 1961 Bernard Duhamel proposed, in the article “From mermaid to anal imperforation: the syndrome of caudal regression”, that sirenomelia and lesser forms of sacral agenesis and anal imperforation represented extremes of a single spectrum of caudal mesoderm deficiency [5, 26]. The association with maternal diabetes was described in 1965 by Rusnak and Driscoll, who reported congenital vertebral anomalies in infants of diabetic mothers, and by Kucera, Lenz and Maier, who described malformations of the lower limbs and caudal spine in the same population; in 1966 Passarge and Lenz consolidated the designation caudal regression syndrome in infants of diabetic mothers, reporting further cases [6, 27, 28]. In 1971 Kucera published an analysis of the rate and type of congenital anomalies in the offspring of diabetic women, which became a classic reference for the spectrum [29].

Three milestones shaped clinical understanding of the disorder. In 1979, Mills, Baker and Goldman dated the most common malformations in infants of diabetic mothers morphologically and showed that they occur before the 7th week of gestation, meaning that any intervention to reduce them must be in place during this early period [16]. In 1981, Miller and colleagues first associated elevated haemoglobin A1c (HbA1c) in early pregnancy with major congenital anomalies [30]. In 1983, Fuhrmann and colleagues, in the former German Democratic Republic, showed in a non-randomised study that intensive metabolic control begun before conception was associated with a malformation rate of 0.8% (1 in 128), compared with 7.5% (22 in 292) when strict control began after 8 weeks [31]. Not all studies agreed: the 1988 US Diabetes in Early Pregnancy Study found neither significantly higher mean blood glucose nor HbA1c during organogenesis in mothers of malformed newborns, although women enrolled early had fewer malformations than those enrolled late [32]. The large population registries of subsequent decades, discussed in section 3, established a continuous relationship between periconceptional HbA1c and anomaly risk [15, 33, 34].

Associated entities were delineated in parallel. Daentl delineated femoral hypoplasia–unusual facies syndrome in 1975 on the basis of 4 patients, following isolated cases described by Franz and O'Rahilly (1961) and Kucera (1965), and Johnson and colleagues showed in 1983 that 12 of 36 published cases (33%) were infants of diabetic mothers [6, 35]. In 1986 Stevenson proposed the vascular steal theory for sirenomelia, separating it pathogenetically from caudal regression [5]. In 1999 Splitt and colleagues described the specific association between left isomerism sequence and maternal type 1 diabetes [36]. Experimental studies by Eriksson, Reece and other groups, synthesised in 2016, shifted the mechanistic explanation from hyperglycaemia alone to a set of metabolic and oxidative stress pathways in the embryo and yolk sac [13, 37].

1.2. Nomenclature and reclassifications

Current nomenclature distinguishes maternal disease from the condition of the conceptus. The World Health Organization International Classification of Diseases (ICD-10, 2019 version) codes syndrome of infant of a diabetic mother as P70.1, including a fetus or newborn affected by pre-existing maternal diabetes, and syndrome of infant of a mother with gestational diabetes as P70.0; each malformation is additionally coded in chapter Q [7]. The Brazilian textbook of genetic syndromes uses code P70.1 for infant of a diabetic mother syndrome [2]. Pregestational, or pre-existing, diabetes denotes any type of diabetes diagnosed before the current pregnancy, chiefly type 1 diabetes (T1D) and type 2 diabetes (T2D) [14]. Gestational diabetes mellitus (GDM) diagnosed after the first trimester has a small malformative effect, which becomes relevant when it represents T2D that was undiagnosed before pregnancy [17, 21].

Orphanet classifies caudal regression syndrome as a rare congenital malformation of the lower spinal segments, with a high, truncated conus medullaris, aplasia or hypoplasia of the sacrum and lumbar spine, and associated gastrointestinal, genitourinary, skeletal and nervous system malformations [3]. It groups it with sirenomelia in the caudal regression–sirenomelia spectrum (ORPHA:444941), defined as a group of developmental defects with varying degrees of malformation of the caudal abdomen, pelvis, kidneys, anorectal region, urogenital tract and lumbosacral spine, with or without lower limb fusion, ranging from isolated coccygeal agenesis to a single rudimentary limb [4]. Smith's Recognizable Patterns of Human Malformation (8th edition, 2022) records that recent evidence suggests sirenomelia and caudal dysplasia are not pathogenetically related, although the issue remains controversial [27].

1.3. Definition of the entities covered

Diabetic embryopathy (ORPHA:1926) is defined by one or more congenital malformations from the diabetes-associated spectrum in the conceptus of a mother with pregestational diabetes, in the absence of another explanation, with no formal criteria published in the sources consulted [1, 19]. Diabetic fetopathy is defined by excessive, disproportionate fetal growth and the neonatal complications of fetal hyperinsulinaemia (hypoglycaemia, polycythaemia, hypocalcaemia, hyperbilirubinaemia, pulmonary immaturity and hypertrophic cardiomyopathy) [2]. Caudal regression syndrome (ORPHA:3027), treated as a subtype in this article, is the anomaly with the highest relative risk associated with diabetes, although most cases occur in infants of non-diabetic mothers [18, 19]. Sirenomelia and femoral hypoplasia–unusual facies syndrome are included as differential diagnoses with partial aetiological overlap [5, 6].

2. Molecular Genetics and Pathophysiology

2.1. Nature of the exposure and critical window

Diabetic embryofetopathy has no causal gene: it results from exposure of the embryo and fetus to a metabolically altered intrauterine environment. The review by Eriksson and Wentzel (2016) describes it as an interaction between maternal diabetic status and genetic predisposition in the embryo, with malformations induced in early gestation and risk increasing with the degree of maternal metabolic dysregulation [13]. Malformations that are significantly more frequent in infants of diabetic mothers develop before the 7th week of gestation [16], and organogenesis occurs mainly between 5 and 8 weeks, the interval used by the American Diabetes Association (ADA) to justify glycaemic optimisation before conception [21]. The Brazilian textbook of genetic syndromes places the critical period within the first 6 to 8 weeks [2]. Caudal regression results from disruption of structures derived from the caudal eminence before 4 weeks of gestation, at a stage when many women are still unaware of the pregnancy [26].

2.2. Maternal teratogens: glucose and ketone bodies

The two most extensively studied maternal teratogens are hyperglycaemia and hyperketonaemia [13]. Glucose injection into non-diabetic animals and embryo culture in high-glucose medium or serum from diabetic animals cause dysmorphogenesis, whereas L-glucose, which is not metabolised, has no effect; damage therefore depends on glucose metabolism rather than osmolarity [13]. This effect requires increased glucose uptake by embryonic transporters, sometimes overexpressed, with greater glycolytic flux, increased Krebs cycle and oxidative phosphorylation activity, and mitochondrial superoxide production, probably the component with the greatest teratogenic potential; this is followed by increased pentose pathway activity, inhibition of glyceraldehyde-3-phosphate dehydrogenase by reactive oxygen species (ROS), formation of α-oxoaldehydes and advanced glycation end products (AGEs), and reduced availability of arachidonic acid and prostaglandins [13]. In the preimplantation embryo, diabetes reduces the inner cell mass of the blastocyst through intracellular hypoglycaemia, consequent upon downregulation of glucose transporters, and apoptosis [13].

β-Hydroxybutyrate and acetoacetate cross the placenta. Serum from diabetic rats remains teratogenic after glucose normalisation, ketone bodies disrupt development in vitro, and the combination of glucose and β-hydroxybutyrate at individually subteratogenic concentrations is teratogenic, suggesting synergy through hyperactivity of oxidative phosphorylation [13]. The Oxford Desk Reference summarises the mechanism as simultaneous disruption of interrelated teratological pathways, including inositol metabolism, prostaglandins and ROS [19]. Clinical studies have not established an association between maternal hypoglycaemia and embryopathy in humans, although animal models indicate that hypoglycaemia is potentially teratogenic during embryogenesis [19]; in the Diabetes in Early Pregnancy Study, maternal hypoglycaemia (≤50 mg/dL) was not more frequent in mothers of malformed newborns [32].

2.3. Cellular and molecular pathways demonstrated in animal models

Oxidative stress is the mechanism with the strongest experimental support. Embryos and yolk sacs from diabetic rats have reduced antioxidant enzyme activity and lower concentrations of vitamins C and E, lower still in malformed embryos, and diabetes-prone Cohen rats are unable to increase antioxidant enzymes in response to hyperglycaemia [38]. Supplementation with superoxide dismutase (SOD), catalase, glutathione peroxidase, N-acetylcysteine, vitamins E and C, folic acid, α-lipoic acid or butylated hydroxytoluene reduces malformations in vitro and in vivo; mice transgenic for SOD1 or thioredoxin-1 have fewer malformations, and antioxidants prevent cardiac outflow tract defects in a murine model by preserving neural crest cell migration [13]. Excess ROS is a mechanism shared with other teratogens, such as phenytoin and ethanol [13]. Hyperglycaemia reduces oxygen availability in the mouse embryo by approximately 30%; 12% ambient oxygen increases the neural tube defect (NTD) rate by a factor of 8 and 30% suppresses it, linking hypoxic and oxidative stress [13].

Other pathways converge on apoptosis of neuroepithelial and neural crest cells. 3-Deoxyglucosone induces malformations reversible with SOD, mice lacking the AGE receptor (RAGE) are protected, inducible nitric oxide synthase promotes nitrosative stress and apoptosis, and endoplasmic reticulum stress, with activation of the unfolded protein response, JNK1/2 kinase and caspases, causes NTDs that are blocked by 4-phenylbutyric acid [13]. Diabetes suppresses autophagy in the neuroepithelium, an effect prevented by trehalose in an animal model, and endoplasmic reticulum stress in the endocardial cushions is associated with atrioventricular septal defect (AVSD) [13]. The hexosamine pathway, through increased UDP-N-acetylglucosamine and O-GlcNAcylation, inhibits Pax3, and the ASK1–FoxO3a–TRADD–caspase-8 pathway is activated; blocking these pathways reduces NTDs, and activation of the same genes has been observed in human NTD tissues [13]. Reduced prostaglandin E2 due to glucose-induced downregulation of cyclo-oxygenase 2 has been documented in rodent embryos during neural tube closure and in yolk sacs from embryos of diabetic women, and supplementation with arachidonic acid or prostaglandin E2 blocks dysmorphogenesis in animal models [13].

The yolk sac theory adds a vascular explanation. The yolk sac vascular system is the first to form and the most sensitive to hyperglycaemia; yolk sac vasculopathy, mediated by hypoxia-inducible factor 1α, the ASK1–thioredoxin-1 pathway, JNK, nitric oxide, arachidonic acid and myo-inositol, is inversely related to malformations in animal models, and human embryonic vasculogenesis is also affected by maternal diabetes [37].

2.4. Gene expression, epigenetics and genetic predisposition

Pax3 expression is reduced in embryos from diabetic mice, and the Splotch mutant, lacking functional Pax3, has NTDs; Pax3 blocks p53 during the neural crest migration required for cardiac outflow tract septation, linking the same alteration to NTDs and conotruncal heart defects [13]. Oxidative stress activates the DNA methyltransferase Dnmt3b, which methylates and silences Pax3; CuZnSOD, MnSOD and glutathione peroxidases 1 and 2 are also reduced, and TGF-β and Wnt signalling, implicated in heart defects and in a model of caudal dysgenesis, is altered [13]. Greater variability in gene expression between exposed embryos has been proposed as a correlate of incomplete penetrance, and methylation changes in Peg3 and H19 have been described in oocytes and placenta [13].

Genetic predisposition has been demonstrated in rodents, not in humans. In rat substrains, skeletal malformations varied between 0%, 3–5% and 17–19% according to the maternal and fetal genomes; embryo transfer between NOD and ICR mice produced malformation rates of 14%, 40% and 0%; a catalase variant modified risk; and linkage analysis identified loci on 16 chromosomes, near genes such as Shh, Tgfb3, Vegfa, Folr1, Akr1b1 and Tp53 [13]. No universal susceptibility gene has been identified. Men with diabetes have no increased risk of offspring with malformations, showing that diabetes predisposition genes do not cause the malformations and that the effect depends on the maternal environment [13].

2.5. Pathophysiology of fetopathy

Fetopathy follows the sequence of maternal hyperglycaemia, fetal hyperglycaemia and fetal hyperinsulinaemia. Fetal hyperinsulinism produces macrosomia and disproportionate adipose growth of the abdomen, shoulders and chest relative to the head; hyperglycaemia and hyperinsulinaemia delay pulmonary maturation; reduced oxygen exchange stimulates erythropoietin, causing polycythaemia and hyperviscosity; persistent hyperinsulinism after cord ligation causes neonatal hypoglycaemia; and prematurity and polycythaemia contribute to hypocalcaemia and hyperbilirubinaemia [2]. Postprandial hyperglycaemia is the main determinant of macrosomia and is poorly captured by HbA1c [21].

2.6. Genes relevant to differential diagnosis and special populations

Four gene groups are relevant to this article without causing embryopathy. MNX1 (HGNC:4979; OMIM 142994; 7q36.3), formerly HLXB9, causes Currarino syndrome in the heterozygous state and a form of permanent neonatal diabetes with caudal malformations in the biallelic state [8]. The MANE Select reference transcript is ENST00000252971.11, corresponding to RefSeq NM_005515.4 (transcript variant 1; CCDS34788), with coordinates 7:156,993,901–157,010,668 on GRCh38; NM_001165255 given in the GeneGuidance curation is not the MANE transcript and should not be used for HGVS nomenclature in reports [8, 39]. No pathogenic MNX1 variant was found in 48 cases of caudal dysplasia sequence, genetically separating caudal regression from Currarino syndrome [27]. VANGL1 (1p13.1; gene MIM 610132) has rare, low-penetrance variants associated with NTD susceptibility and, in a pilot exome study, a heterozygous variant was one of several candidates in a patient with sacral agenesis; the Gene Curation Coalition (GenCC) classifies the evidence as limited, and FUZ, linked by Orphanet to caudal regression, likewise has limited evidence [3, 9]. X-linked ZIC3 causes heterotaxy and VACTERL, the two phenotypes most easily confused with the laterality pattern of embryopathy [10, 40].

In maternal monogenic diabetes, the effect is on fetal growth and depends on the fetal genotype. Diabetes caused by heterozygous GCK (glucokinase) variants accounts for 30–60% of maturity-onset diabetes of the young (MODY, maturity-onset diabetes of the young) and causes mild, stable fasting hyperglycaemia from birth, at 5.5 to 8.0 mmol/L (99–144 mg/dL), with HbA1c around 5.6–7.6%, requiring no treatment outside pregnancy [11, 12]. A fetus that does not inherit the maternal variant perceives maternal blood glucose as elevated, increases insulin secretion and is on average 700 g heavier at birth, whereas a fetus that inherits the variant grows normally [11, 12]. A fetus inheriting an HNF4A variant from either parent produces more insulin in utero, is on average 800 g heavier at birth and has neonatal hypoglycaemia in approximately 10% of cases; median birth weight is 1.8–2.5 kg in neonatal diabetes, 2.6 kg with HNF1B, approximately 3 kg with GCK and 4.3 kg with HNF4A [11]. The Oxford Desk Reference describes no malformative effect of these monogenic forms, and monogenic diabetes accounts for 1–2% of people with diabetes [11].

2.7. Maternal factors modifying risk

In Northern England, independent predictors of major congenital anomaly were periconceptional HbA1c, with an adjusted odds ratio (aOR) of 1.3 (95% confidence interval [95% CI] 1.2–1.4) for each 1% (11 mmol/mol) above 6.3% (45 mmol/mol), and pre-existing nephropathy (aOR 2.5; 95% CI 1.1–5.3); diabetes type and duration, ethnicity, age, body mass index (BMI), preconception care, smoking and fetal sex were not independently associated [15]. In the Danish national cohort, previous acute diabetic complications increased the relative risk (RR) of congenital heart defects to 7.62 (5.23–10.6), compared with 3.49 (2.91–4.13) without complications [41]. In the US National Birth Defects Prevention Study (NBDPS), the odds ratio (OR) pattern suggested a higher risk of diabetes-associated defects in the absence of periconceptional use of vitamins or supplements containing folic acid [42]. Maternal obesity and older age may contribute to risk in T2D, which is similar to that in T1D in large meta-analyses, both with malformation rates of approximately 5–6% [13].

3. Clinical Spectrum and Manifestations by System

3.1. Magnitude of malformation risk

The malformation risk of pregestational diabetes is broad but concentrated in specific anomaly groups. In Northern England, in a population cohort of 401,149 singleton pregnancies between 1996 and 2008, including 1677 in women with pregestational diabetes, which encompassed live births, fetal deaths from 20 weeks and terminations for anomaly, the prevalence of major non-chromosomal congenital anomaly was 71.6 per 1000 (95% CI 59.6–84.9), with an RR of 3.8 (3.2–4.5) and a 3- to 6-fold higher risk across all common anomaly groups [15]; the Oxford Desk Reference rounds these values to 7% and RR 4 [19]. The European EUROCAT registries (18 population-based registries, 1990–2005, 669 anomaly cases born to mothers with pregestational diabetes versus 92,976 cases without diabetes) showed a specific increase in certain anomalies and multiple anomalies (13.6% versus 6.1%), rather than a uniform increase in all anomalies [18]. In the NBDPS, a US case–control study of deliveries between 1997 and 2011 adjusted for BMI, age, education, race or ethnicity and site, pregestational diabetes was associated with significant ORs, ranging from 2.5 to 80.2, for 46 of the 50 defects studied [17].

The measures in the table cannot be compared directly. Cohort RRs, case–control ORs and birth-certificate RRs measure associations with different denominators and populations, and the very high values for rare anomalies have wide confidence intervals [17, 18, 43]. A high RR also does not mean that most cases of the anomaly have a diabetic mother: only 17% of EUROCAT caudal regression cases arose from pregnancies with pregestational diabetes, and the Oxford Desk Reference stresses that, in absolute numbers, caudal dysplasia is as likely to occur in the infant of a non-diabetic woman as in that of a diabetic woman [18, 19].

3.2. Cardiovascular system

Congenital heart defects are the most frequent component of embryopathy. The Oxford Desk Reference estimates the risk of congenital heart defects at 3–5% in infants of diabetic mothers, including ventricular septal defect (VSD), coarctation of the aorta, transposition of the great arteries (TGA), truncus arteriosus, tricuspid atresia and left isomerism, compared with 7 per 1000 live births (0.7%) in the general population [44]. In the Danish cohort of 2,025,727 births between 1978 and 2011, 7296 (0.36%) exposed to pregestational diabetes, congenital heart defect prevalence was 318 per 10,000 live births (n = 232), versus 80 per 10,000 in the background population, with no variation by birth year, age at diabetes onset or diabetes duration, no difference between T1D and T2D, and an association with all heart defect phenotypes (RR 2.74–13.8) [41]. In Sweden between 2003 and 2015, major heart defect prevalence in singleton live births to mothers with T1D rose from 33 per 1000 with periconceptional HbA1c <6.5% to 101 per 1000 with HbA1c ≥9.1%, versus 15 per 1000 without diabetes [34]. The Brazilian textbook of genetic syndromes lists complete AVSD, mitral atresia and hypoplasia of the left chambers and descending aorta [2], and the Oxford Desk Reference includes maternal diabetes in the history to be taken for hypoplastic left heart [45].

The laterality pattern is characteristic. In the congenital anomaly survey of England's Northern Region, 3 of the 6 reported cases of left isomerism sequence over 3 years were infants of mothers with T1D [36]. Left isomerism, or polysplenia, involves two morphologically left atria; absence of the sinoatrial node may cause complete heart block, atrial septal defect (ASD) or AVSD may be associated, cardiac defects are generally less severe than in right isomerism, multiple small spleens are present, and intestinal malrotation may present as obstruction; isomerism sequence accounts for approximately 1% of congenital heart defects, with an incidence of approximately 1 in 24,000 [10].

Hypertrophic cardiomyopathy in an infant of a diabetic mother is a manifestation of fetopathy rather than a malformation. It results from fetal hyperinsulinism, may cause neonatal myocardial failure and is among the endocrine causes in the differential diagnosis of sarcomeric hypertrophic cardiomyopathy [2, 46].

3.3. Central nervous system

Diabetes-associated NTDs are predominantly cranial. EUROCAT showed increased ORs for anencephaly and encephalocele, but not spina bifida [18], and the ADA lists anencephaly and microcephaly among anomalies whose frequency increases with HbA1c during the first 10 weeks [21]. The Brazilian textbook of genetic syndromes includes myelomeningocele and hydrocephalus in the malformation spectrum and maternal diabetes among environmental factors in CNS malformations [2, 47]. In the United Kingdom, the Confidential Enquiry into Maternal and Child Health (CEMACH) survey found a 4.2-fold excess of NTDs [48]. Holoprosencephaly has an overall risk of approximately 1% in infants of diabetic mothers, approximately 200 times that of the general population according to the Oxford Desk Reference, and an aOR of 13.1 in the NBDPS; the neurological prognosis is very guarded, although many patients survive into childhood and adulthood [17, 49, 50]. Optic nerve hypoplasia has a reported association with maternal diabetes, and Eriksson's review mentions a suggested but unestablished association in female offspring [13, 51].

3.4. Caudal regression syndrome

Caudal regression syndrome is the most characteristic subtype. Smith's Recognizable Patterns of Human Malformation describes incomplete development of the sacrum and, to a lesser extent, lumbar vertebrae, of variable severity; absence of the sacral body produces flattened buttocks, a shortened intergluteal cleft and dimples; disruption of the distal spinal cord causes a neurological deficit ranging from urinary and faecal incontinence to complete neurological loss; lower limb hypomobility causes extreme caudal growth failure; the most severe cases have flexed, abducted hips and popliteal pterygia; talipes equinovarus and calcaneovalgus are common; and unilateral renal agenesis is frequent, with vesicoureteric reflux (VUR) and a risk of impaired renal function [27]. Imperforate anus, cleft lip, cleft palate, microcephaly and myelomeningocele occur occasionally [27]. Smith Fig. 1 (p. 897) shows a newborn with a normal upper trunk and short lower segment, and popliteal pterygia secondary to flexion contractures of neurological origin [27].

The Pang classification, used in the clinical case in section 10, distinguishes two types. In type 1 the conus medullaris ends above the lower border of L1, with a club-shaped termination, the sacral deficiency reaches S1 or higher levels, and bladder and bowel dysfunction is severe; in type 2 the conus ends below L1, the sacrum is better preserved and tethered cord is more frequent [26]. Type 1 survivors generally have normal intelligence but motor deficits with hip dislocation, pelvic deformities, knee flexion contractures and clubfoot [26]. Anal anomalies belong to the same spectrum, ranging from anal imperforation to sacral agenesis and sirenomelia, and most caudal regression cases are sporadic or associated with maternal diabetes [52].

3.5. Urinary tract, digestive tract and abdominal wall

EUROCAT found increased ORs for bilateral renal agenesis and omphalocele [18]. In the Genoa series of 69 patients with caudal regression syndrome, 20.3% had a solitary kidney and 37.7% VUR [53]. The Oxford Desk Reference recommends asking about maternal diabetes in anterior abdominal wall defects and includes maternal diabetes and caudal regression among malformation syndromes to consider in oesophageal and intestinal atresias [54, 55]. VACTERL association (vertebral anomalies, anal atresia, cardiac defects, tracheo-oesophageal fistula, renal and limb anomalies) occurs more frequently in infants of diabetic mothers [40]. US National Vital Statistics System (NVSS) birth certificates also recorded an increase in hypospadias with pregestational diabetes (adjusted RR 1.88) and, to a lesser extent, GDM (1.29) [43].

3.6. Limbs, face and ears

Longitudinal limb deficiency (aOR 10.1 in the NBDPS) and radial hypoplasia or aplasia belong to the spectrum, and the Oxford Desk Reference recommends asking about maternal diabetes in limb reduction defects, radial defects and polydactyly [17, 56–58]. Femoral hypoplasia, isolated or in femoral hypoplasia–unusual facies syndrome, is the most typical limb anomaly: in the series by Johnson and colleagues (1983), the 3 children described were infants of diabetic mothers, one with overt diabetes in the first trimester, with upslanting palpebral fissures, a long philtrum with thin upper lip, micrognathia, hypoplastic nasal alae, cleft palate in 2 and cleft lip in 1, and anomalies of the ears, upper limbs, ribs, vertebrae, lower limbs and genitourinary tract; one died at 24 hours with pulmonary hypoplasia, dysplastic kidneys and a patent ductus arteriosus [35]. The bone dysplasia atlas considers diabetic embryopathy a phenocopy of this syndrome, distinguishable by maternal history and the presence of caudal regression [59].

Cleft palate, especially posterior, and auricular anomalies are part of the history to be taken [2, 60, 61]. EUROCAT found an increased OR for anotia [18]. Oculo-auriculo-vertebral spectrum (OAVS) has a reported association with maternal diabetes: in a series of 30 cases in infants of diabetic mothers (Cedars-Sinai and the literature), 50% had hemifacial microsomia, 46.7% hearing loss, 33.3% facial palsy, 33.3% vertebral anomalies, 36.7% cardiovascular anomalies (45% of these conotruncal), 26.7% renal anomalies, 13.3% limb anomalies (all radial ray hypoplasia), 10% DiGeorge sequence, 6.7% laterality anomalies and 6.7% imperforate anus [62]. In the Spanish ECEMC registry (Spanish Collaborative Study of Congenital Malformations), the OR for OAVS was 2.28 (1.03–4.82) with GDM and 1.50 (0.08–9.99; non-significant) with pregestational diabetes [62]. Smith records a recurrence of approximately 2% in first-degree relatives for OAVS overall [63].

3.7. Fetopathy: fetal growth and neonatal complications

Fetopathy manifests as excessive, disproportionate intrauterine growth and neonatal metabolic and respiratory complications. The following table reproduces the synthesis in the Brazilian textbook of genetic syndromes [2].

Source: tables 9.4.1 and 9.4.2 of the Brazilian textbook of genetic syndromes [2].

Quantitative fetal growth data come from cohorts receiving current treatment rather than natural history. In the UK national database (National Pregnancy in Diabetes audit, NPID, 2014–2018), 52.5% of newborns of mothers with T1D (n = 8690) and 26.2% of those of mothers with T2D (n = 8685) were LGA, and 5.4% and 14.1%, respectively, were small for gestational age (SGA) [14]. The Brazilian textbook reports approximately 30% LGA newborns and 60% caesarean deliveries, without specifying the population or source of these figures [2]. Polyhydramnios occurs in 1–2% of all pregnancies and complicates maternal diabetes, whether pre-existing or acquired in pregnancy [64]. The ADA lists miscarriage, fetal anomalies, pre-eclampsia, fetal death, macrosomia, neonatal hypoglycaemia, hyperbilirubinaemia and respiratory distress syndrome as specific risks of diabetes in pregnancy [21].

3.8. Pregnancy outcomes and differences between diabetes types

In Denmark, in a prospective national cohort of 1218 pregnancies in 990 women with T1D between 1993 and 1999, perinatal mortality was 3.1% versus 0.75% in the population (RR 4.1; 2.9–5.6), stillbirth 2.1% versus 0.45% (RR 4.7; 3.2–7.0) and congenital malformations 5.0% versus 2.8% (RR 1.7; 1.3–2.2); 6 perinatal deaths (16%) were related to malformations, caesarean delivery occurred in 55.9% versus 12.6% and preterm delivery in 41.7% versus 6.0% [65]. In CEMACH (England, Wales and Northern Ireland, deliveries from March 2002 to February 2003, 2359 pregnancies), perinatal mortality was 31.8 per 1000, approximately 4 times that of the general population, and major anomaly prevalence was 46 per 1000, with 48 per 1000 in T1D and 43 per 1000 in T2D [48]. The Balsells meta-analysis (33 studies, 1987–2008) found no significant difference in major malformations between T2D and T1D, but T2D had higher perinatal mortality (OR 1.50; 1.15–1.96) despite lower HbA1c [66]; in NPID, malformations occurred in 4.5% in T1D and 4.0% in T2D (p = 0.17) [14]. Pregnancy loss is more frequent in the first trimester in T1D and the third trimester in T2D [21].

GDM has a much weaker association. In NBDPS, only 12 of 56 defects had significant ORs with GDM, between 1.3 and 2.1 [17], and among the 29,211,974 US live births in 2011–2018 the adjusted RR of congenital anomaly was 1.28 (1.24–1.31) in GDM versus 2.44 (2.33–2.55) in prepregnancy diabetes, with the limitation that the data came from birth certificates, which underdetect anomalies [43]. The Oxford Desk Reference stated in 2017 that there is no malformation risk in GDM [19]; the above data show a small excess, compatible with undiagnosed hyperglycaemia in the first trimester.

3.9. Long-term outcomes and neurodevelopment

Intrauterine exposure to hyperglycaemia increases the risk of obesity, hypertension and T2D in offspring [2, 21]. The 2017 Oxford Desk Reference stated that nothing in the literature raised concern about neurodevelopment in the children of diabetic women [19]. In the Kaiser Permanente Southern California cohort (322,323 singleton children born between 1995 and 2009, followed until 2012, 3388 autism spectrum disorder [ASD] diagnoses), unadjusted annual incidence was 3.26 per 1000 with pre-existing T2D, 3.02 with GDM diagnosed by 26 weeks and 1.77 without diabetes; after full adjustment, the hazard ratio (HR) was 1.21 (0.97–1.52; non-significant) for pre-existing T2D and 1.42 (1.15–1.74) for GDM diagnosed by 26 weeks, with no independent association with antidiabetic medication [67]. The same group published a research letter in 2018 devoted to maternal T1D and autism risk [68]. These data are observational and subject to residual confounding, and do not support causal conclusions.

4. Clinical Recognition and Warning Signs

4.1. Before conception and in early pregnancy

Useful recognition of embryopathy is prospective: identifying, before conception, a woman with diabetes of reproductive age and the risk that her current glycaemia confers on a pregnancy. The ADA recommends integrating preconception counselling into routine care from puberty in all people with diabetes and reproductive potential, and the joint Endocrine Society and European Society of Endocrinology guideline (ES/ESE, 2025) suggests asking about pregnancy intention at all reproductive health, diabetes and primary care visits [14, 21]. Warning signs include unplanned pregnancy in a woman with diabetes, periconceptional HbA1c above 6.5% (48 mmol/mol), especially above 9–10%, diabetic nephropathy, previous acute complications (diabetic ketoacidosis [DKA] or severe hypoglycaemia), absence of folic acid supplementation and use of medicines that should be stopped before conception [15, 20, 22, 41, 42]. Women with previous GDM should be screened for diabetes before conception or before 15 weeks, because undiagnosed T2D exposes organogenesis to hyperglycaemia [21].

4.2. In the prenatal period

Prenatal suspicion arises in two settings. In the first, the pregnant woman has known diabetes and ultrasound identifies a spectrum anomaly, particularly sacral agenesis, a conotruncal heart defect or AVSD, heterotaxy, holoprosencephaly, a cranial NTD, renal agenesis, femoral hypoplasia or multiple anomalies [17, 18]. In the second, ultrasound identifies one of these anomalies in a pregnant woman without known diabetes, requiring investigation for undiagnosed hyperglycaemia and assessment of HbA1c [20, 21]. Polyhydramnios and fetal growth above the 90th centile with disproportionate abdominal circumference suggest fetopathy [2, 64]. In a pregnant woman with GCK monogenic diabetes, fetal abdominal circumference above the 75th centile on second-trimester ultrasound is used as an indicator that the fetus has not inherited the maternal variant [12].

4.3. In newborns and children

In newborns, suspicion rests on the malformation pattern and maternal history. According to the Oxford Desk Reference, anomalies that should prompt questions about maternal diabetes include anorectal anomalies, heart defects and hypoplastic left heart, laterality disorders, holoprosencephaly, cleft palate, auricular anomalies, limb reduction defects, radial defects, polydactyly, abdominal wall defects, oesophageal and intestinal atresias, and optic nerve hypoplasia [10, 44, 45, 49, 51, 52, 54–58, 60, 61]. Preauricular tags or pits associated with maternal diabetes warrant renal ultrasound, which is not indicated in the absence of this or other risk factors [69]. Flattened buttocks, a short intergluteal cleft, sacral dimples, a short lower segment, hip and knee contractures, popliteal pterygia and clubfoot suggest caudal regression [27].

Fetopathy is recognised by macrosomia, neonatal hypoglycaemia, respiratory distress, polycythaemia, jaundice and hypertrophic cardiomyopathy [2]. In most newborns, hypoglycaemia has a non-genetic cause, such as prematurity, low birth weight, sepsis or maternal diabetes, and the Oxford Desk Reference recommends always asking about gestational diabetes; overgrowth with hypoglycaemia in an infant of a diabetic mother carries an increased risk of malformation, especially heart defects, NTDs and skeletal defects [70, 71].

4.4. Emergency warning signs

Neonatal emergencies include hypoglycaemia with abnormal clinical signs, respiratory distress, cardiac decompensation due to congenital heart disease or cardiomyopathy, encephalopathy and polycythaemia likely to require partial exchange transfusion, which the National Institute for Health and Care Excellence (NICE) includes among neonatal unit admission criteria [20]. Neonatal hypoglycaemia may cause seizures and neurological sequelae, and hyperbilirubinaemia may cause kernicterus [2]. Complete heart block in left isomerism and intestinal obstruction due to malrotation are emergencies to recognise in heterotaxy [10]. In pregnant women, DKA is an emergency because pregnancy is a ketogenic state, DKA occurs at lower blood glucose levels than outside pregnancy and carries a high risk of fetal death; NICE recommends urgent ketone testing in any pregnant woman with diabetes who has hyperglycaemia or intercurrent illness, and immediate admission to level 2 care if DKA is suspected [20, 21].

4.5. Newborn screening

There is no population newborn screening for diabetic embryopathy. Targeted screening in infants of diabetic mothers involves routine blood glucose measurement 2 to 4 hours after birth and laboratory tests and investigations guided by clinical signs (sections 5 and 11) [2, 20]. A University of Missouri observational study evaluating spinal ultrasound to screen for sacral agenesis in asymptomatic newborns of mothers with pregestational diabetes (n = 60) ended in December 2024 owing to funding cessation, with no results published in the registry [72].

5. Diagnosis: Criteria, Tests and Algorithms

5.1. Absence of formal criteria

The sources consulted provide no formal diagnostic criteria or confirmatory biomarker for diabetic embryopathy. Diagnosis is clinical and history-based, combining three elements: pregestational maternal diabetes or hyperglycaemia in the periconceptional period, a malformation phenotype compatible with the diabetes-associated spectrum, and reasonable exclusion of other causes, particularly genetic causes [1, 19]. Attribution is probabilistic: even among anomalies with the highest RR, such as caudal regression, a substantial proportion of cases occur without maternal diabetes, and an anomaly in a diabetic mother may have an independent chromosomal or monogenic cause [18, 19].

5.2. Clinical assessment

The Oxford Desk Reference recommends a three-generation pedigree to seek a genetic explanation, a detailed history of periconceptional glycaemic control (HbA1c, hypoglycaemic episodes), a brief enquiry about other teratogenic exposures (medicines, infections, alcohol) and, if the conceptus was liveborn, the perinatal history, birth weight and growth parameters [19]. On examination, when an anomaly has been identified, the parents are also examined, for example their hands and feet in a limb defect and their hearts in a cardiac defect [19].

5.3. Maternal biomarkers and units

HbA1c reflects glycaemic control over the preceding 6 to 8 weeks and is the best available indicator of malformation risk [19]. It is expressed as a percentage (NGSP units, National Glycohemoglobin Standardization Program) and in mmol/mol (IFCC units, International Federation of Clinical Chemistry): 6.0% corresponds to 42 mmol/mol, 6.5% to 48 mmol/mol, 7.0% to 53 mmol/mol and 10% to 86 mmol/mol [20, 21]. For diagnosing diabetes outside pregnancy, HbA1c ≥48 mmol/mol (6.5%) defines diabetes and 43–47 mmol/mol defines impaired glycaemia [11]. During pregnancy, HbA1c falls owing to increased erythrocyte turnover and poorly captures postprandial hyperglycaemia; the ADA therefore uses it as a secondary measure after glucose monitoring, and NICE does not recommend it routinely in the second and third trimesters to assess control [20, 21]. The Brazilian Diabetes Society (SBD) notes that iron deficiency may raise HbA1c by 0.1–0.2% and that the ideal target of 6.0% presupposes a high-performance liquid chromatography (HPLC) method [22]. Glucose in mmol/L is converted to mg/dL by multiplying by 18 (for example, 5.3 mmol/L corresponds to 95 mg/dL) [21].

In a series cited by the Oxford Desk Reference (Wong 2002), periconceptional HbA1c above 9% was associated with 14.3% major anomalies [19]. Maternal serum screening for NTDs requires the laboratory to be informed of diabetes, because alpha-fetoprotein (AFP) reference values must be adjusted [19].

5.4. Fetal ultrasound and echocardiography

NICE recommends offering women with diabetes an ultrasound scan at 20 weeks to detect fetal structural anomalies, including cardiac examination with four-chamber, outflow tract and three-vessel views [20]. The Brazilian protocol described in the textbook of genetic syndromes includes fetal echocardiography at 18–20 weeks and anomaly ultrasound at 20–24 weeks [2]. A systematic review of six guidelines and consensus statements published between 2008 and 2023 classified maternal diabetes mellitus as an indication for fetal echocardiography with partial agreement: all guidelines consider it an indication, but with inconsistent definitions of the population covered [73]. The 2014 American Heart Association (AHA) scientific statement on diagnosis and treatment of fetal cardiac disease establishes fetal echocardiography indications with recommendation classes [74]; the full text, including the class assigned to pregestational diabetes and any HbA1c thresholds, was not accessible for this article.

Ultrasound sensitivity is lower in pregnant women with diabetes. In a series cited by the Oxford Desk Reference (Wong 2002), fetal anomaly detection was 30% in diabetic versus 73% in non-diabetic women, with suboptimal image quality chiefly attributable to maternal obesity [19]. In CEMACH, 65% (71 of 109) of children with major anomalies were diagnosed prenatally, but only 54.8% (23 of 42) of heart defects versus 71.6% (48 of 67) of other anomalies [48]. Fetal magnetic resonance imaging (MRI) better defines the extent of caudal regression, spinal cord termination and renal anomalies, as illustrated by the case in section 10 [26].

5.5. Genetic testing: indications, limitations and reporting

Genetic testing does not confirm embryopathy: it is used to exclude an alternative cause. The Oxford Desk Reference recommends genomic microarray when multiple anomalies are present and, in complex heart defects or those associated with other anomalies or learning difficulties, genomic analysis by array, panel, exome or genome, with urgent fluorescence in situ hybridisation (FISH) for 22q11 in an ill newborn and DNA storage before cardiac surgery or extracorporeal membrane oxygenation (ECMO) [19, 44]. In the published case in section 10, the 46,XY karyotype excluded aneuploidy [26].

The limitations of genetic testing should be explicit in counselling. A normal microarray does not exclude sequence variants, a negative exome does not exclude deep intronic variants, balanced rearrangements or low-level mosaicism, and a negative result does not demonstrate diabetic aetiology, which remains a clinical inference. Conversely, maternal diabetes does not obviate genetic investigation when the phenotype suggests a syndrome with a distinctive pattern, such as Currarino syndrome, X-linked heterotaxy or familial holoprosencephaly [8, 10, 49]. In Currarino syndrome, MNX1 variant detection was 57.4% in a review of more than 60 studies, with no variant identified in 7.7% of familial and 68% of sporadic cases, so a negative result does not exclude the radiological diagnosis [8].

When a variant is identified, the report should indicate the gene, reference transcript with version (NM_005515.4 for MNX1), HGVS (Human Genome Variation Society) nomenclature at coding DNA and protein levels, zygosity, ACMG/AMP (American College of Medical Genetics and Genomics/Association for Molecular Pathology) classification and a recommendation for segregation studies [8, 39]. A rare heterozygous VANGL1 variant should be reported as a low-penetrance susceptibility factor rather than a cause or predictor of malformation [9]. Variants of uncertain significance do not alter clinical attribution and should not be used for prenatal or preimplantation diagnosis.

5.6. Postnatal investigations

The Brazilian textbook of genetic syndromes lists the investigations recommended during pregnancy and after birth [2].

Source: table 9.4.4 of the Brazilian textbook of genetic syndromes [2].

NICE is more selective: it recommends routine blood glucose measurement 2 to 4 hours after birth, tests for polycythaemia, hyperbilirubinaemia, hypocalcaemia or hypomagnesaemia only if clinical signs are present, and echocardiography if there are signs of congenital heart disease or cardiomyopathy, including a murmur, with timing determined clinically [20]. This divergence reflects two practices: Brazilian practice routinely assesses all infants of diabetic mothers, whereas British practice assesses according to signs. If caudal regression is suspected, radiography and MRI of the lumbosacral spine and renal ultrasound define the level of deficiency, spinal cord termination and renal anomalies [26, 27]. In Currarino syndrome, diagnosis is mainly radiological: radiography or computed tomography (CT) of the sacrum, showing a scimitar sacrum or agenesis below S2, and lumbosacral and pelvic MRI to assess the presacral mass, communication with the canal, tethered cord and conus lipoma [8].

5.7. Practical algorithm

Sources: [2, 8, 19, 20, 44, 48].

5.8. Prenatal and preimplantation diagnosis

Prenatal diagnosis of embryopathy is morphological. Invasive diagnosis (chorionic villus sampling or amniocentesis) serves to exclude chromosomal or monogenic anomalies when ultrasound shows multiple anomalies or features suggestive of a genetic syndrome [19, 26]. In GCK diabetes, invasive fetal genotyping solely to guide maternal treatment is not recommended and is replaced by ultrasound assessment of fetal growth [12]. Preimplantation genetic diagnosis has no application in diabetic embryopathy, which is not hereditary; it applies only if a monogenic cause in the differential diagnosis is identified within a family, such as a pathogenic MNX1 variant [8].

6. Differential Diagnosis

6.1. Guiding principle

Differential diagnosis has two aspects. A spectrum anomaly in an infant of a diabetic mother requires exclusion of genetic causes and other teratogens producing the same anomaly; a spectrum anomaly without known maternal diabetes requires exclusion of undiagnosed periconceptional hyperglycaemia [2, 19]. The Brazilian textbook of genetic syndromes summarises this rule by recommending an anomaly-specific differential diagnosis, including non-diabetic aetiologies affecting the same region [2].

Sources: [5, 6, 8, 10–12, 19, 27, 40, 44, 46, 49, 52, 59, 70, 71, 75].

6.2. Caudal regression without diabetes and Currarino syndrome

Caudal regression without maternal diabetes is the most frequent differential diagnosis of the diabetic form: in absolute numbers its frequency is comparable to that associated with diabetes, and only 16% of Smith cases and 17% of EUROCAT cases occurred in infants of diabetic mothers [18, 19, 27]. The sequence is aetiologically heterogeneous and usually sporadic, with rare affected siblings of unaffected parents, including a family with recurrence and maternal diabetes [27].

Currarino syndrome is distinguished by the triad of partial sacral agenesis, a presacral mass and an anorectal malformation. Agenesis typically affects S2 to S5 and takes the form of a hemisacrum or scimitar sacrum; the presacral mass, present in 64% of cases, may be a teratoma, anterior meningocele, rectal duplication or a combination; the anorectal malformation, present in 57%, is generally anal stenosis; and a urogenital malformation is frequently present [27, 52]. The triad is often incomplete; in a series of 29 patients, tethered cord or conus lipoma was present in most and communication between the mass and canal in 12, the teratoma may rarely undergo malignant transformation, and the female-to-male ratio is approximately 1.39 to 1 [8]. The cause is heterozygous MNX1 variants or 7q36 microdeletions, and European prevalence is estimated at 1 in 100,000, without Orphanet validation [8]. Smith writes “MNH1”, a typographical error for the symbol MNX1 [27].

Permanent neonatal diabetes due to biallelic MNX1 variants (approximately 1% of permanent neonatal diabetes) produces a reversed phenocopy: the patient, rather than the mother, has diabetes, with developmental delay, sacral agenesis and imperforate anus [8]. A similar situation occurs in Mitchell-Riley syndrome, due to biallelic RFX6 variants, with neonatal diabetes and gastrointestinal atresias, which the Oxford Desk Reference distinguishes from autosomal recessive Martínez-Frías syndrome [55].

6.3. Sirenomelia

Sirenomelia is characterised by a single lower limb with knees and feet aligned posteriorly, owing to failure of cleavage of the lower limb bud field, absence of the sacrum and other vertebrae, imperforate anus and absent rectum, absent external and internal genitalia, renal agenesis and absent bladder; no two cases are identical [5]. Stevenson's vascular steal theory (1986) explains it through a single large-calibre vessel, derived from the vitelline artery complex and arising from the aorta just below the diaphragm, which replaces the two iliac umbilical arteries and diverts flow from caudal structures; the distal aorta is subordinate and has no renal branches or inferior mesenteric artery [5]. Duhamel's alternative theory regards sirenomelia and caudal dysplasia as extremes of a primary caudal mesoderm deficiency, in which the single umbilical artery would be a consequence rather than a cause; cases of sirenomelia with normal umbilical arteries and caudal dysplasia with an aberrant abdominal umbilical artery support this theory [5]. Twickler and colleagues likewise proposed a distinct vascular mechanism involving an aberrant single umbilical artery derived from the vitelline artery [26]. Maternal diabetes increases the RR of sirenomelia, but most cases occur in infants of non-diabetic women [19]. Smith Fig. 1 (p. 895) shows a stillborn infant with sirenomelia, in which the bones of the single limb range from separate bones to a single broad femur with two distal ossification centres and a broad tibia with two centres [5].

6.4. Femoral hypoplasia–unusual facies

Femoral hypoplasia–unusual facies syndrome, or femoral–facial syndrome, combines short stature due to short lower limbs with facies comprising a prominent forehead, short nose with hypoplastic alae, long philtrum, thin upper lip, micrognathia, cleft palate, upslanting palpebral fissures and low-set malformed ears [6]. Involvement is bilateral and asymmetrical, ranging from a hypoplastic to an absent femur, with variable fibula and tibia, Sprengel deformity and talipes equinovarus; the pelvis has hypoplastic acetabula, a narrow iliac base with a vertical ischial axis and large obturator foramina; there is a dysplastic sacrum, absent vertebrae or hemivertebrae, lumbar sacralisation and scoliosis; and associated findings include cryptorchidism, inguinal hernia, small penis, hypoplastic labia majora, polycystic or absent kidneys and an anomalous collecting system [6]. Occasional findings include astigmatism, esotropia, cleft lip, short metatarsals 3 to 5, preaxial foot polydactyly, tapered, fused or absent ribs, humeral hypoplasia with radioulnar or radiohumeral synostosis, heart defects (VSD, pulmonary stenosis, truncus), craniosynostosis and CNS anomalies such as hydrocephalus, agenesis of the corpus callosum, falx hypoplasia, absent septum pellucidum and colpocephaly [6]. Radiology shows variable femoral hypoplasia, from proximal focal femoral deficiency to mild hypoplasia with bowing, and second-trimester prenatal diagnosis has been described [59]. Maternal diabetes is documented in almost half of more than 50 published cases, and diabetic embryopathy is considered a phenocopy; maternal history and the presence of caudal regression point to diabetic aetiology [6, 59]. Severe forms are indistinguishable from femur-fibula-ulna syndrome, and Roberts syndrome is also in the differential diagnosis [59].

6.5. VACTERL, OEIS, heterotaxy and holoprosencephaly

VACTERL association is more frequent in infants of diabetic mothers but has a heterogeneous aetiology, including ZIC3 and Fanconi anaemia [40]. The diagnosis of VACTERL is secure with at least one anomaly in each of three domains and probable with at least two anomalies in each of two domains, and recurrence risk is 2–3% [44]. OEIS complex (omphalocele, cloacal exstrophy, imperforate anus and spinal defects) is predominantly sporadic [52]. Genetic forms of heterotaxy include X-linked ZIC3 and autosomal LEFTY1 (termed LEFTB in the Oxford Desk Reference), ACVR2B, CFC1 and MMP21, and the sibling recurrence risk for non-syndromic isomerism is 5–10% [10, 44]. In holoprosencephaly, SHH variants explain 37% of families with autosomal dominant inheritance, Pallister-Hall syndrome results from GLI3 variants, and trisomy 13 is the main chromosomal cause [49].

6.6. Other teratogens and maternal diseases

Alcohol, anticonvulsants, lithium, retinoic acid and rubella cause congenital heart defects, as do the maternal diseases phenylketonuria (heart defect risk approximately 15%) and systemic lupus erythematosus (congenital heart block in 20–40% according to the Oxford Desk Reference) [44]. Fetal alcohol spectrum disorders, with their distinctive facial phenotype, enter the differential diagnosis. 22q11.2 deletion syndrome is the main genetic differential diagnosis of conotruncal heart defects [44].

6.7. Overgrowth, hypoglycaemia and cardiomyopathy

The differential diagnosis of macrosomia with neonatal hypoglycaemia includes congenital hyperinsulinism and overgrowth syndromes such as Beckwith-Wiedemann, Sotos, Simpson-Golabi-Behmel, Perlman and Costello syndromes [70, 71, 75]. Monogenic diabetes explains some familial macrosomia: a fetus carrying an HNF4A variant and a fetus without the maternal GCK variant grow excessively without malformation [11, 12]. Transient hypertrophic cardiomyopathy in an infant of a diabetic mother is distinguished from sarcomeric cardiomyopathy by regression in the neonatal period, whereas the sarcomeric form persists and frequently has a family history [2, 46].

7. Treatments and Risk Management Strategies

7.1. Principles

There is no treatment capable of reversing an established malformation or specific antiteratogenic therapy demonstrated in humans; the effective strategy is primary prevention through normalisation of glycaemia before conception and during organogenesis [13, 20, 21]. NICE summarises the evidence by stating that good glycaemic control before conception and during pregnancy reduces the risk of miscarriage, congenital malformation, stillbirth and neonatal death, and that these risks can be reduced but not eliminated [20]. Treatment itself targets fetopathy, neonatal complications and malformations, and the Oxford Desk Reference recommends that all pregnant women with diabetes be followed by an obstetrician experienced in maternal–fetal medicine [2, 19].

7.2. Preconception care

Evidence on preconception care is observational but consistent. The Fuhrmann study (1983) showed 0.8% malformations with intensive control begun before conception versus 7.5% with control begun after 8 weeks [31], a meta-analysis of observational studies cited by the ADA associated preconception care with lower HbA1c and reduced risks of congenital defects, preterm delivery, perinatal mortality, SGA and neonatal intensive care unit (NICU) admission [21], and SBD cites a meta-analysis of 5903 women in which preconception care reduced malformations by 71%, perinatal mortality by 54% and prematurity by 5% [22]. In the Danish 1993–1999 cohort, only 34% of women with T1D performed daily blood glucose self-monitoring at conception and 58% had received preconception guidance, and severe outcomes were associated with higher HbA1c and less self-care [65].

The ADA recommends starting preconception counselling at puberty and continuing it in all people with diabetes and reproductive potential (grade A), prescribing effective contraception, considering long-acting reversible contraceptives (LARCs), until the treatment plan and HbA1c are optimised (A), advising blood glucose as close to normal as possible without excessive hypoglycaemia, ideally HbA1c <6.5% (<48 mmol/mol) (A), and organising interprofessional care involving endocrinology, maternal–fetal medicine, nutrition and diabetes education (B) [21]. ES/ESE suggests asking about pregnancy intention at all visits and contraception when pregnancy is not desired, with non-coercive, person-centred counselling [14]. NICE recommends counselling from adolescence at every contact about the benefits of preconception control, structured education, contraception chosen by the woman according to the UK medical eligibility criteria (UKMEC), and preconception care before stopping contraception [20].

Preconception glycaemic targets vary slightly between sources. The ADA advises preprandial glucose of 80–110 mg/dL (4.4–6.1 mmol/L) and 2-hour postprandial glucose <155 mg/dL (<8.6 mmol/L), with prandial insulin reduction if postprandial glucose falls below 100 mg/dL (5.6 mmol/L), and notes that achieving targets after stopping a medicine generally takes several months [21]. For T1D during pregnancy planning, NICE recommends fasting capillary glucose on waking of 5–7 mmol/L and before meals of 4–7 mmol/L, monthly HbA1c during planning and intensified monitoring when intensifying treatment; it recommends blood ketone test strips and a meter in T1D [20].

Sources: [2, 20–22].

Preconception assessment includes complications and comorbidities. The ADA recommends a dilated retinal examination ideally before pregnancy and in the first trimester, then each trimester and up to 1 year postpartum according to retinopathy severity (B), and a checklist including screening for nephropathy (urinary albumin/creatinine ratio, creatinine), thyroid disease (thyroid-stimulating hormone, TSH), lipid profile, electrocardiography from age 35 if signs, symptoms or risk factors are present, anaemia, recessive disease carrier status and infections, as well as review of obstetric history (previous congenital malformations or fetal losses) [21]. NICE recommends retinal assessment at the first preconception visit unless performed within the preceding 6 months, and delaying rapid glycaemic optimisation until retinal assessment and treatment [20]. The Brazilian textbook of genetic syndromes recommends TSH and thyroid peroxidase antibodies (anti-TPO) in women with T1D planning pregnancy, and SBD includes assessment of subclinical hypothyroidism, retinopathy, renal disease, autonomic neuropathy and atherosclerotic cardiovascular disease, vaccination (measles–mumps–rubella 3 to 6 months beforehand, hepatitis B, tetanus–diphtheria–acellular pertussis [Tdap], respiratory syncytial virus, COVID-19, influenza and pneumococcal vaccines) and discontinuation of teratogenic medicines [2, 22].

Weight and lifestyle are part of preparation. NICE advises weight loss if BMI exceeds 27 kg/m², and SBD if it exceeds 25 kg/m² [20, 22]. The ADA checklist includes nutritional assessment, regular moderate exercise, adequate sleep and avoidance of hyperthermia, for example in hot tubs [21]. ES/ESE considers either a carbohydrate-restricted diet (<175 g/day) or the usual diet (>175 g/day) acceptable, avoiding extremes [14].

7.3. Folic acid and other supplements

Guidelines differ on the folic acid dose. NICE advises women with diabetes planning pregnancy to take folic acid 5 mg/day orally until 12 weeks of gestation to reduce NTD risk, with no maximum dose defined in the source [20]. The ADA recommends prenatal vitamins with at least 400–800 μg folic acid and 150 μg potassium iodide before conception, following the U.S. Preventive Services Task Force (USPSTF), which recommends 0.4–0.8 mg/day for all people planning or potentially able to become pregnant (recommendation A, high certainty of substantial net benefit) [21, 76]. Guideline No. 427 of the Society of Obstetricians and Gynaecologists of Canada (SOGC, 2022) reserves the high dose for those with a previous pregnancy affected by an NTD or another folate-sensitive anomaly and allows personalised approaches, including red cell or serum folate measurement, for complex genetic, medical or surgical risks [77]. The SBD recommends folic acid supplementation until the 12th week to reduce NTDs, without specifying a dose in the text consulted [22]. The Brazilian textbook of genetic syndromes recommends starting 3 months before conception, but the printed dose contains a unit error and is not reproduced [2].

Sources: [20–22, 76, 77].

The 5 mg dose in diabetes is a NICE consensus recommendation, without a randomised trial in the diabetic population; the 400–800 μg dose has evidence from the general population [20, 76]. NBDPS data suggesting higher risk of diabetes-associated defects without periconceptional supplementation support supplementation but do not distinguish doses [42]. In practice, the choice follows the jurisdiction's guideline and individual NTD risk; in Portugal and Brazil, the sources consulted do not establish a diabetes-specific dose.

7.4. Medication review before conception

Medication review is a mandatory component of preconception care [20–22]. The table summarises recommendations by class.

Inadvertent first-trimester ACE inhibitor exposure, after adjustment for confounders, does not appear associated with malformations, supporting discontinuation before or as early as possible in the first trimester without indicating termination of pregnancy for this reason [21]. In a cohort of 469 pregnant women exposed to statins cited by SBD, teratogenicity was not increased, but SGA (RR 1.51; 1.05–2.16) and preterm delivery (RR 1.99; 1.46–2.71) were more frequent [22].

GLP-1 RA discontinuation requires planning. In T2D, ES/ESE suggests stopping these agents before conception rather than between the start of pregnancy and the end of the first trimester (conditional recommendation, very low certainty), because abrupt discontinuation may cause hyperglycaemia and weight gain, increasing malformation and miscarriage risk; it recommends timely transition to and titration of alternatives, with timing individualised according to the likelihood of conception after stopping contraception, the GLP-1 RA type and the risk of a prolonged untreated interval [14]. The panel concluded that anomaly risk with GLP-1 RAs is probably low, based mainly on the Cesta cohort (2024) and limited liraglutide pharmacovigilance data (109 exposed pregnancies, 29.4% miscarriages and 8.3% anomalies, difficult to interpret) [14].

7.5. Treatment during pregnancy

Insulin and antidiabetic medicines

Insulin is the reference treatment. The ADA recommends insulin in T1D (A) and as the preferred medicine in GDM (A) and T2D (B), using multiple daily injections (MDI) or an insulin pump (C) [21]. NICE considers that the rapid-acting analogues aspart and lispro have no demonstrated adverse effect in pregnancy, recommends isophane insulin (NPH) as the first-choice basal insulin, and allows continuation of detemir or glargine in women with good previous control [20]. NICE records that insulin detemir (Levemir®) will be discontinued in the United Kingdom, with supply expected to end in December 2026, requiring transition planning for pregnant women treated with this analogue [20]. Insulin doses are individualised and titrated to glycaemic targets; the sources consulted do not establish weight-based doses. The Brazilian textbook of genetic syndromes reports better perinatal prognosis with a progressive increase in insulin dose in the second half of gestation, based on a Brazilian study [2]. T1D carries increased hypoglycaemia risk in the first trimester, rising again postpartum [21].

The ADA recommends that metformin and glibenclamide (glyburide in the USA) should not be first-line treatments because they cross the placenta (A) and may be insufficient (B) [21]. First-trimester metformin was not associated with increased major malformations and early gestational maternal–fetal transfer is minimal, but in late pregnancy cord levels equal or exceed maternal levels; a primate model showed fetal bioaccumulation, growth restriction and renal dysmorphology, children in the MiG TOFU trial (Auckland) were heavier at 9 years and children of mothers with PCOS had higher BMI at 4 and 5–10 years, whereas MiTy Kids found no anthropometric differences at 24 months; metformin should not be used with hypertension, pre-eclampsia or risk of fetal growth restriction [21]. NICE allows metformin as an adjunct or alternative to insulin before and during pregnancy when benefit outweighs risk, and recommends stopping other oral antidiabetic medicines [20]. ES/ESE suggests not routinely adding metformin to insulin in pregnant women with T2D [14].

Glycaemic targets and monitoring

Sources: [20–22]. For the ADA, either the 1-hour or the 2-hour target suffices, lower limits do not apply to diet-only T2D, and targets may be less stringent if significant hypoglycaemia occurs [21].

ES/ESE suggests not using a single continuous glucose monitoring (CGM) target <140 mg/dL (7.8 mmol/L) in place of standard targets of fasting <95 mg/dL (5.3 mmol/L), 1-hour <140 mg/dL (7.8 mmol/L) and 2-hour <120 mg/dL (6.7 mmol/L), and adjusting insulin according to fasting and postprandial targets rather than the 63–140 mg/dL range alone [14]. The ADA recommends CGM in pregnant women with T1D (A), because it reduces LGA newborns and neonatal hypoglycaemia (A), allows for benefit in other diabetes types (E), and advises against estimated HbA1c or the glucose management indicator in pregnancy (C) [21]. In the CONCEPTT trial, real-time CGM added to standard care in T1D slightly improved HbA1c, increased time in range (TIR) and reduced time above range (TAR) without more hypoglycaemia, with fewer LGA newborns, shorter neonatal hospitalisation and less severe neonatal hypoglycaemia; a 5% increase in TIR was associated with lower neonatal morbidity [21]. NICE offers real-time CGM to all pregnant women with T1D, intermittently scanned (flash) CGM if they cannot use it or prefer this, and considers real-time CGM in insulin-treated pregnant women without T1D who have problematic severe hypoglycaemia or unstable glucose levels [20]. In T2D, ES/ESE considers CGM or capillary self-monitoring acceptable [14], and SBD recommends self-monitoring before and 1 hour after the three main meals, at bedtime and occasionally between 2 and 4 in the morning [22].

Automated insulin delivery

Since its 2026 edition, the ADA recommends automated insulin delivery (AID) systems with pregnancy-specific targets in pregnant women with T1D (grade A), and allows systems without specific targets in selected cases, with assistive techniques and experienced teams (B) [21]. ES/ESE suggests hybrid closed-loop (HCL) delivery in pregnant women with T1D rather than a pump with CGM without an algorithm or MDI with CGM, with a conditional recommendation and very low certainty, and notes that not all algorithms are appropriate in pregnancy [14]. SBD recommends advanced hybrid closed-loop delivery in women with T1D planning pregnancy who do not achieve targets, with an experienced team and a trained pregnant woman (class I, level B) [22]. The difference in recommendation strength between the ADA and ES/ESE reflects different grading criteria applied to the same evidence, mainly the AiDAPT trial (section 7.7).

Ketonaemia, aspirin, hypertension and retina

Pregnancy is a ketogenic state, DKA occurs at lower blood glucose levels and carries a high risk of fetal death, so the ADA recommends ketone strips and education, and NICE urgent blood ketone testing in hyperglycaemia or intercurrent illness [20, 21]. The ADA recommends acetylsalicylic acid 100–150 mg/day from 12–16 weeks of gestation in T1D and T2D to reduce pre-eclampsia (E), considers 162 mg/day acceptable (E), the highest value mentioned in the guideline for pregnant women with T1D or T2D, and notes that US tablets contain 81 mg; adverse effects and duration are not detailed in the text consulted [21]. In chronic hypertension, the ADA recommends starting or titrating treatment at the threshold of <140/90 mmHg (A) [21]. NICE recommends retinal photography with tropicamide mydriasis at the first antenatal visit unless performed within the preceding 3 months, at 16–20 weeks if retinopathy is present and at 28 weeks in all women with pre-existing diabetes; retinopathy does not contraindicate rapid glycaemic optimisation in those presenting with high HbA1c in early pregnancy [20].

Antenatal schedule

NICE recommends booking ideally by 10 weeks at a joint diabetes and obstetric clinic, contact with that clinic every 1–2 weeks throughout pregnancy, confirmation of viability and gestational age at 7–9 weeks, HbA1c and retinal and renal assessment at booking, anomaly ultrasound with cardiac assessment at 20 weeks, and growth and amniotic fluid scans at 28, 32 and 36 weeks [20]. The Brazilian protocol described in the textbook of genetic syndromes adds umbilical Doppler from 26 weeks, abdominal circumference and amniotic fluid index every 4 weeks from 28 weeks, and cardiotocography and biophysical profile from 38 weeks [2]. ES/ESE suggests deciding delivery timing through individual risk assessment rather than expectant management (conditional), in the absence of validated tools, and notes that risks may outweigh the benefits of expectant management beyond 38 weeks even with ideal glycaemia [14].

GCK monogenic diabetes in pregnancy

In pregnant women with GCK diabetes, management depends on fetal genotype. If the fetus has inherited the variant, growth is normal and current recommendations do not support maternal insulin; if it has not, it responds to maternal hyperglycaemia with excess insulin and growth, and the mother is treated with insulin to reduce macrosomia, with limited data and often high doses [12]. Fetal abdominal circumference above the 75th centile on second-trimester ultrasound indicates non-inheritance, invasive fetal genotyping solely for this purpose is not recommended, and some authors advocate early insulin for all, an unresolved issue; the International Society for Pediatric and Adolescent Diabetes (ISPAD) 2022 guideline addresses this scenario [12]. Approximately 0.1% of the population has GCK diabetes, recognisable by clinical features in pregnancy according to the Irish Atlantic DIP cohort [12].

7.6. Comparative regulatory status

No medicine has an approved indication for preventing diabetic embryopathy. Relevant regulatory information concerns medicines whose periconceptional exposure must be managed, chiefly GLP-1 RAs, and availability of insulins and devices.

Sources: [20, 21, 78, 79].

The standard wording of the US semaglutide label estimates background major birth defect risk at 6–10% in women with pregestational diabetes and HbA1c >7%, and reports values up to 20–25% with HbA1c >10%, compared with 2–4% major defects and 15–20% miscarriages in clinically recognised pregnancies in the US general population; the label does not identify the primary source of these values [79].

7.7. Intervention and safety studies

These studies do not permit indirect comparisons. The Fuhrmann study and Diabetes in Early Pregnancy Study are not randomised, and groups differ in motivation and access to care; CGM and AID trials have glycaemic primary outcomes, began after organogenesis (from 16 weeks in AiDAPT) and do not evaluate malformation prevention [31, 32, 80]. The MiniMed 780G trial was negative for its primary outcome in an already well-controlled group, limiting the scope for improvement [22]. In GLP-1 RA safety studies, relative risks near 1 have confidence intervals compatible both with no risk and clinically relevant increases, and confounding by previous glycaemic control persists; Lewin's data are unadjusted [81–83]. A prospective cohort of 168 pregnancies exposed to GLP-1 RAs in the first trimester, cited by SBD, found major defects in 2.6% versus 2.3% in the diabetes group, with no increase in losses after adjustment [22]. In absolute terms, a 6% relative reduction in any anomaly for each 0.5% fall in preconception HbA1c above 6.4%, observed in a Canadian cohort of 3459 births cited by SBD (aRR 0.94; 0.89–0.98), corresponds, with a baseline risk of 5–7%, to approximately 3 to 4 anomalies prevented per 1000 pregnancies [22].

7.8. Developments in the last 12 months

In 2025, ES/ESE published its first joint guideline on pre-existing diabetes and pregnancy, using GRADE (Grading of Recommendations Assessment, Development and Evaluation) methodology, with ten conditional recommendations, nine of very low certainty and one of low certainty, simultaneously published in the Journal of Clinical Endocrinology & Metabolism and the European Journal of Endocrinology, with a correction in November 2025 (guideline) [14]. On 8 December 2025, the ADA published its 2026 Standards of Care, upgrading the recommendation for AID with pregnancy-specific targets in T1D to grade A and introducing a recommendation to discontinue dual GIP/GLP-1 agonists before pregnancy (guideline) [21]. SBD published the 2026 edition of its guideline, with an ideal HbA1c target of 6.0% and a threshold of 9% for advising against pregnancy (guideline) [22].

Regarding GLP-1 RA safety, the target trial emulation by Brown and colleagues (2026) did not show a definitively higher risk with first-trimester continuation, with imprecise estimates for malformations and SGA (observational study) [82]; the analysis by Lewin and colleagues (published online in October 2025) is unadjusted (descriptive observational study) [83]; and the 2026 BJOG narrative review concluded that data show no clear harm signal following inadvertent periconceptional exposure, but are insufficient to exclude clinically important risks, recommending individualised counselling without presuming harm or proven safety (review) [84]. An analysis in the BMJ on outcomes of first-trimester GLP-1 exposure and a Brazilian narrative review on GLP-1 and GLP-1/GIP agonists before, during and after pregnancy were published over the same period [85, 86]. A 2025 JAMA review synthesised automated insulin delivery in pregnancy (review) [87].

Among ongoing trials, AID-DM at Rigshospitalet (Denmark; open-label randomised national multicentre trial, estimated n of 305, started on 1 December 2025, completion expected in October 2028, recruiting) compares CamAPS FX initiated during planning or early pregnancy with usual treatment with CGM in women with T1D, with primary outcomes of TIR (3.5–7.8 mmol/L) and birth weight standard deviation; it is the first randomised AID trial initiated before conception and thus during organogenesis, with no results yet (registered trial) [88]. CIRCUIT (University of Calgary, Canada and Australia; randomised trial, n = 94, Tandem t:slim X2 Control-IQ versus standard insulin with CGM, primary outcome TIR between 16 and 34 weeks) was completed on 23 May 2025, with no registry results or publication located (registered trial) [89]. COSCLIP (University of California, San Francisco; observational, estimated n of 1000) compares commercial and open-source AID systems in T1D pregnancy and has been recruiting since March 2025 (registered study) [90].

In preclinical research, Wentzel and colleagues (2025) showed in rat embryos cultured for 48 hours that high glucose increased hexokinase 2 by 42%, glucose consumption by 75% and methylglyoxal formation by 27%, with dysmorphogenesis, and that a glyoxalase 1 inducer (trans-resveratrol with hesperetin, 10 μM) reversed alterations in crown–rump length, somite number and malformation index by 60%, 49% and 47% (preclinical) [91]. Antioxidants, myo-inositol and glyoxalase inducers have no human clinical trials in diabetic embryopathy, and Ornoy's review had already highlighted the need for such trials in 2007 [13, 38].

7.9. Treatment of the newborn and malformations

NICE recommends early newborn feeding to reduce hypoglycaemia, routine blood glucose measurement 2 to 4 hours after birth, tests for polycythaemia, hyperbilirubinaemia, hypocalcaemia or hypomagnesaemia if clinical signs are present, echocardiography if there are signs of heart disease or cardiomyopathy, and neonatal unit admission for hypoglycaemia with abnormal signs, respiratory distress, cardiac decompensation, encephalopathy or polycythaemia likely to require partial exchange transfusion; the woman should be informed of possible neonatal admission within the first 28 days [20]. The Brazilian textbook of genetic syndromes recommends treatment of congenital anomalies and biochemical abnormalities without detailing protocols [2]; glucose doses for neonatal hypoglycaemia and exchange transfusion criteria follow local neonatal protocols, which the sources consulted do not detail.

In caudal regression syndrome, urological and orthopaedic management is needed in the vast majority of survivors [27]. In the Genoa series (398 patients aged 1 to 37 years, 69 with caudal regression, 244 with myelomeningocele and 85 with spinal cord lipoma), neurogenic bladder was present in 61% of patients with caudal regression, versus 98% and 42% in the other groups, and with intermittent catheterisation and medicines only 30% achieved dry intervals longer than 4 hours, versus 45% and 71% [53]. Follow-up is multidisciplinary, involving physical and rehabilitation medicine (PRM), neurology, neurosurgery, urology and orthopaedics, and delivery of a fetus with severe caudal regression should take place in a tertiary centre [26]. Congenital heart defects are treated according to type, with prior DNA storage for genetic investigation when indicated [44].

When differential diagnosis leads to Currarino syndrome, treatment is surgical after defining the four components (anorectal, sacral, presacral mass and spinal cord); in a series of 29 patients, 25 underwent surgery, 7 in a single stage, with histological assessment of the mass and chronic bowel management, and first-degree relatives should undergo sacral imaging [8]. In neonatal diabetes due to biallelic MNX1, treatment is insulin rather than sulfonylureas, with rehydration, intravenous followed by subcutaneous insulin, a high-calorie diet and developmental support [8].

7.10. Pregnancy, lactation, postpartum care and transition

Agents to avoid in pregnancy are GLP-1 RAs and dual agonists, SGLT2 inhibitors, oral antidiabetic medicines other than metformin, ACE inhibitors, ARBs, mineralocorticoid antagonists, atenolol and, in most cases, statins [20–22]. Semaglutide should not be used during breastfeeding [78]. After delivery, including after pregnancy loss or termination, ES/ESE suggests endocrine care in addition to obstetric care [14]. The ADA notes that many pregnancies are unplanned, recommends regular contraception review, considers LARCs the ideal option and states that the risk of an unplanned pregnancy exceeds that of any available contraceptive method [21]. The transition from adolescence to adulthood is when preconception counselling should begin, from puberty according to the ADA and from adolescence according to NICE, and the ADA provides free counselling resources for adolescents [20, 21].

7.11. Divergences between guidelines

Sources: [14, 20–22].

The American College of Obstetricians and Gynecologists (ACOG) recommendations on pregestational diabetes (Practice Bulletin No. 201, 2018, reaffirmed in 2025) could not be incorporated because the full text was inaccessible; only its existence and reaffirmation are documented [92]. The sources consulted did not include a Portuguese guideline specifically on pregestational diabetes.

8. Reproductive Options and Family Counselling

8.1. Nature of risk and recurrence

Diabetic embryofetopathy is not hereditary. Occurrence and recurrence risk depend on maternal glycaemic control during the periconceptional period of each pregnancy rather than a transmitted genotype [2, 13]. Maternal–fetal genetic predisposition has been demonstrated only in animal models, and men with diabetes do not have an increased risk of offspring with malformations [13]. Counselling should therefore explain that a subsequent pregnancy with HbA1c normalised before conception has a much lower risk than the affected pregnancy, without reaching general population risk [20, 34]. Before attributing an anomaly to diabetes, a genetic explanation must be excluded through the pedigree, parental examination and microarray if multiple anomalies are present [19].

8.2. Risk according to periconceptional HbA1c

Risk is continuous and has no absolutely safe threshold. The table compiles available estimates, with each study's population and measure.

In the Swedish cohort, major non-cardiac anomalies (50 cases) did not increase significantly in any HbA1c category, with limited power and inclusion of live births only [34]. The central counselling message is that of NICE: any reduction of HbA1c towards target is likely to reduce risk, and risk can be reduced but not eliminated [20]. The same Swedish study shows that even with HbA1c <6.5%, major heart defect risk is more than twice that of the population without diabetes [34].

8.3. Content of preconception and post-diagnosis counselling

According to the ADA, preconception counselling addresses the natural history of insulin resistance in pregnancy, preconception targets, prevention of DKA, severe hyperglycaemia and severe hypoglycaemia, retinopathy progression, PCOS, fertility, diabetes genetics and risks of miscarriage, stillbirth, congenital malformations, macrosomia, preterm delivery and hypertensive disorders [21]. NICE adds nausea and vomiting as causes of poor control, LGA newborn risk with birth trauma, induction, instrumental delivery or caesarean section, retinal and nephropathy assessment, early newborn feeding, and obesity and diabetes risk in offspring [20]. After an affected pregnancy, counselling includes explaining aetiology as a clinical inference, the absence of hereditary risk from embryopathy itself, recurrence risk conditioned by HbA1c and a plan for the next pregnancy, with preconception diabetes optimisation according to NICE and maternal–fetal medicine follow-up [19].

8.4. Prenatal diagnosis and options in severe anomalies

Prenatal diagnosis relies on ultrasound at 20 weeks with cardiac assessment, fetal echocardiography and fetal MRI when indicated, with the detection limitations described in section 5 [19, 20, 48]. In a severe anomaly, the options presented in the published case in section 10 were medical termination of pregnancy (MTOP), early induction for comfort care or continuation of pregnancy with delivery in a tertiary centre; confirmatory genetic testing was offered and declined, and the postnatal karyotype was normal [26]. Preimplantation genetic diagnosis is not indicated in diabetic embryopathy and applies only if a monogenic cause in the differential diagnosis is identified [8].

8.5. Risks when the differential diagnosis is genetic

Caudal regression is usually sporadic, with rare affected siblings of unaffected parents [27]. Currarino syndrome is autosomal dominant, with a 50% risk for each child of a carrier, highly variable expressivity and a frequently incomplete triad; first-degree relatives should therefore undergo sacral imaging and, if a variant has been identified in the index case, cascade testing [8]. Neonatal diabetes due to biallelic MNX1 has a 25% recurrence risk in each pregnancy of a carrier couple [8]. A heterozygous VANGL1 variant does not predict malformation and should not be used for prenatal diagnosis [9]. In isolated non-syndromic congenital heart defects, sibling recurrence risk is 5–10% for isomerism, 2% for TGA, 1% for truncus, 3% for VSD and 2% for coarctation; in a Danish study of more than 18,000 patients with congenital heart defects, recurrence was 1.4% for the same defect and 1% for a different defect [44]. These values apply when the heart defect is not attributed to diabetes; in embryopathy, recurrence depends on glycaemic control. VACTERL has a recurrence of 2–3% and OAVS approximately 2% in first-degree relatives [44, 63].

8.6. Maternal monogenic diabetes

In women with GCK diabetes, each child has a 50% probability of inheriting the variant; a fetus that inherits it grows normally, whereas one that does not tends towards macrosomia, influencing maternal treatment [11, 12]. In HNF4A diabetes, a carrier fetus, whether the carrier parent is the mother or father, is at risk of macrosomia and neonatal hypoglycaemia [11]. The Oxford Desk Reference recommends warning families with monogenic diabetes about the fetal risks of poorly controlled diabetes in pregnancy and lists Diabetes UK and Diabetes Genes (www.diabetesgenes.org) resources [11].

8.7. Minors and adolescents

Counselling minors does not involve predictive genetic testing, because embryopathy is not hereditary, but does involve reproductive education: the ADA recommends starting preconception counselling at puberty and NICE from adolescence, at every contact, with information on contraception [20, 21]. In genetic forms within the differential diagnosis, testing asymptomatic minors is justified when it alters surveillance, as with sacral imaging of relatives of patients with Currarino syndrome [8].

9. Special Considerations: Population Diversity and Epidemiology

9.1. Frequency of exposure and embryopathy

Orphanet classifies the worldwide prevalence of diabetic embryopathy as unknown and unvalidated, because no registry counts the entity as such [1]. Frequency is estimated indirectly from pregestational diabetes prevalence and the excess of anomalies in exposed pregnancies. ES/ESE estimates that pre-existing diabetes has doubled over 20 years and affects approximately 1% of pregnancies worldwide, with 0.5% in Europe, 2% in the USA and up to 2.4% in Africa and the Middle East [14]. In the Danish 1978–2011 cohort, 0.36% of births had been exposed to pregestational diabetes [41], and in Northern England 1677 of 401,149 pregnancies (0.42%) occurred in women with pregestational diabetes [15]. In NBDPS, 0.6% of control mothers reported pregestational diabetes, versus 2.5% of mothers of congenital defect cases [17].

9.2. Regional data and what each study measures

Values cannot be compared directly. Birth-certificate studies underdetect anomalies, registries including terminations and fetal deaths capture more cases, case–control studies measure associations rather than prevalence, and the US administrative cohort in ES/ESE has an unusually high baseline malformation rate, suggesting broad coding criteria [14, 15, 17, 43]. Australian caudal regression prevalence includes cases with and without diabetes and should not be read as embryopathy frequency [3]. A 2023 case review reports caudal regression incidence as “1 in 5 per 100,000 live births”, an ambiguous expression in the original, 16–22% of cases associated with maternal diabetes and a 200- to 400-fold higher risk in infants of insulin-dependent women, values drawn from secondary reviews [26].

9.3. Founder populations, ethnicity and socioeconomic status

Diabetic embryopathy has no founder variants, because it is not monogenic. Population differences reflect diabetes prevalence, access to preconception care and glycaemic control. In CEMACH, women with T2D were more often from ethnic minorities (48.8% versus 9.1% in T1D) and deprived areas (46.3% versus 22.8% in the most deprived quintile) [48]. In Northern England, unadjusted anomaly risk was higher in deprived areas and without folate supplementation, although ethnicity was not an independent predictor in the adjusted model [15]. GCK diabetes affects approximately 0.1% of the population in the Irish Atlantic DIP cohort [12].

9.4. Portugal and Brazil

The sources consulted do not include Portuguese epidemiological data on diabetic embryopathy, and the EUROCAT study abstract does not allow confirmation of Portuguese registry participation in the analysis [18]. For patients and families in Portugal, the Associação Protectora dos Diabéticos de Portugal (APDP, https://apdp.pt/) is the reference diabetes patient organisation. In Brazil, the textbook of genetic syndromes estimates congenital anomaly prevalence in infants of diabetic mothers at approximately 3 times that of the general population, dependent on early glycaemic control and varying with population, without specifying a cohort or denominator, and bases its surveillance protocol on Brazilian obstetric practice [2]. The 2026 SBD guideline defines targets and planning for Brazil [22].

9.5. Special populations

Adolescents with diabetes are the group in whom prevention begins, with counselling from puberty [20, 21]. Women with T2D have a similar malformation risk to T1D, higher perinatal mortality, older age and greater adiposity, and increasing periconceptional exposure to GLP-1 RAs, SGLT2 inhibitors and DPP-4 inhibitors [13, 48, 66, 81, 82]. Women with previous GDM require diabetes screening before conception or before 15 weeks, and at GDM diagnosis NICE measures HbA1c to identify pre-existing T2D [20, 21]. Women with nephropathy or previous acute diabetic complications have increased risk [15, 41]. Maternal obesity reduces ultrasound detection of anomalies [19]. Men with diabetes have no increased risk of offspring with malformations [13].

9.6. European Reference Networks and patient organisations

The European Reference Network ERN-ITHACA, devoted to rare malformation syndromes and neurodevelopmental disorders, is relevant for dysmorphological assessment of embryopathies (https://ern-ithaca.eu/) [23]. ERNICA covers rare congenital digestive tract anomalies, including anorectal malformations associated with caudal regression (https://ern-ernica.eu/) [24], and eUROGEN covers rare urogenital diseases. EUROCAT is the European network of population-based registries for epidemiological surveillance of congenital anomalies (https://eu-rd-platform.jrc.ec.europa.eu/eurocat_en) [25]. Patient organisations and resources include the American Diabetes Association (https://diabetes.org/living-with-diabetes/pregnancy), Diabetes UK (https://www.diabetes.org.uk/guide-to-diabetes/life-with-diabetes/pregnancy), Breakthrough T1D (https://www.breakthrought1d.org/), APDP in Portugal (https://apdp.pt/) and the Brazilian Diabetes Society (https://diabetes.org.br/) [21, 22]. The sources consulted did not identify a patient organisation dedicated to sacral agenesis or caudal regression.

10. Illustrative Clinical Case

Ayoub, Dixon, Byrd and Bowker, from Rush University Medical Center in Chicago, published in 2023 a case of type 1 caudal regression syndrome documented by fetal MRI and minimally invasive autopsy using CT and MRI [26]. The pregnant woman was a 29-year-old Caucasian primigravida, of short stature and a smoker, with poorly controlled T2D diagnosed at 22 years, treated with metformin before pregnancy. During pregnancy, the obstetrician replaced metformin with glibenclamide 5 mg twice daily and subsequently 10 mg twice daily, with fasting blood glucose of 78–113 mg/dL and 2-hour postprandial levels of 97–332 mg/dL. At 13 weeks and 3 days, she was admitted for glycaemic control, insulin initiation and dietitian input; baseline HbA1c was 8%, with no earlier values available [26].

The 22-week ultrasound showed severe sacral agenesis, bilateral renal pelvic dilatation, a single umbilical artery and signs of pulmonary hypoplasia. Fetal MRI at 29 weeks showed absence of the lower two-thirds of the spine with a spinal cord anomaly compatible with type 1 caudal regression (Fig. 1a, https://pmc.ncbi.nlm.nih.gov/articles/PMC10675958/figure/Fig1/) and a horseshoe kidney (Fig. 1b), without a heart defect or relevant family history [26]. Counselling presented a very guarded prognosis, with possible fetal death, perinatal death and paraplegia or tetraplegia; confirmatory genetic testing was offered and declined by the woman, and MTOP, early induction for comfort care and pregnancy continuation were discussed. The woman chose to continue, with induction planned at 39 weeks and resuscitation if the newborn was viable. At 37 weeks, fasting glucose was below 90 mg/dL and 2-hour postprandial glucose below 120 mg/dL [26].

On admission for induction at 39 weeks and 3 days, the amniotic fluid index was low and there was no fetal heartbeat. Vaginal delivery resulted in a male stillborn infant weighing 2070 g, below the 1st centile (z −3.3 on Fenton charts), with normal facies and upper limbs, a spine palpable only to approximately two-thirds of its length, very small lower limbs and left clubfoot; the karyotype was 46,XY [26]. Conventional autopsy was declined and minimally invasive autopsy was performed. Post-mortem MRI showed a truncated spinal cord, spine ending at T7 and a horseshoe kidney (Fig. 2, https://pmc.ncbi.nlm.nih.gov/articles/PMC10675958/figure/Fig2/), and CT with three-dimensional reconstruction showed a frog-leg position, with hip flexion and abduction and knee flexion, left clubfoot, hypoplastic iliac wings approximated in the midline and 8 pairs of thoracic ribs, with the 8th fused in the midline below T7 (Fig. 3, https://pmc.ncbi.nlm.nih.gov/articles/PMC10675958/figure/Fig3/) [26]. The authors attributed the aetiology to poorly controlled maternal diabetes in early gestation [26].

The case illustrates three points of this article. First, the decisive exposure occurred before pregnancy recognition and the first intervention, in a woman with T2D without preconception care; HbA1c of 8% lies within a range where anomaly risk is several times baseline [33, 34]. Second, ultrasound identified the anomaly at 22 weeks and fetal MRI characterised it, while imaging autopsy documented the phenotype when conventional autopsy was declined [26]. Third, management in the report diverges from current guidelines: replacing metformin with glibenclamide in poorly controlled T2D conflicts with the ADA recommendation to use insulin as the preferred medicine and not to use glibenclamide first-line, and NICE's recommendation to stop oral antidiabetic medicines except metformin and use insulin; absence of preconception counselling and contraception until optimisation conflicts with all guidelines reviewed in section 7 [14, 20, 21]. The figures of the case report are published under a CC BY 4.0 licence [26].

11. Prognosis and Long-Term Follow-up

11.1. Perinatal prognosis

Congenital malformations are the main cause of perinatal mortality in infants of diabetic mothers according to the Brazilian textbook of genetic syndromes, which considers prognosis good with adequate antenatal control, early diagnosis and treatment [2]. In the Danish 1993–1999 T1D cohort, perinatal mortality was 3.1% versus 0.75% and 16% of perinatal deaths were related to malformations; in CEMACH, perinatal mortality was 31.8 per 1000, approximately 4 times that of the general population [48, 65]. In the Danish study of 933 T1D pregnancies published in 2009, perinatal mortality increased even with periconceptional HbA1c below 6.9% [93]. The Brazilian textbook also reports increased risk of miscarriage, preterm delivery and fetal death, and considers jaundice the most frequent neonatal complication [2]. These data describe patients treated in the 1990s and 2000s, rather than the natural history without care.

11.2. Prognosis by anomaly type

Prognosis depends chiefly on the malformation. In caudal regression, it is poor in the most severe cases, unilateral renal agenesis with VUR carries a risk of renal impairment, and urological and orthopaedic management is needed in the vast majority of survivors [27]. In the Genoa series, 20.3% of patients with caudal regression had a solitary kidney, 37.7% VUR, 8.7% impaired renal function and 61% neurogenic bladder, and only 30% achieved dry intervals longer than 4 hours with intermittent catheterisation and medicines [53]. Survivors of Pang classification type 1 generally have normal intelligence, with motor deficits and severe bladder and bowel dysfunction [26]. In femoral hypoplasia–unusual facies syndrome, intelligence is normal, most patients walk, motor delay is common and speech problems may occur [6]. In holoprosencephaly, the neurological prognosis is very guarded, although many patients survive into childhood and adulthood [50]. Heart defect prognosis depends on type and surgical correction; the Brazilian textbook reports increased cardiovascular mortality in this population [2].

11.3. Long-term prognosis in offspring

Intrauterine exposure to hyperglycaemia is associated with childhood obesity, arterial hypertension, diabetes mellitus and metabolic syndrome throughout life [2, 21]. Regarding neurodevelopment, the association with ASD is more consistent for GDM diagnosed by 26 weeks (adjusted HR 1.42) than for pre-existing T2D (adjusted HR 1.21, non-significant) in the Californian cohort, subject to the limitation of its observational design [67].

11.4. Surveillance

Surveillance of a child with MNX1 neonatal diabetes, a differential diagnosis in this article, includes frequent blood glucose measurements, annual microalbuminuria and cystatin C from age 10, annual ophthalmological examination from age 10 and annual developmental assessment [8].

12. Summary and Final Considerations

Maternal diabetic embryofetopathy is a metabolic teratogen without a causal gene, in which periconceptional hyperglycaemia from pregestational diabetes disrupts organogenesis before the 7th week of gestation and hyperglycaemia in the second half of pregnancy produces fetopathy. The malformation spectrum is specific rather than global: conotruncal heart defects, AVSD, single ventricle and heterotaxy with left isomerism, holoprosencephaly, cranial NTDs, renal agenesis, longitudinal limb deficiency, femoral hypoplasia and, with the highest relative risk, caudal regression with sacral agenesis. Fetopathy adds disproportionate macrosomia, hypoglycaemia, polycythaemia, hypocalcaemia, hyperbilirubinaemia, pulmonary immaturity and transient hypertrophic cardiomyopathy.

Major anomaly risk is 4–7% in European cohorts with pregestational diabetes, 1.7 to 3.8 times population risk, without a difference between T1D and T2D, and rises continuously with periconceptional HbA1c, without a safe threshold: even with HbA1c below 6.5%, major heart defect risk remains twice as high. GDM diagnosed after the first trimester has a small malformative effect, chiefly reflecting undiagnosed T2D. A high relative risk does not mean that most cases of an anomaly have a diabetic mother: in caudal regression, only a minority of cases arise from pregnancy with diabetes, and MNX1-related Currarino syndrome, ZIC3-linked heterotaxy and genetic holoprosencephalies must be excluded.

Diagnosis is clinical and history-based, without a confirmatory biomarker, and genetic testing serves to exclude alternative causes. Ultrasound has lower sensitivity in pregnant women with diabetes, especially for heart defects and with obesity, justifying ultrasound at 20 weeks with cardiac assessment and fetal echocardiography. Prevention rests on preconception care starting in adolescence: contraception until optimisation, HbA1c <6.5% (48 mmol/mol) before conception, folic acid (5 mg/day according to NICE, 400–800 μg/day according to the ADA), discontinuation of ACE inhibitors, ARBs, statins, GLP-1 RAs and SGLT2 inhibitors, and insulin with CGM and, in T1D, automated systems with pregnancy-specific targets. Technology trials improved glycaemic control from the second trimester, but no trial has yet demonstrated malformation prevention; the first randomised AID trial initiated before conception is ongoing. Data on periconceptional GLP-1 RAs show no clear teratogenicity signal but are imprecise. Recurrence depends on glycaemic control in the next pregnancy rather than inheritance, and counselling should convey that risk can be reduced but not eliminated.

13. Bibliographic References

1. Orphanet. Diabetic embryopathy (ORPHA:1926). https://www.orpha.net/en/disease/detail/1926. Accessed 9 Oct 2026.
2. Pinheiro SHBCS, Ribeiro EM. Síndrome do filho de mãe diabética. In: Ribeiro EM, coordinator. Síndromes Genéticas: Reconhecer para Tratar. Fortaleza: EdUnichristus; 2021. p. 184–6.
3. Orphanet. Caudal regression syndrome (ORPHA:3027). https://www.orpha.net/en/disease/detail/3027. Accessed 9 Oct 2026.
4. Orphanet. Caudal regression-sirenomelia spectrum (ORPHA:444941). https://www.orpha.net/en/disease/detail/444941. Accessed 9 Oct 2026.
5. Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. Sirenomelia sequence; p. 894–5.
6. Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. Femoral hypoplasia–unusual facies syndrome; p. 438–9.
7. World Health Organization. ICD-10 Version:2019. P70 Transitory disorders of carbohydrate metabolism specific to fetus and newborn. Available from: https://icd.who.int/browse10/2019/en#/P70. Accessed 9 Oct 2026.
8. GeneGuidance. MNX1 — clinical curation (local geneguidance.db database). Accessed 9 Oct 2026. Primary references: Dworschak 2021, PMID 33836786; Cretolle 2006, PMID 16410121; De Leon DD, Pinney SE. Permanent Neonatal Diabetes Mellitus. GeneReviews, NBK1447 (revision of 25 Sep 2025).
9. GeneGuidance. VANGL1 — clinical curation (local geneguidance.db database). Accessed 9 Oct 2026. Primary references: PMID 17409324, 19319979, 25208524, 28007035.
10. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Laterality disorders including heterotaxy and isomerism; p. 206–7.
11. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Diabetes mellitus; p. 390–3.
12. GeneGuidance. GCK — clinical curation (local geneguidance.db database). Accessed 9 Oct 2026. Primary references: PMID 26106223, 24092492, 24430320, 24550216, 11483936; ISPAD 2022 guideline, PMID 36537518.
13. Eriksson UJ, Wentzel P. The status of diabetic embryopathy. Ups J Med Sci. 2016;121(2):96-112. doi:10.3109/03009734.2016.1165317. PMID: 27117607. PMC4900070.
14. Wyckoff JA, Lapolla A, Asias-Dinh BD, Barbour LA, Brown FM, Catalano PM, et al. Preexisting Diabetes and Pregnancy: An Endocrine Society and European Society of Endocrinology Joint Clinical Practice Guideline. J Clin Endocrinol Metab. 2025;110(9):2405-2452. doi:10.1210/clinem/dgaf288. PMID: 40652453. PMC12342393.
15. 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;55(4):936-47. doi:10.1007/s00125-012-2455-y. PMID: 22314812.
16. Mills JL, Baker L, Goldman AS. Malformations in infants of diabetic mothers occur before the seventh gestational week. Implications for treatment. Diabetes. 1979;28(4):292-3. doi:10.2337/diab.28.4.292. PMID: 437367.
17. Tinker SC, Gilboa SM, Moore CA, Waller DK, Simeone RM, Kim SY, et al. Specific birth defects in pregnancies of women with diabetes: National Birth Defects Prevention Study, 1997-2011. Am J Obstet Gynecol. 2020;222(2):176.e1-176.e11. doi:10.1016/j.ajog.2019.08.028. PMID: 31454511. PMC7186569.
18. Garne E, Loane M, Dolk H, Barisic I, Addor MC, Arriola L, et al. Spectrum of congenital anomalies in pregnancies with pregestational diabetes. Birth Defects Res A Clin Mol Teratol. 2012;94(3):134-40. doi:10.1002/bdra.22886. PMID: 22371321.
19. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Maternal diabetes mellitus and diabetic embryopathy; p. 756–7.
20. National Institute for Health and Care Excellence. Diabetes in pregnancy: management from preconception to the postnatal period. NICE guideline NG3. London: NICE; 2015 [updated 2020 Dec 16]. Available from: https://www.nice.org.uk/guidance/ng3. Accessed 9 Oct 2026.
21. American Diabetes Association Professional Practice Committee for Diabetes. 15. Management of Diabetes in Pregnancy: Standards of Care in Diabetes-2026. Diabetes Care. 2026;49(Suppl 1):S321-S338. doi:10.2337/dc26-S015. PMID: 41358885. PMC12690181.
22. Zajdenverg L, Lauand TCG, Façanha C, Dualib PM, Reichelt AJ, Araujo CAR, et al. Planejamento, metas e monitorização do diabetes durante a gestação. In: Bertoluci M, editor. Diretriz da Sociedade Brasileira de Diabetes – Edição 2026. São Paulo: SBD; 2026. doi:10.29327/5937145.2026-1. Available from: https://diretriz.diabetes.org.br/planejamento-metas-e-monitorizacao-do-diabetes-durante-a-gestacao/. Accessed 9 Oct 2026.
23. ERN-ITHACA – European Reference Network on Rare Congenital Malformations and Rare Intellectual Disability. Available from: https://ern-ithaca.eu/. Accessed 9 Oct 2026.
24. ERN ERNICA – European Reference Network for rare Inherited and Congenital (digestive and gastrointestinal) Anomalies. Available from: https://ern-ernica.eu/. Accessed 9 Oct 2026.
25. European Commission Joint Research Centre. EUROCAT – European network of population-based registries for the epidemiological surveillance of congenital anomalies. Available from: https://eu-rd-platform.jrc.ec.europa.eu/eurocat_en. Accessed 9 Oct 2026.
26. Ayoub M, Dixon C, Byrd SE, Bowker RM. Caudal regression syndrome type 1 with minimally invasive computed tomography and magnetic resonance imaging autopsy: a case report. J Med Case Rep. 2023;17(1):491. doi:10.1186/s13256-023-04220-5. PMID: 38001547. PMC10675958.
27. Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. Caudal dysplasia sequence (caudal regression syndrome); p. 896–7.
28. Passarge E, Lenz W. Syndrome of caudal regression in infants of diabetic mothers: observations of further cases. Pediatrics. 1966;37(4):672-5. PMID: 5930030.
29. Kucera J. Rate and type of congenital anomalies among offspring of diabetic women. J Reprod Med. 1971;7(2):73-82. PMID: 5095696.
30. Miller E, Hare JW, Cloherty JP, Dunn PJ, Gleason RE, Soeldner JS, et al. Elevated maternal hemoglobin A1c in early pregnancy and major congenital anomalies in infants of diabetic mothers. N Engl J Med. 1981;304(22):1331-4. doi:10.1056/NEJM198105283042204. PMID: 7012627.
31. Fuhrmann K, Reiher H, Semmler K, Fischer F, Fischer M, Glöckner E. Prevention of congenital malformations in infants of insulin-dependent diabetic mothers. Diabetes Care. 1983;6(3):219-23. doi:10.2337/diacare.6.3.219. PMID: 6347574.
32. Mills JL, Knopp RH, Simpson JL, Jovanovic-Peterson L, Metzger BE, Holmes LB, et al. Lack of relation of increased malformation rates in infants of diabetic mothers to glycemic control during organogenesis. N Engl J Med. 1988;318(11):671-6. doi:10.1056/NEJM198803173181104. PMID: 3344018.
33. Guerin A, Nisenbaum R, Ray JG. Use of maternal GHb concentration to estimate the risk of congenital anomalies in the offspring of women with prepregnancy diabetes. Diabetes Care. 2007;30(7):1920-5. doi:10.2337/dc07-0278. PMID: 17446531.
34. Ludvigsson JF, Neovius M, Söderling J, Gudbjörnsdottir S, Svensson AM, Franzén S, et al. Periconception glycaemic control in women with type 1 diabetes and risk of major birth defects: population based cohort study in Sweden. BMJ. 2018;362:k2638. doi:10.1136/bmj.k2638. PMID: 29976596. PMC6031927.
35. Johnson JP, Carey JC, Gooch WM 3rd, Petersen J, Beattie JF. Femoral hypoplasia-unusual facies syndrome in infants of diabetic mothers. J Pediatr. 1983;102(6):866-72. doi:10.1016/s0022-3476(83)80013-4. PMID: 6854450.
36. Splitt M, Wright C, Sen D, Goodship J. Left-isomerism sequence and maternal type-1 diabetes. Lancet. 1999;354(9175):305-6. doi:10.1016/S0140-6736(99)01942-X. PMID: 10440314.
37. Dong D, Reece EA, Lin X, Wu Y, AriasVillela N, Yang P. New development of the yolk sac theory in diabetic embryopathy: molecular mechanism and link to structural birth defects. Am J Obstet Gynecol. 2016;214(2):192-202. doi:10.1016/j.ajog.2015.09.082. PMID: 26432466. PMC4744545.
38. Ornoy A. Embryonic oxidative stress as a mechanism of teratogenesis with special emphasis on diabetic embryopathy. Reprod Toxicol. 2007;24(1):31-41. doi:10.1016/j.reprotox.2007.04.004. PMID: 17548185.
39. Ensembl. MNX1 (ENSG00000130675), transcript ENST00000252971.11 (MANE Select; RefSeq NM_005515.4). Available from: https://www.ensembl.org/Homo_sapiens/Gene/Summary?g=ENSG00000130675. Accessed 9 Oct 2026.
40. Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. VACTERL association; p. 922–5.
41. Øyen N, Diaz LJ, Leirgul E, Boyd HA, Priest J, Mathiesen ER, et al. Prepregnancy Diabetes and Offspring Risk of Congenital Heart Disease: A Nationwide Cohort Study. Circulation. 2016;133(23):2243-53. doi:10.1161/CIRCULATIONAHA.115.017465. PMID: 27166384. PMC4890838.
42. Correa A, Gilboa SM, Botto LD, Moore CA, Hobbs CA, Cleves MA, et al. Lack of periconceptional vitamins or supplements that contain folic acid and diabetes mellitus-associated birth defects. Am J Obstet Gynecol. 2012;206(3):218.e1-13. doi:10.1016/j.ajog.2011.12.018. PMID: 22284962. PMC4915337.
43. Wu Y, Liu B, Sun Y, Du Y, Santillan MK, Santillan DA, et al. Association of Maternal Prepregnancy Diabetes and Gestational Diabetes Mellitus With Congenital Anomalies of the Newborn. Diabetes Care. 2020;43(12):2983-2990. doi:10.2337/dc20-0261. PMID: 33087319. PMC7770264.
44. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Congenital heart disease; p. 112–4.
45. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Hypoplastic left heart; p. 740–1.
46. Christiaans I, Elliott PM, Michels M. Hypertrophic cardiomyopathy. In: Baars HF, van der Smagt JJ, Doevendans PAFM, van Tintelen JP, editors. Clinical Cardiogenetics. 3rd ed. Cham: Springer; 2020. Chapter 4, Table 4.2, p. 67. doi:10.1007/978-3-030-45457-9_4
47. Ribeiro EM, coordinator. Síndromes Genéticas: Reconhecer para Tratar. Fortaleza: EdUnichristus; 2021. Malformações do SNC; p. 289–90.
48. Macintosh MC, Fleming KM, Bailey JA, Doyle P, Modder J, Acolet D, et al. Perinatal mortality and congenital anomalies in babies of women with type 1 or type 2 diabetes in England, Wales, and Northern Ireland: population based study. BMJ. 2006;333(7560):177. doi:10.1136/bmj.38856.692986.AE. PMID: 16782722. PMC1513435.
49. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Holoprosencephaly (HPE); p. 168–71.
50. Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. Holoprosencephaly sequence; p. 870–3.
51. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Optic nerve hypoplasia; p. 270–1.
52. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Anal anomalies (atresia, stenosis); p. 58–61.
53. Torre M, Buffa P, Jasonni V, Cama A. Long-term urologic outcome in patients with caudal regression syndrome, compared with meningomyelocele and spinal cord lipoma. J Pediatr Surg. 2008;43(3):530-3. doi:10.1016/j.jpedsurg.2007.10.036. PMID: 18358295.
54. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Anterior abdominal wall defects; p. 704–5.
55. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Oesophageal and intestinal atresia; p. 266–9.
56. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Radial ray defects and thumb hypoplasia; p. 296–9.
57. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Limb reduction defects; p. 212–5.
58. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Polydactyly; p. 284–7.
59. Spranger JW, Brill PW, Hall C, Nishimura G, Superti-Furga A, Unger S. Bone Dysplasias: An Atlas of Genetic Disorders of Skeletal Development. 4th ed. Oxford University Press; 2018. Femoral-facial syndrome; p. 876–9.
60. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Cleft lip and palate; p. 102–5.
61. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Ear anomalies; p. 142–5.
62. Wang R, Martínez-Frías ML, Graham JM Jr. Infants of diabetic mothers are at increased risk for the oculo-auriculo-vertebral sequence: A case-based and case-control approach. J Pediatr. 2002;141(5):611-7. doi:10.1067/mpd.2002.128891. PMID: 12410187.
63. Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. Oculo-auriculo-vertebral spectrum; p. 904–7.
64. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Polyhydramnios; p. 774–5.
65. Jensen DM, Damm P, Moelsted-Pedersen L, Ovesen P, Westergaard JG, Moeller M, et al. Outcomes in type 1 diabetic pregnancies: a nationwide, population-based study. Diabetes Care. 2004;27(12):2819-23. doi:10.2337/diacare.27.12.2819. PMID: 15562191.
66. Balsells M, García-Patterson A, Gich I, Corcoy R. Maternal and fetal outcome in women with type 2 versus type 1 diabetes mellitus: a systematic review and metaanalysis. J Clin Endocrinol Metab. 2009;94(11):4284-91. doi:10.1210/jc.2009-1231. PMID: 19808847.
67. Xiang AH, Wang X, Martinez MP, Walthall JC, Curry ES, Page K, et al. Association of maternal diabetes with autism in offspring. JAMA. 2015;313(14):1425-34. doi:10.1001/jama.2015.2707. PMID: 25871668.
68. Xiang AH, Wang X, Martinez MP, Page K, Buchanan TA, Feldman RK. Maternal Type 1 Diabetes and Risk of Autism in Offspring. JAMA. 2018;320(1):89-91. doi:10.1001/jama.2018.7614. PMID: 29936530. PMC6134431.
69. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Minor congenital anomalies; p. 240–1.
70. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Hypoglycaemia in neonate and infant; p. 182–5.
71. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Overgrowth; p. 272–5.
72. University of Missouri-Columbia. Screening for Sacral Agenesis in Offspring of Mothers With Diabetes in Pregnancy. ClinicalTrials.gov identifier NCT05033275. Available from: https://clinicaltrials.gov/study/NCT05033275. Accessed 9 Oct 2026.
73. De Robertis V, Stampalija T, Abuhamad AZ, Bosco M, Chaoui R, Formigoni C, et al. Indications for fetal echocardiography: consensus and controversies among evidence-based national and international guidelines. Ultrasound Obstet Gynecol. 2025;65(6):682-693. doi:10.1002/uog.29224. PMID: 40208627.
74. Donofrio MT, Moon-Grady AJ, Hornberger LK, Copel JA, Sklansky MS, Abuhamad A, et al. Diagnosis and treatment of fetal cardiac disease: a scientific statement from the American Heart Association. Circulation. 2014;129(21):2183-242. doi:10.1161/01.cir.0000437597.44550.5d. PMID: 24763516.
75. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Beckwith–Wiedemann syndrome (BWS); p. 358–61.
76. US Preventive Services Task Force; Barry MJ, Nicholson WK, Silverstein M, Chelmow D, Coker TR, Davis EM, et al. Folic Acid Supplementation to Prevent Neural Tube Defects: US Preventive Services Task Force Reaffirmation Recommendation Statement. JAMA. 2023;330(5):454-459. doi:10.1001/jama.2023.12876. PMID: 37526713.
77. Wilson RD, O'Connor DL. Guideline No. 427: Folic Acid and Multivitamin Supplementation for Prevention of Folic Acid-Sensitive Congenital Anomalies. J Obstet Gynaecol Can. 2022;44(6):707-719.e1. doi:10.1016/j.jogc.2022.04.004. PMID: 35691683.
78. European Medicines Agency. Ozempic (semaglutide): EPAR – Product information (Summary of Product Characteristics), section 4.6. Amsterdam: EMA; first authorised 2018 Feb 8. Available from: https://www.ema.europa.eu/en/documents/product-information/ozempic-epar-product-information_en.pdf. Accessed 9 Oct 2026.
79. Novo Nordisk. RYBELSUS / OZEMPIC (semaglutide) tablets. Prescribing information, section 8.1 Pregnancy (NDA213051; label effective 2026-01-30). DailyMed setid 27f15fac-7d98-4114-a2ec-92494a91da98. Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=27f15fac-7d98-4114-a2ec-92494a91da98. Accessed 9 Oct 2026.
80. Lee TTM, Collett C, Bergford S, Hartnell S, Scott EM, Lindsay RS, et al. Automated Insulin Delivery in Women with Pregnancy Complicated by Type 1 Diabetes. N Engl J Med. 2023;389(17):1566-1578. doi:10.1056/NEJMoa2303911. PMID: 37796241.
81. Cesta CE, Rotem R, Bateman BT, Chodick G, Cohen JM, Furu K, et al. Safety of GLP-1 Receptor Agonists and Other Second-Line Antidiabetics in Early Pregnancy. JAMA Intern Med. 2024;184(2):144-152. doi:10.1001/jamainternmed.2023.6663. PMID: 38079178. PMC10714281.
82. Brown JP, Huybrechts KF, Straub L, Patorno E, Seely EW, Bateman BT, et al. Continuing Glucagon-Like Peptide-1 Receptor Agonists Into the First Trimester of Pregnancy and Pregnancy Outcomes : A Target Trial Emulation Study Using Claims Information. Ann Intern Med. 2026;179(7):929-937. doi:10.7326/ANNALS-25-04820. PMID: 42258827. PMC13430612.
83. Lewin Z, Snow S, Lee JA, Wilkie G. Glucagon-Like Peptide-1 Receptor Agonists among Pregnancies with Pregestational Diabetes and Its Relationship with Congenital Malformations. Am J Perinatol. 2026;43(7):984-988. doi:10.1055/a-2716-1639. PMID: 41052622.
84. Lau KGY, Chopra A, Mullins E, Agha-Jaffar R, Tan T, Sykes L, et al. Glucagon-Like Peptide-1 Receptor Agonists in Pregnancy: Periconception Exposure and Perinatal Outcomes. BJOG. 2026;. doi:10.1111/1471-0528.70339. PMID: 42830218.
85. Heald A, Reynolds RM, Ozanne SE, Dearden L, Wu P. Pregnancy outcomes of GLP-1 exposure in the first trimester. BMJ. 2026;394:e100237. doi:10.1136/bmj-2026-100237. PMID: 42476613.
86. Zajdenverg L, Santos VM, Rodrigues ACM, Previdelli LK, Negrato CA. Impact of GLP-1 and GLP-1/GIP receptor agonists before, during, and after pregnancy: a narrative review. Diabetol Metab Syndr. 2026;18(1):205. doi:10.1186/s13098-026-02235-8. PMID: 42754900. PMC13584491.
87. Polsky S. Automated Insulin Delivery in Pregnancy. JAMA. 2025;334(24):2165-2167. doi:10.1001/jama.2025.19201. PMID: 41134584.
88. Copenhagen University Hospital, Rigshospitalet. Automated Insulin Delivery Versus Usual Insulin Treatment Modality Before and During Pregnancy in Women With Type 1 Diabetes (AID-DM). ClinicalTrials.gov identifier NCT07240012. Available from: https://clinicaltrials.gov/study/NCT07240012. Accessed 9 Oct 2026.
89. University of Calgary. Closed-loop Insulin Delivery by Glucose Responsive Computer Algorithms In Type 1 Diabetes Pregnancies (CIRCUIT). ClinicalTrials.gov identifier NCT04902378. Available from: https://clinicaltrials.gov/study/NCT04902378. Accessed 9 Oct 2026.
90. University of California, San Francisco. Commercial or Open Source Closed Loop Impact on Pregnancy (COSCLIP) Study. ClinicalTrials.gov identifier NCT06654713. Available from: https://clinicaltrials.gov/study/NCT06654713. Accessed 9 Oct 2026.
91. Wentzel P, Xue M, Rabbani N, Eriksson UJ, Thornalley PJ. Overload of Glucose Metabolism as Initiating Factor in Diabetic Embryopathy and Prevention by Glyoxalase 1 Inducer Dietary Supplement. Antioxidants (Basel). 2025;14(8):1022. doi:10.3390/antiox14081022. PMID: 40867918. PMC12383083.
92. ACOG Practice Bulletin No. 201: Pregestational Diabetes Mellitus. Obstet Gynecol. 2018;132(6):e228-e248. doi:10.1097/AOG.0000000000002960. PMID: 30461693.
93. Jensen DM, Korsholm L, Ovesen P, Beck-Nielsen H, Moelsted-Pedersen L, Westergaard JG, et al. Peri-conceptional A1C and risk of serious adverse pregnancy outcome in 933 women with type 1 diabetes. Diabetes Care. 2009;32(6):1046-8. doi:10.2337/dc08-2061. PMID: 19265024. PMC2681038.