Syndromic Neonatal Diabetes (Pancreatic Agenesis, Mitchell-Riley, MEDS, GLIS3 and NEUROG3)

Class of monogenic disorders in which diabetes in the first months of life results from a failure of development of the pancreas and other organs of endodermal origin, or from a disturbance of secretory trafficking shared by the β cell and the neuron; isolated neonatal diabetes, Wolcott-Rallison syndrome and IPEX syndrome enter the differential diagnosis.

Literature and regulatory information updated: 31 October 2026.

0. Initial Reference Table

The class brings together entities with two features in common: insulin-dependent diabetes with onset, in most patients, before six months of age, and extrapancreatic manifestations that define each syndrome and that frequently become evident only after the diagnosis of diabetes [1–3]. The first table identifies each entity; the second gathers the data common to the class. Abbreviations used in the tables: AD, autosomal dominant; AR, autosomal recessive; SD, standard deviation; ISPAD, International Society for Paediatric and Adolescent Diabetes; IUGR, intrauterine growth restriction.

1. Introduction

1.1. Definition of the class and scope

Neonatal diabetes (ND) is defined by onset of hyperglycaemia in the first six months of life, an age at which polygenic autoimmune type 1 diabetes (T1D) is exceptional; the permanent form (PNDM, permanent neonatal diabetes mellitus) appears, on average, at 7 weeks (from birth to 26 weeks) and is associated with partial or complete insulin deficiency, already present in utero, which explains the frequent intrauterine growth restriction (IUGR) at birth [1]. As some patients are only diagnosed between 3 and 6 months, the terms "diabetes of infancy" and "congenital diabetes" have been proposed [1]. ND is syndromic when the same genetic alteration affects other organs: the Oxford Desk Reference already noted that the different PNDM subtypes have distinct associated features and that the genetic diagnosis is increasingly made before those features are recognised [51].

This article covers the class of syndromic forms whose mechanism is abnormal development of the pancreas and of neighbouring endodermal organs (pancreatic agenesis due to GATA6, GATA4, PDX1, PTF1A, CNOT1 and ZNF808; Mitchell-Riley syndrome due to RFX6; GLIS3 syndrome; enteric anendocrinosis due to NEUROG3) or disturbance of transport from the endoplasmic reticulum (ER) to the Golgi apparatus in the β cell and the neuron (microcephaly, epilepsy and diabetes syndromes, MEDS, due to IER3IP1, YIPF5 and TMEM167A). Wolcott-Rallison syndrome (WRS, EIF2AK3), the most frequent syndromic form in consanguineous populations, is included in the class table; the forms due to ATP-sensitive potassium (KATP) channels, the 6q24 locus and the insulin gene, IPEX syndrome (FOXP3) and the other monogenic autoimmune forms, and isolated heart disease due to GATA4 and GATA6 enter the differential diagnosis.

1.2. Historical overview and milestones of molecular discovery

In 1972, Wolcott and Rallison described the association of infancy-onset diabetes with multiple epiphyseal dysplasia, and in 2000 Delépine and colleagues identified the EIF2AK3 gene, according to the bibliography of the bone dysplasia atlas [36]. In 1992, Martínez-Frías and colleagues described two siblings born to consanguineous Spanish Roma parents with hypoplastic pancreas, duodenal atresia and atresia of the extrahepatic bile ducts, one of them with tracheo-oesophageal fistula [52]. In 2004, Mitchell and colleagues brought together five infants with ND, hypoplastic or annular pancreas, duodenal and jejunal atresia and absent gallbladder, whose pancreatic immunohistochemistry showed scant clusters of chromogranin A-positive cells and absence of insulin, glucagon and somatostatin; Chappell and colleagues proposed in 2008 that these cases and that of Martínez-Frías were a single syndrome with six key features [21]. In 2010, Smith and colleagues showed that the transcription factor Rfx6 directs islet differentiation downstream of Neurog3 and identified, by homozygosity mapping, RFX6 variants in these infants [21, 53].

In 1999, Hoveyda and colleagues described ND with cerebellar agenesis and dysmorphism in Pakistani patients, and in 2004 Sellick and colleagues identified truncating PTF1A variants in that family and in a family from northern Europe [13, 14]. In 2003, Taha and colleagues described two Saudi siblings with ND, congenital hypothyroidism, congenital glaucoma, hepatic fibrosis and polycystic kidneys, and in 2006 Senée and colleagues attributed the syndrome to GLIS3 [24, 25]. In the same year, Wang and colleagues linked congenital malabsorptive diarrhoea without enteroendocrine cells to NEUROG3 variants [54]. In 2011, Allen and colleagues identified de novo GATA6 variants in 15 of 27 patients with pancreatic agenesis, the most frequent cause of this malformation, and Poulton and colleagues described the syndrome of microcephaly with simplified gyration, epilepsy and infantile diabetes caused by IER3IP1 [30, 40]. There followed variants in a distal PTF1A enhancer as a cause of isolated pancreatic agenesis (Weedon and colleagues, 2014), GATA4 (Shaw-Smith and colleagues, 2014), CNOT1 with holoprosencephaly (De Franco and colleagues, 2019), YIPF5 (De Franco and colleagues, 2020) and TMEM167A, the third MEDS gene (Virgilio and colleagues, published online on 17 November 2025) [7, 33, 34, 55, 56].

The international Exeter cohort, with 1020 patients referred from 79 countries between 2000 and 2013, showed that early comprehensive genetic testing identified the cause in 82% and came to precede the appearance of extrapancreatic features; the authors concluded that genetic testing defines, rather than confirms, the clinical diagnosis [3]. In July 2026, the same team published the analysis of 125 probands with pancreatic agenesis followed over 21 years, with a genetic diagnosis in 98% [8].

1.3. Nomenclature and current classifications

The most widely used clinical classification is that of the International Society for Paediatric and Adolescent Diabetes (ISPAD), which groups the monogenic subtypes of neonatal and early infancy diabetes into three mechanisms: abnormal pancreatic development (PDX1, PTF1A and its enhancer, HNF1B, RFX6, GATA6, GATA4, GLIS3, NEUROG3, NEUROD1, PAX6, MNX1, NKX2-2, CNOT1, ONECUT1), abnormal β-cell function (KCNJ11, ABCC8, recessive INS, GCK, SLC2A2, SLC19A2, KCNMA1) and β-cell destruction (dominant INS, EIF2AK3, IER3IP1, FOXP3, WFS1, EIF2B1, YIPF5 and the monogenic autoimmunity genes) [2]. The Exeter cohort uses a complementary functional classification, which separates insulin-treated diabetes from disorders of exocrine pancreatic development that also require enzymes (GATA6, PTF1A, PDX1, GATA4) [3].

The nomenclature of the entries has changed over time. OMIM entry 600001, originally "pancreatic agenesis and congenital heart defects", is now titled "heart defects, congenital, and other congenital anomalies" (HDCA), which reflects the fact that heart disease is the most constant manifestation of GATA6 [5, 41]. The DIAR4 entry now includes diabetes and combined pituitary hormone deficiency in its title [27]. Orphanet names Mitchell-Riley syndrome "hypoplastic pancreas-intestinal atresia-hypoplastic gallbladder syndrome" and groups the first two MEDS into a single entity (ORPHA 306558) [23, 31]. Martínez-Frías syndrome retains an OMIM entry with no molecular basis (symbol %), and Cruz and colleagues proposed treating it together with Mitchell-Riley syndrome as a continuum, the "RFX6 malformation complex" [52, 57]. TMEM167A has an OMIM gene entry (620000) but, at the time of consultation, still no phenotype entry [35]. Under the dyadic nomenclature proposed for Mendelian disorders, the CNOT1 syndrome may be called CNOT1-related neurodevelopmental disorder [19].

1.4. Operational definition of the entities

Pancreatic agenesis, in the 2026 Exeter cohort, is defined by diabetes diagnosed before 6 months and at least one of three criteria: biochemically confirmed exocrine pancreatic insufficiency (faecal elastase below 200 µg/g or faecal fat), treatment with pancreatic enzymes, or a reduced or absent pancreas on magnetic resonance imaging (MRI) or ultrasound [8]; in 2013, the definition required insulin-treated diabetes and exocrine insufficiency with enzyme replacement [4]. Mitchell-Riley syndrome is defined by ND associated with hypoplastic pancreas, intestinal atresia and aplasia or hypoplasia of the gallbladder, with biallelic RFX6 variants [21]. GLIS3 syndrome combines non-autoimmune ND with congenital hypothyroidism and, in the complete form, congenital glaucoma, liver disease and renal cysts [24]. Enteric anendocrinosis is defined by malabsorptive diarrhoea with absence or paucity of chromogranin A-positive enteroendocrine cells and normal villous architecture [41, 54]. MEDS are defined by the triad of microcephaly, epilepsy and early-onset diabetes with biallelic variants of IER3IP1, YIPF5 or TMEM167A [34].

2. Molecular Genetics and Pathophysiology

2.1. Genetic architecture of the class

The list of ND genes has exceeded 40; between 2018 and early 2024 six new genes were described (CNOT1, FICD, ONECUT1, PDIA6, YIPF5, ZNF808) and three genes already known in other disorders became ND genes (EIF2B1, NARS2, KCNMA1) [58], joined in 2025–2026 by TMEM167A and the de novo ACTB variant NM_001101.5:c.1043C>T p.(Ser348Leu) [34, 59]. In GeneReviews Table 1, which refers to PNDM in general and not only to the syndromic forms, GATA6, GLIS3 and RFX6 account for about 5% each, PDX1 for about 4%, EIF2AK3 for less than 10% and PTF1A for less than 1%, and an unknown cause persists in less than 20% [1]. The distribution depends above all on consanguinity: in the Exeter cohort, recessive causes accounted for 81% (158/195) of diagnoses in children of consanguineous parents and 13% (81/645) in the others, and among the consanguineous the most frequent cause was WRS (56/230; 24%) [3]. Among the 125 probands with pancreatic agenesis in the same cohort, recessive PTF1A variants accounted for 48% (60/125; 45 in the distal enhancer and 14 in the gene body), GATA6 30% (37), PDX1 6% (7), GATA4 4% (5), RFX6 4% (5), ZNF808 4% (5) and CNOT1 2% (3); there were no cases due to ONECUT1 and 3 probands remained without a cause [8].

Table 2.1 brings together the types of the class, with the features that distinguish them and the test that separates them.

2.2. Transcription factors of pancreatic and endodermal development

GATA6 is a zinc-finger transcriptional activator expressed in tissues derived from endoderm and mesoderm, where it directs the organogenesis of the pancreas, heart and gut [41, 66]. The Gata6−/− mouse dies between embryonic days 6.5 and 7.5 with a defect of endodermal differentiation, and the distribution of the human malformations, all in organs of endodermal origin, is consistent with that role [4]. In the bicuspid aortic valve chapter of the cardiogenetics book, the authors recall that GATA6 is highly expressed in the embryonic pancreas and that de novo variants have been found in patients with pancreatic agenesis, 92% of whom had congenital heart disease [67]. The causal variants are heterozygous and inactivating (haploinsufficiency); in the 2013 series, all missense variants affected residues of the second zinc finger, required for DNA binding and protein interaction, and there were also splicing, frameshift and nonsense variants [1, 4]. Sequencing detects about 90% of variants and targeted deletion/duplication analysis about 10% [1]. Expressivity is highly variable, and a non-coding polymorphism 3′ of GATA6, more frequent in patients with agenesis, reduced gene expression and the efficiency of pancreatic differentiation in edited induced pluripotent stem cells (iPSCs), which makes it a candidate modifier of penetrance (preclinical study) [68]. The 2011 series that established the association assembled the arguments for pathogenicity: in the 12 patients with both parents tested, the variant had arisen de novo in all; of the 14 variants, five were insertions or deletions with a premature stop codon and three altered splicing; the six missense variants hit residues of the DNA-binding surface, conserved in vertebrate orthologues and in all human GATA proteins; four of these mutant proteins no longer bound to GATA6 recognition sites or activated a responsive promoter; and no variant was present in 1094 population controls [40]. Until then, GATA6 variants had only been described in nine patients with congenital heart disease, 1.5% of those tested in those studies, with no mention of pancreatic failure [40]. The heterozygous Gata6+/− mouse has no obvious phenotype, unlike the human, a discordance already known in the haploinsufficient forms due to HNF1A, HNF4A and HNF1B [40].

GATA4 is the master regulator of cardiac development, acts with TBX5 and NKX2-5, and is also required for testicular and pancreatic development; ClinGen classifies the gene as haploinsufficient [9, 41]. The de novo variant p.Asn273Lys, in the second zinc finger, was found in a patient with complete absence of the pancreas confirmed at autopsy and reduced DNA binding and transactivation; 8p23.1 deletions including the gene were the most common alteration in the series that established the association [7]. The mutant protein had a 4.3-fold reduction in transactivation of a GATA4-responsive promoter, and another substitution at the same residue (p.N273S, historical numbering, = NM_001308093.3:c.821A>G p.(Asn274Ser)) had been found in the cardiac tissue of a deceased patient with severe heart disease [7]. The deletions in the series ranged from 1 to 17 Mb, with a common interval including 8 further genes, and the phenotype did not correlate with deletion size [7]. GATA4 haploinsufficiency had been described in more than 140 patients with congenital heart disease and diabetes in only 6, which indicates lower pancreatic than cardiac penetrance; in an earlier Italian family with the p.R319W variant (published designation), the proband had pancreatic agenesis, and the father and sister, both carriers, had heart disease without diabetes [7]. The Gata4+/− mouse is normal, and pancreatic agenesis only appears in the mouse with simultaneous loss of Gata4 and Gata6, whereas in humans haploinsufficiency of either gene suffices [7]. Transient remission of diabetes, rare in pancreatic agenesis, was more frequent with GATA4 (3 of 5 patients), which suggests persistence of functional β cells in some of these patients [8].

PDX1 is the master regulator of pancreatic development and β-cell differentiation and, in the mature cell, regulates insulin, glucokinase and the GLUT2 transporter [1, 69]. Biallelic variants cause pancreatic agenesis or hypoplasia, with more severe insulin deficiency than in the forms due to ABCC8, GCK or KCNJ11, and a milder phenotype with hypomorphic variants [1]; in a series of 19 individuals, 8 (42%) had exocrine insufficiency requiring replacement and 12 (63.2%) extrapancreatic features, 8 of them in the duodenum or hepatobiliary tract, consistent with the abnormal development of the rostral duodenum in the Pdx1-null mouse [11]. Heterozygotes may have MODY4 [1, 10].

PTF1A encodes the p48 subunit, of basic helix-loop-helix type, of the PTF1 complex, which decides whether cells of the pancreatic buds proceed to pancreas or revert to a duodenal fate and which also specifies cerebellar and retinal neurons [41, 70]. Biallelic coding null variants abolish both functions and cause pancreatic and cerebellar agenesis; recessive variants in an enhancer of about 400 base pairs (bp), located about 25 kb downstream of the gene, abolish only the pancreatic function [1, 55]. The enhancer was identified in 2014 by genome sequencing guided by epigenomic annotation of pancreatic progenitors derived from human embryonic stem cells; the variants abolish binding of the FOXA2 and PDX1 factors and enhancer activity [16, 55]. Enhancer variants account for more than 60% of PNDM variants in the gene, according to GeneReviews, and 75% (45/60) of the probands with PTF1A in the Exeter agenesis cohort [1, 8]. A Japanese child compound heterozygous for a single-nucleotide deletion in exon 1 and an enhancer substitution had pancreatic hypoplasia without neurodevelopmental abnormalities, which shows that the combination of a coding allele with an enhancer allele behaves like the isolated pancreatic form [71].

The recurrent de novo CNOT1 variant, NM_016284.5:c.1603C>T p.(Arg535Cys), causes holoprosencephaly with pancreatic agenesis through a defect of pancreatic and neurological development; CNOT1 is a transcriptional repressor associated with maintenance of pluripotency [56, 72]. The 2026 GeneReviews on CNOT1-related Vissers-Bodmer syndrome states that holoprosencephaly and pancreatic agenesis have only been described with this variant, and that other variants in the gene cause a neurodevelopmental disorder without diabetes [19]. ZNF808 is a primate-specific gene: biallelic loss of function causes a spectrum ranging from pancreatic agenesis to adult diabetes without exocrine insufficiency [44]. ONECUT1 (formerly HNF6) causes recessive syndromic diabetes with pancreatic and gallbladder hypoplasia; heterozygous null variants are associated with type 2 diabetes in the UK Biobank (p = 0.006), and a deletion upstream of the gene, in an enhancer specific to pancreatic development, caused the same picture [60–62].

2.3. RFX6 and Mitchell-Riley syndrome

RFX6 is a winged-helix transcription factor of the RFX family, which acts downstream of NEUROG3, heterodimerises with RFX3 at X-box elements and is required for differentiation of four of the five islet cell types and for insulin production; in the mature β cell, it maintains transcription of insulin and of the L-type calcium channels that sustain glucose-stimulated secretion [22, 41, 73]. In the Rfx6-null mouse only pancreatic polypeptide-producing cells form [53]. RFX6 is also required for production and secretion of glucose-dependent insulinotropic polypeptide (GIP) by enteroendocrine cells [22]. In human stem cell cultures and Xenopus embryos, retinoic acid from the lateral mesoderm activates rfx6 in the foregut endoderm, and Rfx6 activates onecut1, pdx1 and Wnt pathway antagonists and represses hindgut and pharyngeal genes, which provides an explanation for the atresias and hepatobiliary malformations of the syndrome (preclinical study published in 2026) [74]. In stem cell-derived islets, biallelic absence of RFX6 prevented the formation of insulin-secreting β cells, with accumulation of precursors and apoptosis, whereas haploinsufficiency (54.2% reduction in protein) reduced insulin secretion by 62% under basal conditions and by 54% with high glucose, without changing β-cell number (preclinical study) [75].

Biallelic truncating variants or missense variants of the DNA-binding domain cause the syndrome; truncating variants at the 3′ end of the gene, with incomplete inactivation, were associated with diabetes only at 3 and 6 years, and the combination of a variant with residual activity in the dimerisation domain with a truncating variant gave milder forms with diabetes at two and a half years [42, 76]. In the 2022 French series, the p.(Arg181Trp) and p.(Val506Gly) variants did not activate the insulin promoter or the L-type calcium channel genes, although the protein localised to the nucleus [22]. GeneReviews does not recognise clinically useful genotype-phenotype correlations for RFX6 [1].

2.4. NEUROG3 and absence of the common endocrine precursor

NEUROG3 is a basic helix-loop-helix transcription factor that acts as a switch for endocrine fate in the gut and pancreas: with NKX2-2 it initiates transcription of NEUROD1 and specifies the common precursor of the four pancreatic endocrine lineages and of enteroendocrine cells, sparing the absorptive epithelium and exocrine pancreas [41, 77]. Patient variants prevent activation of NEUROD1 and binding to the E-box of its promoter [54]. In human organoids, most pathogenic alleles were hypomorphic or null in both tissues, but NEUROG3 is very unstable in the intestinal epithelium, which explains why the gut is more sensitive than the pancreas [78]. Two patients with functionally null biallelic variants only developed diabetes at school age, which indicates that, unlike in the mouse, generation of human pancreatic endocrine cells does not depend entirely on NEUROG3 [43].

2.5. GLIS3

GLIS3 is a nuclear transcription factor of the GLI family with five zinc fingers, able to activate and repress transcription, required for development of the β cell, thyroid, eye, liver and kidney [41, 79]. It is expressed in the pancreas from early stages, more in the β cell than in other pancreatic tissues; in the original description, a frameshift variant caused the complete syndrome and two deletions of the 11 or 12 5′ exons caused incomplete forms, explained by loss of a major transcript in the pancreas and thyroid, or of an eye-specific transcript, with residual expression of other transcripts [25]. About half of pathogenic alleles are exonic deletions, detectable only by copy-number analysis [1]. Monoallelic pathogenic GLIS3 variants increase the risk of type 2 diabetes (odds ratio 3.9; 95% confidence interval 1.4 to 12), and all carriers with type 2 diabetes were sensitive to sulfonylureas [80].

2.6. MEDS and endoplasmic reticulum-to-Golgi trafficking

The three forms of MEDS share a mechanism: delayed export of proteins from the ER to the Golgi, which in the β cell retains proinsulin and sensitises the cell to ER stress-induced apoptosis, and which in the developing brain disturbs secretion of extracellular matrix by neural progenitors [34, 41, 81, 82]. The metabolic diseases handbook includes IER3IP1 among the vesicular transport genes whose variants cause microcephaly, with epilepsy and diabetes [83]. At autopsy of a patient with an IER3IP1 variant there was increased apoptosis in the cerebral cortex and β cells, and patient fibroblasts were more susceptible to cell death under stress [30]; loss of IER3IP1 reduced ER-to-Golgi proinsulin trafficking to one third in stem cell-derived β cells [64]. Loss of function of YIPF5, a component of COPII-coated vesicles, caused proinsulin retention in the ER, marked ER stress and β-cell failure [33]; in 2026, YIPF5 was described as a negative regulator of SURF4 receptor-mediated export, and its reduction in the embryonic mouse cortex caused premature neuronal migration (preclinical study) [84]. TMEM167A is a 72-residue transmembrane protein located in the Golgi, highly expressed in the pancreas and developing brain; in iPSC-derived β cells with the p.(Val59Glu) variant, differentiation into α, β and δ cells was preserved, but proinsulin and insulin content fell to about half, the arrival of proinsulin at the Golgi was delayed (40% colocalisation versus more than 60% at 10 minutes), and grafts in immunodeficient mice produced almost no human C-peptide [34]. IER3IP1 forms a complex with TMEM167A, which strengthens the hypothesis of a common mechanism (preclinical studies) [34].

2.7. Wolcott-Rallison syndrome

EIF2AK3 encodes PERK, an ER kinase that phosphorylates eIF2α when misfolded proteins are present; in its absence, the pancreas develops, but the accumulation of misfolded proteins after birth leads to β-cell apoptosis, and ER secretory dysfunction affects bone, liver and kidney [2, 36, 85]. Sequencing detects about 97% of variants and targeted copy-number analysis about 3% [1].

2.8. Allelic disorders, heterozygotes and two inheritance patterns in the same gene

Several genes of the class have different phenotypes depending on allelic dosage, which is relevant to reporting and counselling. In RFX6, biallelic variants cause Mitchell-Riley syndrome, and heterozygous truncating variants cause MODY-type diabetes with reduced penetrance: 27% at 25 years, versus 70% in HNF1A and 55% in HNF4A, with lower fasting and stimulated GIP [63]. A heterozygous truncating RFX6 finding should therefore not be reported merely as carrier status [41]. In Finland, the RFX6 variant p.His293LeufsTer7 (published designation) has about 1 carrier in 250 people and is associated with type 2 and gestational diabetes [75]. In PDX1, the biallelic form causes pancreatic agenesis and the heterozygous form MODY4 [10]. In GATA6 and GATA4, the same haploinsufficiency causes isolated heart disease, pancreatic agenesis with heart disease or adult diabetes, and the variant may be inherited from a mildly affected parent [4, 41]. In GLIS3 and ONECUT1, monoallelic alleles are associated with type 2 diabetes [61, 80]. Heterozygotes for GCK, INS or PDX1 may have milder diabetes, whereas GeneReviews does not describe allelic phenotypes for EIF2AK3, GLIS3, PTF1A and SLC19A2 [1].

GeneReviews Table 4 classifies the STAT3 form as autosomal recessive; ISPAD describes it as spontaneous, and it is due to heterozygous activating variants, so the recessive classification in the table should not be followed [1, 2].

2.9. Variant nomenclature and recurrent variants

The variants cited were normalised in VariantValidator to the MANE Select transcript (Table 2.2). For variants that the sources described only at the protein level, the c. description is the one VariantValidator confirms as corresponding to the published protein consequence [72].

The enhancer variant chr10:g.23508437A>G (GRCh37) was the most frequent in the series of 30 patients and had already been described in Turkey, which suggests a founder effect; the same variant was found in the homozygous state in 9 patients from 6 Syrian, Lebanese, Kurdish and Turkish families in the original description [16, 17]. In three unrelated Bedouin families, four infants with enteric anendocrinosis were homozygous for the novel NEUROG3 variant p.(Gln137Arg), a possible founder variant [93]. The GATA4 variant p.Asn273Lys was described on transcript NM_002052.3 (c.819C>A); on the MANE transcript, with numbering shifted by one codon, it corresponds to NM_001308093.3:c.822C>A, p.(Asn274Lys), and the designation from the original article is kept in the text to facilitate searching [7, 72].

3. Clinical Spectrum and Manifestations by System

3.1. Features common to the class

Insulin deficiency begins in utero and is reflected in birth weight, which is markedly reduced in ND in general (1.8 to 2.5 kg) [51]. In the international Exeter cohort, median birth weight was 2460 g (standard deviation, SD, −1.7) and diabetes was diagnosed at a median of 6 weeks [3]; in pancreatic agenesis, the weight is even lower (median z-score −3.16) and diabetes appears at a median of 2 weeks, with no differences between genetic subtypes, which suggests a shared mechanism of absence or paucity of β cells [8]. Presentation includes hyperglycaemia, glycosuria, osmotic polyuria, severe dehydration and poor weight gain, with low or undetectable insulin and C-peptide [1]. Extrapancreatic features were present in 81% (105/129) of patients with pancreatic agenesis, and neurological manifestations were, across the whole ND cohort, the most common extrapancreatic features (n = 184), with nine genes involved, including IER3IP1, PTF1A and EIF2AK3 [3, 8].

Diabetes in these forms has two particular features that distinguish it from T1D. The first is lability, with frequent fasting and postprandial hypoglycaemia; in a North American registry of 11 children with GATA6, GATA4 or PDX1, mean glycated haemoglobin (HbA1c) was 8.7%, but there was only 1 episode of diabetic ketoacidosis (DKA), which the authors partly attribute to lack of glucagon, the driver of ketosis [89]. In that registry, families described blood glucose so unstable that it required 10 or more measurements a day, rapid-onset hypoglycaemia, especially at night, and a tendency to keep glucose above target to avoid it; insulin sensitivity was high in the first years and the daily dose per kilogram varied widely, with no trend with age [89]. Malabsorption worsens the lability, because the subcutaneous insulin bolus is absorbed normally while food is absorbed slowly or incompletely, which caused hypoglycaemia after conventional boluses [89]. The only DKA occurred at diagnosis, at 4 months, in a child compound heterozygous for PDX1 without exocrine insufficiency, and no child had DKA after diagnosis [89]. The second is growth failure, which persists despite replacement doses of insulin and enzymes: in the cohort of 30 patients with PTF1A enhancer variants, mean height at last follow-up was −2.4 SD, with short stature in 69% and no improvement with age [16].

3.2. Pancreatic agenesis and hypoplasia

Pancreatic agenesis manifests as insulin-dependent ND and exocrine pancreatic insufficiency, whose symptoms (steatorrhoea, malabsorption, poor weight gain, loose, foul-smelling stools) are non-specific and often go unnoticed in the newborn [1, 8]. In the 2026 Exeter cohort, 32% (40/125) of probands did not yet have a diagnosis of agenesis when referred for genetic testing, because they had had neither biochemical assessment nor imaging; they were referred earlier (median 3.5 versus 7 months) and the genetic diagnosis triggered the exocrine assessment [8]. Remission of diabetes (at least two weeks without insulin) occurred in 6 of 129 patients (3 with GATA4 deletions, 2 with GATA6 and 1 with the PTF1A enhancer), at a median of 14.5 weeks, and there was relapse in 5 of the 6 [8].

GATA6. In the 2013 Exeter series, with 24 probands and 5 parents, diabetes was diagnosed at a median of 2 days (interquartile range, IQR, 1 to 7) and median birth weight was 1588 g (SD −3.22); imaging showed complete absence of the pancreas in 8 and marked hypoplasia in 10 of 18 patients studied, and faecal elastase or fat confirmed severe insufficiency in 15 [4]. Two probands had PNDM without ever having received enzymes, one with elastase below 15 µg/g; another, with the de novo variant NM_005257.6:c.1435A>G p.(Arg479Gly), had diabetes at 8 weeks treated with insulin for one week, and thereafter only hyperglycaemia during a rotavirus infection and after cardiac surgery, being at 5 years without insulin and without enzymes, with elastase of 61 µg/g [4]. Four heterozygous parents had diabetes diagnosed between 12 and 46 years, treated with insulin (3) or metformin (1), and one mosaic parent had no diabetes at 43 years but had undergone surgery for a patent ductus arteriosus in childhood [4]. Congenital heart disease was present in 83% (21 of 24 probands and 3 parents), with surgical correction in 19; this was followed by congenital hypothyroidism (6), hepatobiliary malformations, mainly gallbladder agenesis and biliary atresia (5), intestinal anomalies such as malrotation and hernias (6), and significant neurocognitive deficit (11) [4]. The initial series of Allen and colleagues (2011) included 27 patients with pancreatic agenesis, defined by insulin-treated ND and exocrine insufficiency requiring enzymes, born to parents without diabetes; in the 21 with imaging, the pancreas was absent in 16 and markedly hypoplastic in the rest [40]. The 15 with GATA6 variants (56%) did not have consanguineous parents (0/15), unlike 5 of the 11 without a variant; of these 11, one had a homozygous PTF1A splicing variant, none had PDX1 variants, and only 2 had extrapancreatic features, which led the authors to propose at least one as yet unidentified recessive subtype of isolated pancreatic agenesis [40]. In that series, heart disease was present in 14 of the 15 patients with GATA6, mainly outflow tract malformations, atrial and ventricular septal defects and tetralogy of Fallot [40], and, according to OMIM, neurocognitive deficit in 47%, gallbladder and biliary tract pathology in 33%, intestinal anomalies, including diaphragmatic hernia, in 33% and other endocrine anomalies in 20% [5]. The cardiac defects described are septal and outflow tract defects (truncus arteriosus, tetralogy of Fallot, atrial septal defect, atrioventricular septal defect), and the pancreas ranges from complete agenesis to a normal appearance [41]. Two patients with IUGR, minor heart disease and ND, with pancreatic agenesis confirmed by imaging and de novo variants identified by exome sequencing, were stable on insulin and enzymes [94]. A Japanese family described in 1994 illustrates the expressivity: the mother, operated on for a patent ductus arteriosus and an atrial septal defect at 19 years, developed diabetes at 28 years and had only the head and uncinate process of the pancreas, and one of her children had diabetes and cyanotic heart disease with severe pancreatic hypoplasia [5].

North American registry (GATA6, GATA4 and PDX1). The US Monogenic Diabetes Registry (University of Chicago) included 11 children with diabetes before 1 year and at least 2 years of age (8 with GATA6, 1 with GATA4 and 2 with PDX1), with data from questionnaires, interviews and medical records; of the 14 eligible, one under 2 years and two with variants of uncertain significance were excluded [89]. All GATA6 and GATA4 variants were heterozygous and de novo [89]. Diabetes began in the first week in 8 of the 11, and three children (two with GATA6 and the child with GATA4) were temporarily off insulin with normal glucose, for up to about 1 year in the longest case, and resumed it when glucose of 200 to 300 mg/dL appeared or after cardiac surgery; one of the children with agenesis due to GATA6 only had diabetes at 9 months [89]. Seven had complete agenesis (six confirmed by imaging or surgery) and two hypoplasia; one child with GATA6 reported a normal-appearing pancreas and the child with GATA4 never had imaging [89]. Frequencies were: congenital heart disease in 10/11, initial poor weight gain in 10/11 (persisting into childhood in 6/10), exocrine insufficiency in 8/11, gallbladder agenesis in 7/11, intestinal malrotation in 4/11, recurrent infections (otitis with repeated tympanostomy tubes, intestinal infections) in 3/10, short stature in 3/10, renal anomalies in 3/10 and epilepsy in 2/10 [89]. All children were small for gestational age, with a mean birth weight of 1532 g, and 9 had diagnosed IUGR; of the three with height below −2 SD for at least 3 years, one had growth hormone deficiency treated from 4 years, another severe infantile scoliosis operated on at 9 years and the third low insulin-like growth factor 1 (IGF-1), untreated [89]. The six children with exocrine insufficiency and available records had difficulty gaining weight, four with repeated weights below the 5th centile, whereas in the three without exocrine insufficiency the poor weight gain resolved between 1 and 2 years, in two after correction of the heart disease and resolution of feeding difficulties, or weight catch-up occurred at 8 months [89]. The authors note that recurrent infections, epilepsy, scoliosis and short stature had not been frequently described and that their relationship with the variants remains to be demonstrated, and that the small sample, with three genes, limits generalisation [89].

GATA4. In five patients with deletions (4) or a variant (1) of GATA4, diabetes began on day 1 to 7 of life in four, with a variable exocrine phenotype; the p.Asn273Lys variant was associated with complete absence of the pancreas [7]. Three patients came from a cohort of 13 with diabetes and chromosomal anomalies (terminal 8p deletions of 12.5, 15 and 17 Mb), and two from 186 probands with ND before 6 months without a diagnosis, selected from 867 patients from 73 countries, a frequency the authors estimate at 0.5% (4/867) of their international series [7]. Two neonatal diabetes cases remitted and relapsed after 6 months and 7 years; one was permanent from the first week; one girl had diabetes at 13 years, after 6 months of polyuria and polydipsia, with HbA1c of 12.6% (114.2 mmol/mol), without ketosis or autoantibodies and with stimulated C-peptide of 3.8 ng/mL three years later; and a preterm infant born at 34 weeks with diabetes from day 1 died of multiorgan failure on day 4 [7]. The four patients with ND had birth weight below the 3rd centile, and all were treated with insulin [7]. The exocrine phenotype ranged from complete agenesis at autopsy to insufficiency treated with enzymes (faecal elastase 17 µg/g), to a reduced pancreas on imaging in two asymptomatic patients and a low elastase (132 µg/g; normal above 200) in one of them [7]. Heart disease (from septal defects with pulmonary stenosis to atrioventricular canal defect) was present in the four patients with a deletion and not in the patient with the missense variant, and there was developmental delay or neurocognitive deficit in all five, with epilepsy in one [7]. In the North American registry, a girl with the de novo variant p.(Cys292Trp) (numbering of NM_002052.5:c.876C>G; = NM_001308093.3:c.879C>G p.(Cys293Trp)) had diabetes at 7 months, without exocrine insufficiency, with minimal insulin requirements in the first 3 years, complete discontinuation at 3 years and return of hyperglycaemia at 4 years, shortly after surgical correction of an atrioventricular septal defect [89]. GeneGuidance lists septal heart disease as the main phenotype, the contiguous 8p23.1 deletions with diaphragmatic hernia and developmental delay, and 46,XY difference of sex development [41].

PTF1A (distal enhancer). The largest series brings together 30 patients (19 boys) from 25 families, with consanguinity in 92%, followed for a median of 4.08 years: birth weight had a median SD of −3.42, with IUGR in 86.2%; diabetes was diagnosed at a median of 5 days, but in 2 patients only at 19 months and 11 years; all had exocrine insufficiency, biochemically confirmed in 19 (low elastase, lipase and pancreatic amylase, hypoalbuminaemia, low vitamins A, D and E), and all had agenesis or hypoplasia on imaging [16]. Frequent extrapancreatic features were transiently raised ferritin (12/12 measured in the neonatal period), normocytic anaemia (19/25; 12 with multiple transfusions), transient elevation of transaminases (18/25) and cholestasis (14/24); cholestasis resolved with enzymes in 3 and with ursodeoxycholic acid in 11 [16]. There were neurological signs in 7: mild delay in 4 (2 of them with epilepsy and 1 with congenital cataract), isolated epilepsy in 2 and microcephaly with global delay and epilepsy in 1, without cerebellar aplasia on the 16 available MRIs [16]. In 6 families, siblings with ND or raised perinatal glucose had died without a diagnosis [16]. In the original description, 3 of 14 patients only had diabetes at 8, 10 and 22 years and one died of cholestatic liver failure [17].

PTF1A (coding region). Coding null variants cause severe IUGR with very little subcutaneous tissue, a triangular "aged" face, beaked nose, low-set dysplastic ears, contractures, club foot, hypoplasia or atrophy of the optic nerve and agenesis of the cerebellum and vermis; 4 cases with central hypoventilation and survival up to 4 months have been described [13, 16]. Prenatal diagnosis was possible in one family through absence of the cerebellum and severe IUGR [13].

PDX1. PAGEN1 manifests as IUGR and endocrine and exocrine insufficiency in the first weeks or months, with exocrine deficiency sometimes subclinical [10]. In the series of 19 patients, 42% had treated exocrine insufficiency and 42% anomalies of the duodenum or hepatobiliary tract [11]. The two patients in the North American registry illustrate the extremes: a boy compound heterozygous for NM_000209.4:c.533A>G p.(Glu178Gly) and c.502A>C p.(Asn168His), the latter a VUS, had hypoplasia with a normal-appearing head and body of the pancreas, diabetes with DKA at 4 months and faecal elastase below 50 µg/g at 13 months without steatorrhoea, which was not treated; at 3 years, faecal fat (0.7 g/day) and elastase (above 500 µg/g) were normal [89]. The other, homozygous for c.188del, born to parents with MODY-type diabetes and remote kinship, had severe hyperglycaemia at birth, pancreatic agenesis, exocrine insufficiency at 23 days, gallbladder agenesis, epilepsy, repeated intestinal infections and eczema that prevented the use of sensors and pumps [89].

CNOT1. Of the seven known carriers of the p.(Arg535Cys) variant, four had pancreatic agenesis with ND and six holoprosencephaly (five semilobar, one lobar), with gallbladder agenesis, deafness and seizures in some of them; one patient also had diabetes insipidus [18, 19]. Vissers-Bodmer syndrome, with 45 individuals described, includes developmental delay or intellectual disability, hypotonia, feeding difficulties and epilepsy [19].

ZNF808 and ONECUT1. In the Exeter cohort, 31 individuals had biallelic loss of function of ZNF808, 30 born to related parents: 19 with PNDM, 5 with early infancy diabetes, 4 with transient diabetes and 3 with diabetes at 10, 14 and 23 years; 3 were treated with a sulfonylurea and 17 had no signs of exocrine insufficiency [44]. A Saudi infant with a homozygous variant also had skeletal anomalies (fixed flexion of the knee and oligodactyly), an association not yet confirmed in other patients [95]. Biallelic ONECUT1 variants cause IUGR, pancreatic hypoplasia and agenesis or hypoplasia of the gallbladder [60].

3.3. Mitchell-Riley syndrome

The syndrome combines ND with malformations of several segments of the digestive tract: small bowel (malrotation, duodenal atresia, jejunal atresia), pancreas (annular or hypoplastic) and hepatobiliary system (gallbladder agenesis, extra- and intrahepatic biliary atresia) [96]. Cholestasis, liver failure, heterotopic pancreas, anaemia, thrombocytopenia, haemochromatosis, hypospadias, coagulopathy and acidosis have also been described [22]. By 2022, 16 cases had been published (12 homozygous, 4 compound heterozygous); 7 of the 12 homozygous patients died of sepsis or liver failure before 6 months [22]. In the description by Mitchell and colleagues, the two sibling pairs died in the first year, whereas a girl with a milder form was off insulin at 1 year [21]. A preterm infant born at 31 weeks weighing 880 g, homozygous for p.(Arg525Ter), had duodenal atresia, hyperglycaemia from day 4, chronic diarrhoea, exocrine insufficiency and cholestasis, and died on day 195 from metabolic instability and progressive liver dysfunction [88].

The French series of four patients who survived more than two years describes the treated disease [22]. All were small for gestational age, with duodenal atresia diagnosed before birth and operated on in the first week; hyperglycaemia (300 to 500 mg/dL) appeared on day 1 or 2, with low C-peptide. Chronic diarrhoea was severe in all, and dependence on parenteral nutrition (PN) at the last assessment ranged from none (stopped at one year) to 119% of non-protein energy requirements. Three had intrahepatic cholestasis, self-limiting in two and progressing to liver failure and transplantation at 7 months in the third. All were born with normocytic anaemia and needed transfusions and intravenous iron in the first year; all had vitamin K deficiency and, in two, factor VII remained between 15% and 30% despite supplementation; three had persistent acidosis requiring bicarbonate; all had infections with extended-spectrum beta-lactamase-producing Enterobacterales and catheter-associated sepsis. Motor development was appropriate or slightly delayed, and one patient was 13 and a half years old at the last assessment [22].

Two observations broaden the spectrum. In two cousins born to double first-cousin parents, with intestinal atresias at birth, diabetes only appeared at 3 and 6 years [42]; two compound heterozygous sisters had diabetes at 2 years and 10 months and at 2 years and 7 months, and surgical removal of the small bowel segments with heterotopic gastric mucosa improved the malabsorption, poor weight gain and anaemia [76]. Ectopic gastric mucosa was also found in the duodenum of a French patient and in extensive areas of the explanted small bowel of a Romani child who underwent multivisceral transplantation [22, 97]. A newborn with a dual diagnosis of Mitchell-Riley syndrome and Temple syndrome has also been described [98]. At Hospital Dona Estefânia (Lisbon), two girls born to consanguineous parents, both homozygous for p.(Arg181Trp) and with no known kinship despite geographical proximity, had protracted diarrhoea with plasma glucagon-like peptide 1 (GLP-1) below the limit of quantification (below 1 pmol/L, fasting and throughout a mixed-meal test in the first, in a single sample in the second), undetectable glucagon (below 1.7 pmol/L) and, in duodenal, sigmoid and rectal biopsies, depletion of chromogranin A-positive cells and almost complete absence of GLP-1- and gastric inhibitory polypeptide (GIP)-positive enteroendocrine cells [87]. The first, already described up to 7 months [86], had at 6 years difficult-to-control diabetes on continuous subcutaneous insulin infusion (CSII) and intractable diarrhoea, with 5 to 8 watery stools a day, perineal dermatitis, poor weight gain and dependence on overnight PN; at 12 months, low plasma citrulline and the D-xylose test (6 mg/dL at the first hour, with a reference above 20 mg/dL) showed malabsorption, and severe exocrine insufficiency (faecal elastase below 30 µg/g) only became evident from 30 months; semi-elemental and elemental formulas, pancreatic enzymes, pectin, racecadotril and loperamide did not control the diarrhoea [87]. The second was born at 32 weeks with IUGR and 1185 g, had diabetes with very low C-peptide from day 1, correction on day 2 of type III duodenal atresia, malrotation and gallbladder agenesis, with a retroperitoneal biliopancreatic cyst that required cystoduodenal anastomosis; cholestasis appeared on day 8 and worsened, watery diarrhoea began on day 37, faecal elastase was normal at 5 months and below 30 µg/g from 7 months, and the patient died at 8 months from hepatic encephalopathy [87].

3.4. Neonatal diabetes with congenital hypothyroidism (GLIS3)

Newborns are small for gestational age and have non-autoimmune insulin-dependent diabetes in the first days or weeks, with congenital hypothyroidism frequently detected by newborn screening [24, 41]. The complete form includes congenital glaucoma (with high hypermetropia and short axial length), cholestatic liver disease with fibrosis and bile duct paucity that may progress to biliary cirrhosis, and renal cysts or cystic dysplasia; exocrine insufficiency, developmental delay, dysmorphism (low-set ears, epicanthus, flat nasal bridge, long philtrum, thin upper lip), sensorineural deafness and osteopenia have been described [24, 41]. By 2022, 22 patients from 16 families in 11 countries had been published [99]. Hypothyroidism presents with an agenetic, hypoplastic or normal-sized thyroid, and resistance to thyroid-stimulating hormone (TSH), with raised thyroglobulin, is frequent [39, 100]. The course depends on parenchymal involvement: the Saudi siblings of the original description died at 16 months, 6 months and 10 days, whereas two French Roma siblings without liver, renal or ocular disease were alive at 22 and 14 years [24]. A Turkish patient followed up to 17 years, without liver, renal or exocrine disease, had glaucoma, facial dysmorphism, multiple naevi and an intelligence quotient of 69, and the authors note that the three oldest known patients had no parenchymal disease [100]. Two children with chronic liver disease required liver transplantation for biliary cirrhosis with ductal paucity [101]. A compound heterozygous child (published as c.1330delC, p.Pro444fsdelG, = NM_001042413.2:c.1335del p.(Ala446ArgfsTer110), and NM_001042413.2:c.1940A>G p.(His647Arg)) also had cardiac anomalies and polycystic kidneys [99].

3.5. Enteric anendocrinosis (NEUROG3)

The disease begins in the first weeks with severe watery diarrhoea, vomiting, dehydration and hyperchloraemic metabolic acidosis after cow's milk formula; the diarrhoea stops with fasting and returns with glucose or amino acid solutions [27]. Small bowel biopsy shows normal villi and Paneth, goblet and absorptive cells present, but paucity or absence of chromogranin A-positive enteroendocrine cells; exocrine function is normal [27, 54]. Most develop insulin-dependent diabetes in the neonatal period or in childhood; in the three boys of the original description, the two survivors had diabetes at 8 years, and the third died of sepsis at 3 years [27, 41]. The endocrine features appear after the diarrhoea: short stature, delayed puberty, hypogonadotropic hypogonadism, subclinical hypothyroidism with a hypoplastic thyroid and, in three Thai patients, pituitary hypoplasia with multiple deficiencies and proximal tubulopathy [27, 102]. By 2022, 17 cases had been described [102]. In a girl with the p.(Glu123Ter) variant, the pancreas was present, but serum amylase and faecal elastase were low, and persistent hyperglycaemia appeared at 5 months [92]. The earliest known diagnosis was made at 4 weeks by an ND panel, in an infant with severe diarrhoea, malnutrition and ND, with a reduction of GLP-1-positive cells in the gut [103].

3.6. Microcephaly, epilepsy and diabetes syndromes

MEDS1 (IER3IP1). The triad is primary microcephaly with a simplified gyral pattern, severe infantile-onset epileptic encephalopathy and permanent diabetes in the first months [30]. In 4 Egyptian patients from 2 consanguineous families, diabetes appeared soon after birth, seizures in the first months with a burst-suppression pattern, microcephaly ranged from −3.6 to −9 SD, and there was dysmorphism, cortical atrophy, hypoplasia or agenesis of the corpus callosum and vermis hypoplasia; in the first patient described, head circumference was −2.5 SD at birth and the electroencephalogram showed hypsarrhythmia, with death at 18 months [29]. A compound heterozygous boy, born to non-consanguineous parents, also had cortical blindness, severe delay and recurrent respiratory infections, and died at 8 years [90]. In a 2017 review, 8 cases of MEDS1 had been described [104]. The 12th case worldwide, in Morocco, had simplified gyration with partial agenesis of the corpus callosum [91]; a Turkish case, homozygous for NM_016097.5:c.53C>T p.(Ala18Val), did not have simplified gyration on early MRI and had later-onset diabetes [105]. In an infant girl with infantile spasms, diabetes became clinically evident after corticosteroid therapy [106].

MEDS2 (YIPF5). In the 6 patients from 5 consanguineous families, microcephaly was severe (median −6.2 SD), epilepsy began between the 1st and 7th month and diabetes, treated with full replacement doses of insulin, between 4 weeks and 20 months; birth weight was low in all [32, 33]. The proband of an Indian family died at 1.3 years and had four affected siblings who died in infancy; a 5-year-old Turkish boy, homozygous for NM_030799.9:c.542C>T p.(Ala181Val), had normal neuromotor development and resolution of epilepsy at 2 years [32]. A novel variant caused only short episodes of severe hyperglycaemia during common illnesses, fully reversible [65].

TMEM167A MEDS. The six patients described had ND diagnosed at a median of 34 days (1 day to 22 weeks), treated with insulin from diagnosis; microcephaly was very severe (median −8.9 SD; −4.17 to −10.3), five had neonatal epilepsy, three had IUGR, and the two patients with MRI had lissencephaly [34]. Five died between 2 months and 14 years; one patient died at 59 days without confirmed epilepsy, and a sister of another patient, with a similar picture, died in the neonatal period [34].

3.7. Other syndromic forms

The de novo ACTB variant p.(Ser348Leu) was found in 2 of 326 patients with undiagnosed ND; both had diabetes on day 1 and day 8 and birth weight of −3.22 and −3.98 SD, one with deafness and developmental delay and the other with intestinal atresia; among the 9 known carriers, 5 had ND or hyperglycaemia and 6 of 7 from the literature had gastrointestinal atresia [59]. The remaining forms in GeneReviews Table 4 (NEUROD1, NKX2-2, MNX1, PAX6, KCNMA1, EIF2B1, WFS1 and the autoimmunity genes) belong to non-syndromic monogenic neonatal diabetes and the autoimmune forms [1].

3.8. Natural history and treated disease

The untreated natural history of these forms is not described in modern series: older reports and the siblings who died without a diagnosis in families with PTF1A enhancer variants indicate early mortality without insulin and without a diagnosis [16, 21]. With treatment, mortality continues to depend on the associated malformations: the forms with cerebellar agenesis, holoprosencephaly or MEDS have short survival in most described cases, Mitchell-Riley syndrome has high mortality in the first months from sepsis and liver failure, and GLIS3 syndrome depends on liver and renal disease [13, 22, 34, 100]. In treated isolated agenesis, survival is the rule, but glycaemic control and growth fall short of what is desirable: only 2 of 28 patients with PTF1A enhancer variants had HbA1c below 7.0% [16].

4. Clinical Recognition and Warning Signs

4.1. When to suspect

Any diabetes diagnosed before 6 months is, until proven otherwise, monogenic, and should be studied genetically at once; between 6 and 12 months, testing should be considered especially when islet autoantibodies are negative or there are suggestive features [2]. The Oxford Desk Reference advises referral of all cases diagnosed before 9 months [51]. Suspicion of a syndromic form arises from the combination of diabetes with one of the following contexts: marked IUGR, parental consanguinity or siblings who died in the neonatal period; digestive malformations operated on in the neonatal period (duodenal or jejunal atresia, malrotation, annular pancreas, gallbladder agenesis); congenital heart disease; steatorrhoea, foul-smelling stools, poor weight gain or low faecal elastase; hypothyroidism detected by newborn screening, corneal clouding or glaucoma; severe watery diarrhoea that stops with fasting; microcephaly, seizures or infantile spasms; holoprosencephaly or absent cerebellum on prenatal ultrasound or MRI [1, 4, 13, 19, 21, 24, 27, 34].

The same features, in reverse order, should prompt a search for diabetes. A child with the intestinal atresias typical of Mitchell-Riley syndrome should have glucose monitored, because diabetes may appear only at 2 to 6 years [42, 76]. In duodenal atresia, trisomy 21 is the first condition to exclude (30% to 50% of duodenal atresias) [96]. An infant with congenital hypothyroidism should be examined for dysmorphism and malformations, especially cardiac, and concomitant diabetes suggests GLIS3 [24, 45]. A child with a genetic diagnosis of GATA4 or GATA6 made because of heart disease needs surveillance of glucose and exocrine function [41].

4.2. Warning and emergency signs

The emergency signs are severe dehydration and DKA at presentation, severe hypoglycaemia that does not respond to oral glucose or glucagon, catheter-associated sepsis in patients on PN, progressive cholestasis or liver failure (Mitchell-Riley syndrome, GLIS3, PTF1A enhancer, WRS), refractory seizures in MEDS and central hypoventilation in the coding PTF1A forms [1, 13, 16, 22, 29, 39, 101]. In Mitchell-Riley syndrome, oral glucose was of little effect (20 to 30 g were needed, twice the usual dose) and glucagon was relatively ineffective; only intravenous glucose (10%, 2 mL/kg) corrected the hypoglycaemia [22]. In a child with GLIS3, severe and prolonged hypoglycaemia during gastroenteritis did not respond to three intramuscular injections of 1 mg glucagon and was only corrected with intravenous glucose in the intensive care unit [39]. Insulin requirements may triple or quadruple during infections [39].

4.3. Newborn screening

There is no newborn screening for ND. Screening for congenital hypothyroidism detects the thyroid component of GLIS3 syndrome, and coexisting hyperglycaemia in a newborn with raised TSH should prompt study of GLIS3 [39, 41]. Low birth weight or preterm newborns under routine glucose monitoring are often detected by hyperglycaemia in the first days, as occurred in the four French patients with Mitchell-Riley syndrome [22].

5. Diagnosis: Criteria, Tests and Algorithms

5.1. Criteria and recommendations

The findings suggestive of PNDM are persistent hyperglycaemia (plasma glucose above 250 mg/dL, or 13.9 mmol/L) for more than 7 to 10 days in an infant under 6 months, glycosuria, ketonuria and hyperketonaemia, and insulin and C-peptide low or undetectable relative to glucose; HbA1c cannot be used to diagnose diabetes before 6 months, because of the high proportion of fetal haemoglobin [1]. The diagnosis of PNDM is established by diabetes diagnosed before 6 months that does not resolve, and molecular testing is recommended because it guides treatment [1]. ISPAD recommends immediate molecular genetic testing in all infants diagnosed before 6 months (grade B) and comprehensive next-generation sequencing (NGS), rather than phenotype-directed testing (grade B), and states that there is no need to wait for remission or for other features to appear, because laboratories offer comprehensive testing of all subtypes and rapid testing of those that change treatment [2]. The result should be reported clearly (grade E), and referral to a monogenic diabetes specialist or a genetics unit is suggested (grade E) [2].

5.2. Biochemical, functional and imaging assessment

The initial assessment includes glucose, C-peptide and insulin, islet autoantibodies (to exclude autoimmunity), thyroid function, echocardiogram and abdominal ultrasound [1, 41]. Exocrine assessment is mandatory in agenesis and hypoplasia: GeneReviews indicates pancreatic imaging and faecal elastase and fat-soluble vitamins in the forms due to HNF1B, PDX1, PTF1A and RFX6 [1]. ISPAD notes that pancreatic imaging is unreliable in the newborn and that functional tests (faecal elastase and faecal fat) are preferable for assessing pancreatic aplasia [2]. In the North American GATA6, GATA4 and PDX1 registry, exocrine insufficiency was diagnosed, in most, by steatorrhoea followed by faecal elastase and, in children with absence of the pancreas seen at surgery, confirmed by faecal fat and elastase and by serum amylase and lipase; the most common presenting symptoms were abdominal distension, feeding intolerance and loose, foul-smelling stools [89]. Elastase alone can mislead in both directions: in a child with PDX1, a value below 50 µg/g at 13 months, without steatorrhoea, normalised at 3 years, and in a child with GATA6 and hypoplasia the diagnosis remained uncertain throughout childhood, but each attempt to stop the enzymes brought diarrhoea, vomiting and weight loss [89]. In enteric anendocrinosis, the decisive non-molecular test is small bowel biopsy with chromogranin A immunohistochemistry, which shows absence or paucity of enteroendocrine cells with normal villi, unlike coeliac disease, microvillus inclusion disease and tufting enteropathy; stool osmolality and volume during a feeding pause and pituitary assessment complete the work-up [41, 54]. Brain MRI distinguishes the coding PTF1A form (cerebellar agenesis) from the enhancer form (normal cerebellum) and shows the simplified gyration or lissencephaly of MEDS and the holoprosencephaly of CNOT1 [19, 34, 41]. In GLIS3 syndrome, assessment includes intraocular pressure and axial length, liver tests, renal ultrasound and thyroglobulin [41, 100]. In Mitchell-Riley syndrome, ferritin, full blood count, coagulation (vitamin K and factor VII) and acid-base balance are part of the initial assessment [22]; with PTF1A enhancer variants, ferritin was raised in all patients measured in the neonatal period, and the authors propose measuring it as a marker of cholestasis or hepatocellular dysfunction [16].

5.3. Genetic testing: order and methods

GeneReviews describes two options: a multigene ND panel including the genes of the class and sequencing, deletion/duplication and methylation methods, or exome or genome sequencing [1]. Three technical aspects are specific to this class. First, the distal PTF1A enhancer is non-coding and is not captured by the exome or by panels restricted to exons; a laboratory that analyses only the coding region gives a negative result in most true cases [1, 8, 41]. Second, half of pathogenic GLIS3 alleles and about 10% of GATA6 alleles are exonic deletions, and GATA4 deletions are often part of 8p23.1 deletions, so copy-number analysis, or a chromosomal microarray, is essential [1, 7, 41]. The authors of the GATA4 series recommend looking for intragenic variants and deletions of the gene in patients with neonatal or childhood diabetes and congenital heart disease without a GATA6 variant [7]. Third, recently described genes (TMEM167A, ZNF808, ACTB) may not be included in older panels; the authors of the TMEM167A study recommend including it in ND and congenital microcephaly panels [34]. In the Exeter pancreatic agenesis cohort, the diagnosis was obtained in 98% of probands with a combination of Sanger sequencing, targeted NGS, exome and genome, and the three probands without a cause had trio genome sequencing without result [8]. The authors recommend including all agenesis genes in ND panels, regardless of apparent exocrine function, because 77 patients with variants in these genes did not yet have confirmed agenesis [8].

5.4. Limits of testing

A negative panel result does not exclude the class. The causes of false negatives are absence of the PTF1A enhancer or of ONECUT1 regulatory regions from the panel, lack of copy-number analysis, recently described genes, parental mosaicism (described in GATA6, with about 20% of the variant in the leucocytes of a parent) and variants of uncertain significance (VUS), which neither establish nor exclude the diagnosis [1, 4, 62]. In the North American registry, a child with diabetes at 7.5 months and the GATA4 VUS NM_001308093.3:c.436C>G p.(Arg146Gly) (same numbering in NM_002052.5), without pancreatic agenesis or heart disease, had a T1D genetic risk score between the 50th and 75th centiles, which made autoimmune T1D more likely than a monogenic cause [89]. The proportion of variants detectable by each method, given in GeneReviews, is not the clinical sensitivity of the test [1]. The genetic cause remained unidentified in 18% of non-consanguineous and 15% of consanguineous patients in the cohort of 1020 patients, and there are patients with the full Mitchell-Riley/Martínez-Frías picture without an RFX6 variant, which suggests genetic heterogeneity [3, 57]. GeneReviews advises storing DNA from probands without a molecular diagnosis [1].

5.5. Interpretation and report content

The report should state the gene, the transcript with version, the variant in HGVS nomenclature, zygosity, the phase of variants in compound heterozygosity and the class according to the American College of Medical Genetics and Genomics and Association for Molecular Pathology (ACMG/AMP) criteria, for which there are ClinGen expert panels in monogenic diabetes [1, 2]. In genes with two phenotypes by allelic dosage, the report should state that a heterozygous truncating RFX6 variant has significance of its own (reduced-penetrance MODY) and not merely carrier significance, and that a GATA6 or GATA4 variant implies cardiac and exocrine assessment [41, 63]. PTF1A enhancer variants are described in genomic coordinates, with the reference genome version [16, 72].

5.6. Practical algorithm

5.7. Prenatal and preimplantation diagnosis

Once the familial variants have been identified, prenatal diagnosis and preimplantation genetic testing are possible [1, 8]. In a family with GLIS3, a sibling born after the affected twins was identified by amniocentesis as heterozygous [39]. In the coding PTF1A form, prenatal ultrasound allowed diagnosis through absence of the cerebellum and severe IUGR [13], and duodenal atresia ("double bubble") and IUGR led to a suspected prenatal diagnosis in Mitchell-Riley syndrome [22, 86]. The Oxford Desk Reference states that prenatal diagnosis of ND is requested mainly for the syndromic forms that may be fatal in infancy [51].

6. Differential Diagnosis

The differential is made at two levels: between the forms of the class, by extrapancreatic features and exocrine function, and between the class and other causes of hyperglycaemia in the infant or of digestive, cardiac and brain malformations. Table 6.1 summarises the main differentials.

The first step distinguishes diabetes with exocrine insufficiency (agenesis due to PTF1A, GATA6, PDX1, GATA4, CNOT1; Mitchell-Riley syndrome; some GLIS3 forms; WRS in 10.6%) from diabetes without exocrine insufficiency [3, 41]. Within the agenesis forms, heart disease points to GATA6 or GATA4, cerebellar agenesis to coding PTF1A, holoprosencephaly to CNOT1, duodenal and hepatobiliary anomalies to PDX1 or RFX6, and absence of extrapancreatic features with consanguinity to the PTF1A enhancer or ZNF808 [8, 11, 44]. Mitchell-Riley syndrome is distinguished from Martínez-Frías syndrome by ND and the usual absence of tracheo-oesophageal fistula, although one French patient with RFX6 variants also had tracheo-oesophageal fistula [22, 96]. In MEDS, the differential includes WRS (one of the IER3IP1 families had been published as WRS), coding PTF1A and NEUROD1 [30, 41, 90]. In enteric anendocrinosis, diarrhoea that stops with fasting and biopsy with chromogranin A separate it from the structural enteropathies, and the association with ND separates it from the other congenital diarrhoeas; in a series of 129 infants with congenital diarrhoea, NEUROG3 accounted for 3 of the 62 diagnoses [54, 93]. Before 6 months, the monogenic autoimmune forms (FOXP3, STAT3, LRBA) are at least as likely as T1D, and trisomy 21 is associated with an autoimmune diabetes distinct from common T1D [2].

7. Treatments and Risk Management Strategies

7.1. General principles and guideline recommendations

No form in this class has a treatment directed at the cause: sulfonylureas, which replace insulin in KATP channel ND, do not apply, because the β cell is absent, has not formed or dies from ER stress [1, 2, 41]. ISPAD states that, except for KATP forms and some patients with SLC19A2, all causes of ND are treated with subcutaneous insulin, and that children with pancreatic aplasia or hypoplasia also need exocrine supplements, without specifying doses [2]. GeneReviews states that there are no published clinical practice guidelines for PNDM, refers to ISPAD 2022 and bases its recommendations on the authors' experience [1]. Treatment is therefore multidisciplinary and organ-directed: paediatric endocrinology for diabetes and the thyroid; gastroenterology, nutrition and paediatric surgery for exocrine insufficiency, atresias and intestinal failure; hepatology; paediatric cardiology; paediatric neurology in MEDS; ophthalmology in GLIS3 syndrome; and medical genetics [1, 22, 41]. The Exeter cohort showed that the genetic diagnosis defines endocrine and exocrine treatment requirements: of the patients with an identified cause, 265 had diabetes treated with insulin alone and 50 disorders of exocrine pancreatic development requiring enzymes in addition to insulin [3].

7.2. Acute treatment

Presentation is treated with rapid rehydration and intravenous insulin infusion, especially if there is ketoacidosis, switching to subcutaneous insulin when the infant is stable and tolerating oral feeding [1]. The 2022 ISPAD consensus on insulin therapy recommends starting insulin soon after diagnosis, generally within 6 hours if there is ketonuria, to prevent metabolic decompensation and DKA (grade A) [108]. In infants on PN (Mitchell-Riley syndrome, enteric anendocrinosis), intravenous insulin or continuous subcutaneous infusion is adjusted to the glucose delivery rate of the PN: in the Portuguese case, insulin was given by continuous infusion with PN at a glucose rate of 5–6 mg/kg/min until 5 months and thereafter by pump [86]. In the French series, the basal rate during PN was more than ten times the rate without PN, and hypoglycaemia occurred when the PN was switched off [22]. Hypoglycaemia in these forms responds poorly to oral glucose and glucagon, and intravenous glucose (10%, 2 mL/kg) is the treatment that proved effective [22, 39].

7.3. Insulin therapy in the infant

Preparations and routes. Data on the most suitable preparations in small infants are scarce. Long-acting insulins without a marked peak (glargine, detemir) may work better, and some centres prefer continuous subcutaneous insulin infusion (CSII) as safer, more physiological and more precise; rapid-acting analogues (lispro, aspart) and regular insulin should generally be avoided as boluses, except in intravenous or continuous subcutaneous infusion, because of the risk of severe hypoglycaemia; NPH insulin has a shorter duration of action in infants; and, with very low requirements, insulin diluted to 5 or 10 U/mL may be used, with extreme caution to avoid dosing errors [1]. In the series of this class, CSII was started between 2 weeks and 5 months, with systems that suspend delivery when hypoglycaemia is predicted (MiniMed 640G), and was described as necessary in the first two years of Mitchell-Riley syndrome because of recurrent hypoglycaemia [22, 39]. In the North American GATA6, GATA4 and PDX1 registry, 8 of 11 children used CSII (one returned to injections) and 5 of 10 continuous glucose monitoring (CGM); the obstacles described were lack of subcutaneous tissue, eczema that prevented application of pump and sensor, and more unstable glucose with the pump than with injections in one child; another child had the pump disconnected overnight for much of early childhood [89].

Diluted insulin and CSII according to the ISPAD consensus. For infants and very young children who need fractional doses of less than 1 unit, the 2022 ISPAD consensus allows diluted insulin, prepared with the manufacturer's diluent (there are specific diluents for aspart, lispro and NPH); rapid-acting insulin can be diluted 1/10 (10 U/mL) or to 50 U/mL with sterile NPH diluent and stored for 1 month for use in pumps in infants and very young children, and 1/5 dilution (20 U/mL) has been used in automated insulin delivery systems in children aged 3 to 6 years; there are reports of dilution with normal saline when no manufacturer's diluent exists [108]. Among the consensus references for these dilutions are CSII in two patients with the homozygous PTF1A enhancer variant and a review of insulin therapy in ND [108]. The consensus stresses that dosing errors with non-standard concentrations can be serious, that the family should be trained before starting, that the same concentration must be ensured with each supply and that the family should inform the new team of the concentration in use when changing centre; syringes and devices must be calibrated for the concentration used (grade E) [108]. CSII is considered acceptable and successful even in small infants; insulin in the reservoir should be changed with the infusion set every 48 to 72 hours, and manufacturers recommend not keeping aspart and lispro in the reservoir, at room temperature, for more than 6 and 7 days, respectively [108]. As to authorised age, the consensus lists lispro with no lower age limit, aspart from 1 year, glulisine from 6 years, detemir from 1 year in the EU and from 2 years in the USA, and degludec from 1 year in the EU and the USA [108].

Doses. No guideline sets an insulin dose per kilogram for syndromic ND; the dose is titrated to glucose, and the ISPAD consensus notes that the daily dose varies widely between patients and over time and requires regular review (grade E), with the distribution through the day adapted to the glycaemic pattern (grade B) [108]. In type 1 diabetes, the same consensus gives usual requirements of less than 0.5 IU/kg/day in the partial remission phase, 0.7 to 1.0 IU/kg/day in prepubertal children outside that phase and 1 to 2 IU/kg/day during puberty; in basal-bolus regimens, basal insulin accounts for about 30% to 45% of the total daily dose and prandial insulin for about 55% to 70% [108]. These ranges have not been established for newborns or for ND, and their application to these forms is an extrapolation. The doses observed in the cohorts of this class, which serve as a reference and not as a recommendation, are in Table 7.1; the medians for pancreatic agenesis (0.7 U/kg/day) and the PTF1A enhancer (1 U/kg/day at the start and 0.8 U/kg/day at follow-up) lie within the prepubertal range for type 1 diabetes [8, 16].

Targets. GeneReviews proposes glucose between 70 and 180 mg/dL for more than 70% of the day and HbA1c below 7.0%, weighed against the risk of prolonged hypoglycaemia in children under 6 years, and notes that the American Diabetes Association has individualised the HbA1c target in children who cannot express hypoglycaemia symptoms [1]. In these forms, the targets were rarely achieved: mean HbA1c of 8.7% in the North American registry and only 2 of 28 patients with the PTF1A enhancer below 7.0% [16, 89]. In Mitchell-Riley syndrome, HbA1c may be falsified by repeated transfusions [22]. The 2022 ISPAD consensus on glycaemic targets and monitoring recommends, for all young people with diabetes, HbA1c below 7.0% (53 mmol/mol) (grade A), measured every 3 months, with a minimum of four measurements a year (grade E); on CGM, analysed over 14-day periods, more than 70% of the time between 70 and 180 mg/dL, less than 4% below 70 mg/dL, less than 1% below 54 mg/dL, less than 25% above 180 mg/dL and less than 5% above 250 mg/dL, with a coefficient of variation of 36% or less (grade B); and, in those on insulin, capillary self-monitoring at least 6 times a day (grade B), with a target of 70 to 180 mg/dL and, fasting, 70 to 144 mg/dL (grade E) [109]. The same consensus names ND among the situations in which strict targets may be unattainable or add treatment burden without benefit, and allows less strict targets when the standard target harms the well-being of the child or carers [109]. The minimum frequency of self-monitoring differs between sources: at least 4 times a day in GeneReviews and at least 6 in the ISPAD consensus [1, 109]. When red cell turnover is altered (acute blood loss, haemolysis, deficiency of or recovery from deficiency of iron, vitamin B12 or folate), HbA1c no longer reflects mean glucose, and fructosamine, which reflects the previous 2 to 3 weeks, or 1,5-anhydroglucitol may be the only laboratory alternative [109].

Regulatory status of insulins. No insulin is authorised for newborns; use before the minimum age in the Summary of Product Characteristics (SmPC) is outside the terms of the marketing authorisation. Table 7.2 summarises the labels consulted.

7.4. Pancreatic enzyme replacement

Indication. Pancreatic enzyme replacement is indicated in the forms with exocrine insufficiency: GeneReviews indicates it for HNF1B, PDX1, PTF1A and RFX6, and the Exeter cohort for GATA6, PTF1A, PDX1 and GATA4 [1, 3]. Exocrine insufficiency may be subclinical and detected only by faecal elastase, even in patients who have never received enzymes [4]. In the Exeter agenesis cohort, 82 patients were receiving enzymes (Creon®) at the last assessment [8]. In Mitchell-Riley syndrome, efficacy is debated: some authors consider that the small pancreas retains exocrine function, and at autopsy the exocrine pancreas appeared normal; in the French series, enzymes had no effect in two patients and a partial effect in two, with about 5000 IU/kg/day and persistent steatorrhoea (6 g/24 h; normal 0.5–3 g/24 h) [22]. In the PTF1A enhancer series, replacement was given to all, and cholestasis resolved with enzymes in 3 patients [16]. In the North American GATA6, GATA4 and PDX1 registry, the 8 children with exocrine insufficiency received enzymes (Creon® in most) and vitamins; in the 6 with longitudinal data, the dose ranged from 500 to 10,000 U lipase/kg/meal, generally remained constant with age, and 4 children, all with complete agenesis, at some point exceeded 2500 U lipase/kg/meal [89]. Parents reported better weight gain with higher doses, and two considered the dose insufficient [89]. The authors propose repeated laboratory reassessment of dose adequacy and teaching families the signs of worsening exocrine function [89].

Dose. There is no published dose specific to pancreatic agenesis. The reference adopted is the guideline for exocrine pancreatic insufficiency in cystic fibrosis (CF), an analogous condition treated with the same medicines, with the extrapolation declared; the labels are given for comparison (Table 7.3). The doses above 2500 U lipase/kg/meal described in the North American registry are clinical practice observations, above the maximum of the guideline and labels, and do not constitute a validated dose; the CF guideline assesses adequacy of replacement by growth and nutritional status, at every visit in infants and every 3 months in children and adolescents, and reserves annual faecal elastase for patients with pancreatic sufficiency [89, 107].

Administration and safety. Enzymes are taken during or immediately after meals; in infants, the capsule is opened immediately before each breastfeed or each 120 mL of formula, and the microspheres are given with a small amount of soft acidic food (pH 4.5 or less in the FDA label, below 5.5 in the Portuguese SmPC; apple purée, banana, yoghurt) or directly into the mouth, followed by milk, without mixing in the bottle, without crushing and checking that nothing remains in the mouth [115, 116]. The guideline allows unprotected powder temporarily when the infant refuses the microspheres, and states that proton pump inhibitors may improve efficacy [107]. Doses above 6000 U lipase/kg/meal have been associated with colonic stricture (fibrosing colonopathy) in children under 12 years with CF, and patients on more than 10,000 U/kg/day should be assessed for this complication; very high doses have been associated with hyperuricosuria and hyperuricaemia [115, 116]. Hydration should be adequate, because dehydration worsens constipation [116]. Adequacy of replacement is assessed by growth and nutritional status at every visit in infants, every 3 months in children and adolescents and every 6 months in adults (low grade of evidence) [107].

Regulatory status. Creon® (pancrelipase; AbbVie) is indicated in the United States for exocrine pancreatic insufficiency in adults and children, with no age restriction, with paediatric efficacy supported by two controlled trials in CF patients aged 7 years or more and an open-label study of 18 children aged 4 months to 6 years (7500 U lipase/kg/day on average) [115]. In Portugal, Kreon® 10,000 and Kreon® 25,000 (pancreatin; Viatris Healthcare) have been authorised since 1985 and 1999 for exocrine pancreatic insufficiency in adults and children, in a non-exhaustive list of causes that does not include pancreatic agenesis [116]. In Brazil, Creon® (pancreatin; Abbott) has had active registration since 30 May 2011 [111]. Use in pancreatic agenesis falls within the generic indication "exocrine pancreatic insufficiency" in all three labels; the dose in infants follows the CF guideline by extrapolation.

7.5. Fat-soluble vitamins and nutrition

GeneReviews recommends measuring fat-soluble vitamins every 6 months in patients with known exocrine insufficiency [1]. In the absence of a dose for pancreatic agenesis, the CF guideline (analogous condition) proposes: initial vitamin D3 400 IU/day in infants (up to 1000 IU/day) and 800 IU/day at other ages (up to 2000 IU/day from 1 to 10 years and 4000 IU/day thereafter), with a 25-hydroxyvitamin D target of at least 20 ng/mL (50 nmol/L); vitamin E 50 IU/day before 12 months and 100–400 IU/day thereafter, with a target α-tocopherol:cholesterol ratio above 5.4 mg/g; vitamin K1 0.3–1.0 mg/day in infants and 1–10 mg/day at other ages; vitamin A adjusted to serum levels, with β-carotene 1 mg/kg/day (maximum 50 mg/day) for 12 weeks and then maintenance up to 10 mg/day; annual reassessment and 3 to 6 months after each change [107]. In the North American registry, 7 of 9 children assessed had fat-soluble vitamin deficiency, mainly vitamin D at the lower limit, but irregular screening and the lack of values prior to supplementation prevent estimation of the true prevalence [89]. In Mitchell-Riley syndrome, all French patients had vitamin K deficiency, and two maintained low factor VII despite oral and intravenous supplementation [22]. Caloric intake should be high to achieve weight gain [1]; in the CF guideline, energy requirements range from 1.1 to 2 times those recommended for the healthy population [107]. In the North American registry, 6 of 10 children received caloric supplements and 5 of 10 a gastrostomy; continuous overnight gastrostomy feeding added calories and stabilised overnight glucose, was maintained in two children to prevent nocturnal hypoglycaemia and made it easier to correct hypoglycaemia, which families found more effective when they gave enzymes with the food [89]. The authors recommend close monitoring of weight and growth and considering enteral nutrition when progress is insufficient [89].

7.6. Parenteral nutrition, surgery and transplantation

Enteric anendocrinosis. The diarrhoea and malabsorption require prolonged PN from the neonatal period, with careful fluid and electrolyte management, and enteral autonomy is rarely achieved; the diabetes requires lifelong insulin, and deficient pituitary axes are replaced when there is combined deficiency [41]. The three boys of the original description were discharged on PN and limited enteral feeding [27], and the infant diagnosed at 4 weeks started PN at 3 months [103]. Management is based on reports and series, without trials or guidelines [41].

Mitchell-Riley syndrome. Duodenal atresia is corrected in the first week (duodenoduodenostomy; Ladd procedure if there is malrotation) [22, 86]. PN was required in the first year to maintain growth and sometimes for longer; early weaning attempts caused glycaemic decompensation, and enteral nutrition was introduced slowly, with monitoring of intestinal tolerance and with the aim of protecting the liver and preventing feeding disorders [22]. Surgical removal of small bowel segments with heterotopic gastric mucosa improved malabsorption, poor weight gain and anaemia in two sisters, and the authors consider gastric heterotopia an actionable component of the syndrome [76]. Intermittent cholestasis was treated with ursodeoxycholic acid; an infant with liver failure received a liver transplant at 7 months and, 8 days later, about 700,000 islets from the same donor injected into the spleen, without success (C-peptide 0.12 nmol/L and unchanged insulin requirements) [22]. A Romani child with severe intestinal failure received a multivisceral transplant at 2 years (stomach, duodenum, small bowel, colon, liver and pancreas), without surgical complications, with normal gastrointestinal, hepatic and pancreatic function more than 10 years later [97]. Multivisceral transplantation is described as an option in selected cases, on the basis of a single patient [97].

GLIS3 syndrome. Chronic liver disease with biliary cirrhosis led to isolated liver transplantation in two children, after multiorgan transplantation (liver, pancreas and, in some cases, kidney) had been considered; the postoperative course was difficult because of infections, renal disease and labile diabetes, and the authors recommend considering transplantation when hepatic complications are life-limiting [101]. Congenital glaucoma is treated with surgery (bilateral trabeculotomy at 6 weeks and reoperation at 18 months in one of the twins) and, according to GeneGuidance, with ocular hypotensive drops (timolol, latanoprost, dorzolamide), under paediatric ophthalmology guidance [39, 41].

PTF1A enhancer. Cholestasis resolved with pancreatic enzymes in 3 patients and with ursodeoxycholic acid in 11 [16]. Anaemia required repeated transfusions in 12 of 25 patients and had no identified cause [16].

7.7. Levothyroxine for congenital hypothyroidism in GLIS3 syndrome

The levothyroxine dose follows the 2020–2021 European consensus on congenital hypothyroidism (ENDO-ERN, endorsed by the European Society for Paediatric Endocrinology): levothyroxine alone, orally, once daily, started as early as possible and at the latest by 2 weeks of life; initial dose up to 15 µg/kg/day, 10 to 15 µg/kg/day in severe hypothyroidism (free T4 below 5 pmol/L) and about 10 µg/kg/day in mild hypothyroidism (free T4 above 10 pmol/L), with 5 to 10 µg/kg if free T4 is normal before treatment (recommendations 1/++0); the consensus prefers the branded product to the generic (2/++0) [45]. The first reassessment is 1 to 2 weeks after starting (at the latest 1 week if the dose is 50 µg/day or more), then every 2 weeks until TSH normalises, every 1 to 3 months up to 12 months, every 2 to 4 months between 12 months and 3 years and every 3 to 6 months until the end of growth, with reassessment 4 to 6 weeks after each dose change [45]. In the published GLIS3 cases, the initial dose was 15 µg/kg/day in the twins, with normal function within 2 weeks, and 10 µg/kg/day in the Turkish patient, who in adolescence needed more than 3 µg/kg/day (more than 100 µg/m²/day) because of TSH resistance, with a normal-sized thyroid and raised thyroglobulin [39, 100]. Levothyroxine is a routinely used medicine in paediatrics; the dose follows the guideline.

7.8. Epilepsy and neurodevelopment in MEDS

Epilepsy in MEDS is often refractory to several antiseizure medications; one patient with MEDS1 had slow and incomplete improvement with numerous drugs [41, 90]. In an infant girl with MEDS1 and refractory infantile spasms, the ketogenic diet was titrated up to a 3:1 ratio and maintained ketosis without DKA; during intercurrent illnesses there was rapid decompensation and mild DKA due to delayed treatment, and the diet was stopped after 5 months; the authors stress the need for close monitoring [106]. In the same patient, the diabetes became apparent after corticosteroid therapy for the spasms [106]. In the YIPF5 forms with hyperglycaemia only during intercurrent illness, glucose should be monitored during fever [41, 65]. Early intervention and rehabilitation are the measures for developmental delay [41].

7.9. Liraglutide for protracted diarrhoea in Mitchell-Riley syndrome (off-label)

Of the two Portuguese patients with congenital GLP-1 deficiency (section 3.3), only the first received liraglutide, off-label and with ethics committee authorisation, as rescue therapy for intractable diarrhoea [87]. Treatment began at 5 years, as an inpatient, in the morning, after the overnight PN had been stopped and before daytime oral feeding, taking into account the half-life of about 12 hours in children: initial dose 0.12 mg subcutaneously once daily, with increments of 0.12 mg on alternate days; at 0.36 mg/day vomiting appeared, which settled at 0.30 mg/day, and after discharge the dose returned to 0.36 mg/day without apparent adverse effects; there was no abdominal pain, headache, hypoglycaemia or signs of pancreatitis [87]. Six days after the start, stools fell from 5 to 8 to 1 to 3 a day, with firmer stools, and the perineal dermatitis disappeared; at 12 months there were still 2 to 3 stools a day, but appetite remained poor and weight gain insufficient, the insulin dose was similar, with no improvement in glycaemic control, and the patient remained predominantly dependent on PN [87]. This is a single patient, with no comparison group, and reduction of PN dependence and improvement in growth were not demonstrated; the authors suggest that somatostatin receptor agonists may be of interest, without having tested them [87]. Liraglutide (Victoza®) is authorised in the United States and the EU for type 2 diabetes from 10 years of age, with an initial paediatric dose of 0.6 mg subcutaneously once daily for at least one week, increasing to 1.2 mg and, if needed, to 1.8 mg; it should not be used in T1D [117, 118]. The dose used in the Portuguese patient (0.12 to 0.36 mg/day) was lower than the initial label dose for T2D (0.6 mg/day), and use in a 5-year-old child and for a gastrointestinal indication is outside the terms of the authorisation, with no established dose [87, 117], and the French series proposes studying modulators of the enteroendocrine system as a therapeutic line [22, 87].

7.10. Regulatory table

7.11. Clinical studies and recent developments

There are no interventional clinical trials directed at these syndromes. A ClinicalTrials.gov search for neonatal diabetes, pancreatic agenesis, Mitchell-Riley, Wolcott-Rallison, enteric anendocrinosis and monogenic diabetes, on 31 October 2026, identified registries and observational studies, including the ENDURE study (NCT06478121, University of Exeter), which since 11 November 2025 has been recruiting about 50 participants by genotype to study biomarkers and physiology, without therapeutic intervention [119]. A negative search does not prove the absence of trials. Table 7.4 therefore brings together the observational and preclinical studies that underpin practice and the developments of the last 12 months, with the level of evidence of each.

These studies are observational and small; the doses and glycaemic control rates described reflect the practice of different centres and do not allow comparisons between treatments. The compounds that protected iPSC-derived β cells from ER stress (ISRIB, imeglimin, GLP-1 receptor agonists) have not been tested in patients with these syndromes, and there is no approved disease-modifying treatment for ER stress diabetes [34]. No regulatory authorisation, opinion or application specific to these disorders was identified in the sources consulted.

7.12. Agents and circumstances to avoid, pregnancy, perioperative care and transition

Rapid-acting analogues and regular insulin as boluses should generally be avoided in infants, because of the risk of severe hypoglycaemia, except by continuous infusion [1]. Diluted insulin requires protocols to avoid dosing errors [1]. Enzyme doses above the label maximums require investigation, because of the risk of fibrosing colonopathy [115]. Corticosteroid therapy may unmask or worsen diabetes in MEDS, and the ketogenic diet increases the risk of DKA during intercurrent illness [106]. In WRS, general anaesthesia and drugs or vaccines that are not strictly necessary may precipitate decompensation [36, 41].

Women with PNDM who become pregnant should be followed by endocrinology and maternal-fetal medicine, with tight glycaemic control and high-resolution ultrasound and fetal echocardiography; GeneReviews notes that a meta-analysis associated glyburide (glibenclamide) with more neonatal hypoglycaemia [1]. In dominant forms with adult expression (GATA6, heterozygous RFX6), pregnancy may reveal or worsen diabetes, and offspring have a 50% risk of inheriting the variant (section 8) [4, 63].

In the perioperative period (correction of atresias, cardiac surgery in GATA6 forms, transplants), insulin should be given by intravenous infusion with glucose, and requirements increase with surgical and infectious stress: in a patient with GATA6 without chronic treatment, cardiac surgery triggered hyperglycaemia, and in the twins with GLIS3 infections tripled or quadrupled insulin requirements [4, 39]. Transition to adult care should include review of the genetic diagnosis, reproductive counselling and maintenance of exocrine, hepatic, renal and cardiac surveillance according to the gene [1].

8. Reproductive Options and Family Counselling

8.1. Recurrence risks by mechanism

In the autosomal recessive forms (PDX1, PTF1A and its enhancer, RFX6, GLIS3, NEUROG3, IER3IP1, YIPF5, TMEM167A, ZNF808, ONECUT1, EIF2AK3), the parents of an affected child are presumed heterozygous, and each sibling has, at conception, a 25% risk of being affected, 50% of being heterozygous and 25% of inheriting neither variant; the offspring of an affected individual are obligate heterozygotes, and the parents' siblings have a 50% risk of being heterozygous [1]. Heterozygotes may have milder diabetes in some genes: MODY4 in PDX1, reduced-penetrance MODY in RFX6 (27% at 25 years) and an increased risk of type 2 diabetes in GLIS3 and ONECUT1 [1, 61, 63, 80]. In the family of a patient with TMEM167A, both parents had diabetes diagnosed at 40 and 48 years, without the relationship with the variant having been studied [34]. Consanguinity is the rule: 92% of families with PTF1A enhancer variants and 30 of 31 individuals with ZNF808 had related parents [16, 44].

In the autosomal dominant forms (GATA6, GATA4, CNOT1, ACTB), most cases are de novo, but GATA6 may be inherited from a parent with adolescent- or adult-onset diabetes, with isolated heart disease or without manifestations; in the 2013 series, five of eight variants with parents tested were inherited, and the parents' variants had arisen de novo in germ cells [4]. Each child of a carrier has a 50% risk of inheriting the variant, with unpredictable expression, from pancreatic agenesis to isolated heart disease [1, 4]. In GATA4, pancreatic penetrance is lower than cardiac penetrance, and the molecular diagnosis indicates, for the offspring of a carrier, a risk of ND, of congenital heart disease or of both; in a family with the p.R319W variant, the father and sister of the proband with pancreatic agenesis were carriers with heart disease and without diabetes [7]. When the variant is not found in the parents' blood, the risk to siblings is low but higher than that of the population, because of the possibility of gonadal mosaicism; a parent mosaic for a GATA6 variant (about 20% in leucocytes) did not have diabetes but had undergone surgery for a patent ductus arteriosus [1, 4]. The CNOT1 variants associated with pancreatic agenesis have always been de novo [19]. In IPEX syndrome, which is X-linked, counselling follows the pattern of X-linked inheritance.

8.2. Testing of parents and relatives at risk

GeneReviews recommends, in all probands, molecular testing of the parents and clinical assessment for diabetes (HbA1c, glucose monitoring or oral glucose tolerance test) [1]. In asymptomatic relatives at risk, targeted molecular testing identifies those who need surveillance; if the familial variant is not known, screening is done with HbA1c or fasting and postprandial glucose, and continuous glucose monitoring has been used [1]. In families with GATA6 and GATA4, assessment of relatives should also be cardiac and exocrine, because normal glucose in a carrier does not exclude the diagnosis [41]. In families with RFX6, heterozygotes should know that they may develop MODY-type diabetes [41, 63].

8.3. Prenatal diagnosis, preimplantation testing and planning

The best time to define the risk and discuss prenatal diagnosis and preimplantation genetic testing is before pregnancy, and counselling should be offered to young adults who are affected or at risk [1]. Once the familial variants have been identified, both options are possible, and the genetic diagnosis is the basis for access to them [1, 8]. In ND, prenatal diagnosis is requested mainly for the syndromic forms that may be fatal in infancy [51]. Ultrasound may suggest the diagnosis in the malformative forms (absent cerebellum and severe IUGR in the coding PTF1A form; "double bubble" and IUGR in Mitchell-Riley syndrome; holoprosencephaly in the CNOT1 form; heart disease in GATA6) [4, 13, 19, 86]. Differences in perspective on prenatal diagnosis between professionals and families should be discussed, and the decision belongs to the parents [1]. GeneReviews advises storing DNA from probands without a molecular diagnosis, for future reanalysis [1].

8.4. Content of counselling

Counselling should explain the mechanism, the lifelong nature of insulin and, where applicable, of enzyme replacement, the recurrence risk, the manifestations that may appear later (seizures and developmental delay in MEDS, liver and renal disease in GLIS3, childhood diabetes in RFX6, pituitary deficiencies in NEUROG3), and the significance of results in heterozygous relatives [1, 3].

9. Special Considerations: Population Diversity and Epidemiology

9.1. Incidence and what each figure measures

The incidence of ND estimated by GeneReviews is 1 in 90,000 to 1 in 260,000 live births, half of it permanent [1]; the Oxford Desk Reference gives about 1 in 200,000 births [51], and the 2015 Exeter cohort cites about 1 in 100,000 [3]. Syndromic forms account for about 10% or more of ND, according to the references cited in two case reports [39, 100]. Pancreatic agenesis accounted for at least 4.8% (125/2601) of probands referred to Exeter for ND, a figure the authors consider an underestimate, because exocrine insufficiency is difficult to detect in the newborn and 77 patients with variants in agenesis genes had no clinical confirmation [8]. In the series that established the genes, GATA6 accounted for 56% (15/27) of patients with pancreatic agenesis born to parents without diabetes, versus 1 of 27 with PTF1A and none with PDX1 [40], and GATA4 for 0.5% (4/867) of an international ND series [7]; these are proportions in cohorts referred to a laboratory, not population prevalences. The global prevalence of pancreatic agenesis and of each syndrome in the class is estimated by Orphanet at less than 1 per 1,000,000 [8, 41]. In the 2013 Exeter cohort, GATA6 accounted for 3% (24/795) of ND and 54% (21/39) of pancreatic agenesis cases [4]; in Table 1 of GeneReviews, GATA6, GLIS3 and RFX6 each account for about 5% of PNDM [1]. The denominators differ (all ND, PNDM, agenesis), so the percentages cannot be compared directly.

The number of published patients per entity gives the order of magnitude of rarity: 12 cases of MEDS1 by 2025, 6 of MEDS2 in the original series, 6 of TMEM167A MEDS, 16 cases of Mitchell-Riley by 2022, 22 of GLIS3 from 16 families and 11 countries by 2022, 17 cases of NEUROG3 by 2022, 45 individuals with CNOT1 variants and 31 with ZNF808 in the Exeter cohort [19, 22, 33, 34, 44, 91, 99, 102].

9.2. Distribution by region and founder populations

Consanguinity alters the distribution of causes. In the international cohort, KATP channel forms accounted for 46% of cases with non-consanguineous parents and 12% of those with consanguineous parents, and in the latter WRS was the most frequent cause (24%) [3]. Case reports estimate ND at about 1 in 100,000 live births in Europe and 1 in 30,000 in the Middle East, with PNDM about 10 times more frequent in that region [39]. Recent regional cohorts confirm the predominance of recessive and syndromic forms in populations with high consanguinity. In Sudan, with consanguinity of 63.6% among 88 patients with monogenic diabetes, EIF2AK3 was the most frequent cause of ND, and ZNF808, NARS2 and FICD accounted for 8.5%, 4.2% and 1.4% of the total [120]. In Bangladesh, among 28 patients with ND, 13 from consanguineous families, EIF2AK3 accounted for 19% of PNDM, after ABCC8 and KCNJ11 [121]. In Morocco, 10 infants all presented in DKA, with consanguinity in 55% and severe syndromic forms [122]. In Thailand, among 14 patients, GATA6 and EIF2AK3 variants were found among other causes [123].

The founder or recurrent variants described are the PTF1A enhancer variant chr10:g.23508437A>G (GRCh37), in Turkish, Syrian, Lebanese and Kurdish families [16, 17]; the NEUROG3 variant p.(Gln137Arg) in Bedouin families [93]; and the RFX6 variant p.His293LeufsTer7 in Finland, with about 1 carrier in 250 people, associated with type 2 and gestational diabetes in heterozygotes [75]. Mitchell-Riley syndrome and Martínez-Frías syndrome have been described repeatedly in Roma families from Portugal, Spain and France [21, 22, 52, 86, 97]; the sources do not allow a founder effect in this population to be asserted.

9.3. Data from Portugal and Brazil

In Portugal, the team at Hospital Dona Estefânia (Lisbon) described in 2015 an infant girl born to Roma first-cousin parents with Mitchell-Riley syndrome and the RFX6 variant p.(Arg181Trp) in the homozygous state, with genetic testing performed in Exeter [86], and in 2021 the same team reassessed this patient at 6 years and described a second girl, born to consanguineous parents, homozygous for the same variant and with no known kinship, both with congenital GLP-1 deficiency; only the first was treated with liraglutide [87]. One of the patients in the 2022 French series was of Portuguese and Spanish origin, with the same variant [22]. The sources consulted contain no Portuguese or Brazilian epidemiological data on these syndromes. In Brazil, pancreatin (Creon®), basal and rapid-acting insulins and liraglutide have active ANVISA registration [111].

9.4. Special populations

Preterm and very low birth weight newborns with digestive malformations have the most severe course, such as the 880 g preterm infant with Mitchell-Riley syndrome who died on day 195 [88]. In children on prolonged PN, the risk of catheter-associated sepsis, cholestasis and vitamin K deficiency dominates the course [22]. In children with severe developmental delay and epilepsy (MEDS, CNOT1), communication of hypoglycaemia symptoms is limited, which justifies individualised glycaemic targets and continuous monitoring [1, 34]. In resource-limited settings, such as that described in Morocco, access to genetic testing depends on the families' financial means [122].

9.5. European Reference Networks and patient associations

The relevant European network is Endo-ERN, which brings together more than 100 reference centres in 27 Member States and Norway and has a thematic group dedicated to genetic disorders of glucose and insulin homeostasis, with a subgroup for rare diabetes, and another for the thyroid [46]. Depending on the manifestations, ERNICA (congenital digestive anomalies and intestinal failure), ERN RARE-LIVER (liver disease), EpiCARE (rare epilepsies) and ERN ITHACA (malformation syndromes) are involved [47–50]. The Exeter laboratory maintains the Diabetes Genes website, with information for patients and professionals and access to genetic testing for ND [51, 124], and the University of Chicago maintains the North American monogenic diabetes registry [1, 125]. In Portugal, the Associação Protectora dos Diabéticos de Portugal (APDP) and, in Brazil, the Sociedade Brasileira de Diabetes provide information and support in diabetes [126, 127].

10. Illustrative Clinical Case

10.1. Twins with GLIS3 syndrome

London and colleagues described in 2021 dizygotic twins, conceived by in vitro fertilisation, born to Arab Muslim first-cousin parents at 37 weeks [39]. The boy weighed 1984 g (−2.53 SD) and was admitted to the neonatal intensive care unit because of low weight; he had glucose between 61 and 443 mg/dL in the first 5 days and received PN for the first 4 days. Insulin was below 0.2 µIU/mL, C-peptide 0.1 ng/mL and pancreatic autoantibodies negative. He started intravenous regular insulin on day 5 (0.24–1.2 U/kg/day), replaced by diluted subcutaneous insulin lispro before meals (0.03–0.05 U) and NPH (0.2–0.4 U/kg); because of glycaemic lability, he switched at 2 weeks, weighing 1730 g, to CSII with a system that suspends delivery when hypoglycaemia occurs, with the catheter in the buttock and a basal rate of 0.3–0.6 U/kg/day of regular insulin and boluses of 0.025 U. In the first week he had primary hypothyroidism (TSH above 150 µIU/mL; free T4 below 1.3 pmol/L) and started levothyroxine at 15 µg/kg/day, with normal thyroid function within 2 weeks; the thyroid was of normal size and position. Corneal clouding led to the diagnosis of bilateral congenital glaucoma, operated on by trabeculotomy at 6 weeks and again at 18 months. Renal ultrasound showed normal-sized, echogenic kidneys with small cortical cysts, with normal renal function; there was transient jaundice with raised gamma-glutamyltransferase, which normalised, the pancreas was normal on ultrasound and faecal elastase (396 µg/g) excluded exocrine insufficiency. Insulin requirements tripled or quadrupled during admissions for infection at 4 and 9 months, and at 2.4 years, during gastroenteritis in which he was 3 days without the pump, he had severe and prolonged hypoglycaemia that did not correct with three intramuscular injections of 1 mg glucagon and required intravenous glucose in intensive care; adrenal insufficiency was excluded. At 3 years he weighed 14.4 kg (40th centile), measured 98 cm (60th centile) and had HbA1c of 6.5% on CSII with lispro diluted to 10 U/mL (0.7 U/kg/day), mild global developmental delay and a normal brain MRI [39].

His sister weighed 1616 g (−3.5 SD) and had a similar course: ND, primary hypothyroidism and congenital glaucoma, CSII at 16 days, recurrent hypoglycaemia even after a few hours without insulin, transient elevation of liver enzymes and bilirubin, and positive anti-GAD antibodies (7.2 U/mL) with negative anti-insulin and anti-islet antibodies at 23 days. At 3 years she had HbA1c of 7.0% on 0.9 U/kg/day and continuous monitoring, and mild to moderate global developmental delay [39]. GLIS3 sequencing identified, in both, the nonsense variant NM_001042413.2:c.2392C>T p.(Gln798Ter) in the homozygous state, reported at 5 weeks; the parents were heterozygous, and a sibling born a year later was identified by amniocentesis as heterozygous [39, 72]. Figure 1 of the article shows the boy's pump and continuous monitoring report, with several episodes of pump disconnection (https://pmc.ncbi.nlm.nih.gov/articles/PMC8169976/figure/f1/) [39].

The case brings together the practical points of the class: the combination of diabetes with hypothyroidism on newborn screening and corneal clouding, which directed testing towards GLIS3; the genetic diagnosis in the first weeks; diluted insulin and CSII in infants under 2 kg; and the inadequate response to glucagon in severe hypoglycaemia. The insulin doses and regimens described reflect the centre's practice and do not constitute a recommendation; GeneReviews allows insulin diluted to 5 or 10 U/mL with extreme caution [1, 39].

10.2. Other published cases with figures and a Portuguese case

The Turkish patient with a homozygous deletion of exons 10–11 of GLIS3, followed up to 17 years, illustrates the form without parenchymal disease, with facial dysmorphism and multiple naevi shown in Figure 1 of the article (https://pmc.ncbi.nlm.nih.gov/articles/PMC10683540/figure/f1/) [100]. Figure 3 of the French Mitchell-Riley syndrome series shows the continuous monitoring traces of two patients, with daytime hypoglycaemia when disconnected from PN and nocturnal hyperglycaemia during PN, with a basal rate about ten times higher (https://pmc.ncbi.nlm.nih.gov/articles/PMC9257252/figure/f3/) [22]. In the Portuguese case of Mitchell-Riley syndrome, ultrasound at 33 weeks showed IUGR and "double bubble"; the girl was born at 35 weeks weighing 1370 g, had hyperglycaemia of 415 mg/dL on day 2 with insulin of 0.12 µIU/mL and C-peptide below 0.1 ng/mL, and laparotomy on day 4 showed type II duodenal atresia, annular pancreas, malrotation, gallbladder agenesis and perijejunal tissue resembling ectopic pancreas; she had transient cholestasis, diarrhoea with enteral advancement, no response to enzymes and dependence on amino acid formula and PN (about 60% of energy) at 7 months, with insulin by infusion and then by pump [86]. The article has no figures. At 6 years, the same patient remained on CSII and overnight PN, and the response of the diarrhoea to liraglutide is described in section 7.9 [87].

11. Prognosis and Long-Term Follow-up

11.1. Prognosis by entity

Prognosis depends mainly on the extrapancreatic manifestations. In isolated pancreatic agenesis (PTF1A enhancer, some cases of GATA6, PDX1, ZNF808), survival with insulin and enzymes is the rule, but growth is compromised (short stature in 69% of patients with the PTF1A enhancer, oldest patient aged 18 years) and glycaemic control is difficult, with median HbA1c of 9.3% [16]. In GATA6, heart disease and neurocognitive deficit (present in 11 of 24 probands) determine the course, and diabetes may only appear in adolescence or adulthood in carriers with mild expression [4]. In the North American registry of 11 patients with GATA6, GATA4 or PDX1, with current recorded ages between 3 and 35 years, HbA1c remained above target with no trend with age, none had DKA after diagnosis, and difficulty gaining weight was the most persistent problem, in one patient with GATA6 only stabilising at around 30 years [89]. In the GATA4 series, a preterm infant with agenesis died on day 4, and the other four were aged between 1 and 17 years at the last assessment [7]. In the coding PTF1A forms, the cases described survived up to 4 months [13, 16]. In Mitchell-Riley syndrome, 7 of 12 homozygous patients published by 2022 died before 6 months, but with multidisciplinary intensive care, PN, CSII and, in one case, multivisceral transplantation, there are survivors into adolescence and beyond 10 years after transplantation [22, 97]. In GLIS3 syndrome, liver and renal disease limit survival, and patients without parenchymal disease reach adulthood [24, 100]. In enteric anendocrinosis, PN dependence and the risk of sepsis dominate childhood, and endocrine deficiencies appear progressively [27, 41]. In MEDS, mortality is high (5 of 6 patients with TMEM167A died between 2 months and 14 years; deaths at 18 and 27 months and at 8 years in MEDS1; death at 1.3 years and of four siblings in infancy in a family with MEDS2), with exceptions of preserved development [29, 32, 34, 90]. The microvascular complications described in PNDM in general (microalbuminuria in 4% to 19% and retinopathy in 6%) come mainly from the KATP channel forms [1].

11.2. Surveillance table

11.3. Education and prevention

Families should learn to recognise hypoglycaemia, know that oral glucose and glucagon may not suffice and that it is necessary to attend the emergency department for intravenous glucose, and increase monitoring during infections, when insulin requirements rise several-fold [22, 39]. Pump disconnection and intercurrent illnesses are the times of highest risk [39, 106]. Diluted insulin requires written protocols to avoid dosing errors [1].

12. Summary and Final Considerations

Syndromic neonatal diabetes brings together rare monogenic forms in which diabetes in the first months results from a failure of development of the pancreas and neighbouring endodermal organs or from a defect of secretory trafficking shared by the β cell and the neuron. Pancreatic agenesis has a genetic diagnosis in 98% of cases, with a predominance of recessive variants of the distal PTF1A enhancer and dominant GATA6 variants; Mitchell-Riley syndrome associates diabetes with intestinal atresias and hepatobiliary malformations; GLIS3 syndrome associates it with congenital hypothyroidism, glaucoma and liver and renal disease; enteric anendocrinosis due to NEUROG3 begins with malabsorptive diarrhoea; and the microcephaly, epilepsy and diabetes syndromes, caused by IER3IP1, YIPF5 and, since 2025, TMEM167A, combine diabetes with severe microcephaly and refractory epilepsy.

Genetic testing should be requested at the diagnosis of any diabetes before 6 months, with a method that includes copy-number analysis, the non-coding PTF1A enhancer and recently described genes, because the result precedes the extrapancreatic features and defines endocrine and exocrine treatment. None of these forms responds to sulfonylureas. Treatment combines insulin, at doses of about 0.7 to 1 U/kg/day in the cohorts and often by pump with diluted insulin in small infants, enzyme replacement and fat-soluble vitamins with doses extrapolated from cystic fibrosis, levothyroxine according to the European consensus, parenteral nutrition, surgery for the malformations and, in selected cases, liver or multivisceral transplantation. Hypoglycaemia responds poorly to oral glucose and glucagon. Counselling depends on the mechanism: a 25% risk in the recessive forms, frequent in consanguineous families, and variable expression in the dominant GATA6 and GATA4 forms, in which a parent may have only adult diabetes or heart disease. There are no interventional trials; studies of stem cell-derived β cells point to drugs that reduce ER stress, still without data in patients.

13. References

1. De León DD, Pinney SE. Permanent Neonatal Diabetes Mellitus. 2008 Feb 8 [updated 2024 Nov 14; revised 2025 Sep 25]. In: Adam MP, Bick S, Mirzaa GM, et al., editors. GeneReviews®. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1447/. PMID: 20301620
2. Greeley SAW, Polak M, Njølstad PR, Barbetti F, Williams R, Castano L, et al. ISPAD Clinical Practice Consensus Guidelines 2022: The diagnosis and management of monogenic diabetes in children and adolescents. Pediatr Diabetes. 2022;23(8):1188-1211. doi:10.1111/pedi.13426. PMID: 36537518. PMC10107883.
3. De Franco E, Flanagan SE, Houghton JA, Lango Allen H, Mackay DJ, Temple IK, et al. The effect of early, comprehensive genomic testing on clinical care in neonatal diabetes: an international cohort study. Lancet. 2015;386(9997):957-963. doi:10.1016/S0140-6736(15)60098-8. PMID: 26231457. PMC4772451.
4. De Franco E, Shaw-Smith C, Flanagan SE, Shepherd MH, International NDM Consortium, Hattersley AT, Ellard S. GATA6 mutations cause a broad phenotypic spectrum of diabetes from pancreatic agenesis to adult-onset diabetes without exocrine insufficiency. Diabetes. 2013;62(3):993-997. doi:10.2337/db12-0885. PMID: 23223019. PMC3581234.
5. OMIM® Online Mendelian Inheritance in Man. #600001 HEART DEFECTS, CONGENITAL, AND OTHER CONGENITAL ANOMALIES; HDCA. Baltimore (MD): Johns Hopkins University; last edited 29/01/2018. Available from: https://omim.org/entry/600001. Accessed 31 October 2026.
6. Orphanet. Pancreatic hypoplasia-diabetes-congenital heart disease syndrome (ORPHA:2255). https://www.orpha.net/en/disease/detail/2255
7. Shaw-Smith C, De Franco E, Lango Allen H, Batlle M, Flanagan SE, Borowiec M, et al. GATA4 mutations are a cause of neonatal and childhood-onset diabetes. Diabetes. 2014;63(8):2888-2894. doi:10.2337/db14-0061. PMID: 24696446. PMC6850908.
8. Hinton KE, Flanagan SE, Houghton JAL, Hattersley AT, Owens NDL, De Franco E. Genetics of pancreatic agenesis: a 21-year international cohort study. Lancet Diabetes Endocrinol. 2026;14(7):539-541. doi:10.1016/S2213-8587(26)00072-0. PMID: 42225116.
9. OMIM® Online Mendelian Inheritance in Man. *600576 GATA-BINDING PROTEIN 4; GATA4. Baltimore (MD): Johns Hopkins University; last edited 01/12/2025. Available from: https://omim.org/entry/600576. Accessed 31 October 2026.
10. OMIM® Online Mendelian Inheritance in Man. #260370 PANCREATIC AGENESIS 1; PAGEN1. Baltimore (MD): Johns Hopkins University; last edited 23/10/2024. Available from: https://omim.org/entry/260370. Accessed 31 October 2026.
11. Jeffery N, Al Nimri O, Houghton JAL, Globa E, Wakeling MN, Flanagan SE, et al. Widening the phenotypic spectrum caused by pathogenic PDX1 variants in individuals with neonatal diabetes. BMJ Open Diabetes Res Care. 2024;12(6):e004439. doi:10.1136/bmjdrc-2024-004439. PMID: 39542526. PMC11575358.
12. Orphanet. Partial pancreatic agenesis (ORPHA:2805). https://www.orpha.net/en/disease/detail/2805
13. OMIM® Online Mendelian Inheritance in Man. #609069 PANCREATIC AND CEREBELLAR AGENESIS; PACA. Baltimore (MD): Johns Hopkins University; last edited 03/11/2023. Available from: https://omim.org/entry/609069. Accessed 31 October 2026.
14. Sellick GS, Barker KT, Stolte-Dijkstra I, Fleischmann C, Coleman RJ, Garrett C, et al. Mutations in PTF1A cause pancreatic and cerebellar agenesis. Nat Genet. 2004;36(12):1301-1305. doi:10.1038/ng1475. PMID: 15543146.
15. Orphanet. Permanent neonatal diabetes mellitus-pancreatic and cerebellar agenesis syndrome (ORPHA:65288). https://www.orpha.net/en/disease/detail/65288
16. Demirbilek H, Cayir A, Flanagan SE, Yıldırım R, Kor Y, Gurbuz F, et al. Clinical Characteristics and Long-term Follow-up of Patients with Diabetes Due To PTF1A Enhancer Mutations. J Clin Endocrinol Metab. 2020;105(12):dgaa613. doi:10.1210/clinem/dgaa613. PMID: 32893856. PMC7526731.
17. OMIM® Online Mendelian Inheritance in Man. #615935 PANCREATIC AGENESIS 2; PAGEN2. Baltimore (MD): Johns Hopkins University; last edited 22/10/2024. Available from: https://omim.org/entry/615935. Accessed 31 October 2026.
18. OMIM® Online Mendelian Inheritance in Man. #618500 HOLOPROSENCEPHALY 12 WITH OR WITHOUT PANCREATIC AGENESIS; HPE12. Baltimore (MD): Johns Hopkins University; last edited 29/09/2020. Available from: https://omim.org/entry/618500. Accessed 31 October 2026.
19. de Brouwer APM, Vissers L. CNOT1-Related Vissers-Bodmer Syndrome. 2026 Mar 26. In: Adam MP, Bick S, Mirzaa GM, et al., editors. GeneReviews®. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK621477/. PMID: 41911374.
20. Orphanet. Pancreatic agenesis-holoprosencephaly syndrome (ORPHA:556955). https://www.orpha.net/en/disease/detail/556955
21. OMIM® Online Mendelian Inheritance in Man. #615710 MITCHELL-RILEY SYNDROME; MTCHRS. Baltimore (MD): Johns Hopkins University; last edited 22/11/2024. Available from: https://omim.org/entry/615710. Accessed 31 October 2026.
22. Passone CGB, Vermillac G, Staels W, Besancon A, Kariyawasam D, Godot C, et al. Mitchell-Riley Syndrome: Improving Clinical Outcomes and Searching for Functional Impact of RFX-6 Mutations. Front Endocrinol (Lausanne). 2022;13:802351. doi:10.3389/fendo.2022.802351. PMID: 35813646. PMC9257252.
23. Orphanet. Hypoplastic pancreas-intestinal atresia-hypoplastic gallbladder syndrome (ORPHA:293864). https://www.orpha.net/en/disease/detail/293864
24. OMIM® Online Mendelian Inheritance in Man. #610199 DIABETES MELLITUS, NEONATAL, WITH CONGENITAL HYPOTHYROIDISM; NDH. Baltimore (MD): Johns Hopkins University; last edited 12/08/2016. Available from: https://omim.org/entry/610199. Accessed 31 October 2026.
25. Senée V, Chelala C, Duchatelet S, Feng D, Blanc H, Cossec JC, et al. Mutations in GLIS3 are responsible for a rare syndrome with neonatal diabetes mellitus and congenital hypothyroidism. Nat Genet. 2006;38(6):682-687. doi:10.1038/ng1802. PMID: 16715098.
26. Orphanet. Neonatal diabetes-congenital hypothyroidism-congenital glaucoma-hepatic fibrosis-polycystic kidneys syndrome (ORPHA:79118). https://www.orpha.net/en/disease/detail/79118
27. OMIM® Online Mendelian Inheritance in Man. #610370 DIARRHEA 4, MALABSORPTIVE, CONGENITAL, WITH DIABETES MELLITUS AND COMBINED PITUITARY HORMONE DEFICIENCY; DIAR4. Baltimore (MD): Johns Hopkins University; last edited 16/04/2025. Available from: https://omim.org/entry/610370. Accessed 31 October 2026.
28. Orphanet. Enteric anendocrinosis (ORPHA:83620). https://www.orpha.net/en/disease/detail/83620
29. OMIM® Online Mendelian Inheritance in Man. #614231 MICROCEPHALY, EPILEPSY, AND DIABETES SYNDROME 1; MEDS1. Baltimore (MD): Johns Hopkins University; last edited 16/04/2021. Available from: https://omim.org/entry/614231. Accessed 31 October 2026.
30. Poulton CJ, Schot R, Kia SK, Jones M, Verheijen FW, Venselaar H, et al. Microcephaly with simplified gyration, epilepsy, and infantile diabetes linked to inappropriate apoptosis of neural progenitors. Am J Hum Genet. 2011;89(2):265-276. doi:10.1016/j.ajhg.2011.07.006. PMID: 21835305. PMC3155199.
31. Orphanet. Primary microcephaly-epilepsy-permanent neonatal diabetes syndrome (ORPHA:306558). https://www.orpha.net/en/disease/detail/306558
32. OMIM® Online Mendelian Inheritance in Man. #619278 MICROCEPHALY, EPILEPSY, AND DIABETES SYNDROME 2; MEDS2. Baltimore (MD): Johns Hopkins University; last edited 16/04/2021. Available from: https://omim.org/entry/619278. Accessed 31 October 2026.
33. De Franco E, Lytrivi M, Ibrahim H, Montaser H, Wakeling MN, Fantuzzi F, et al. YIPF5 mutations cause neonatal diabetes and microcephaly through endoplasmic reticulum stress. J Clin Invest. 2020;130(12):6338-6353. doi:10.1172/jci141455. PMID: 33164986. PMC7685733.
34. Virgilio E, Tielens S, Bonfield G, Nian FS, Sawatani T, Vinci C, et al. Recessive TMEM167A variants cause neonatal diabetes, microcephaly, and epilepsy syndrome. J Clin Invest. 2025;135(22):e195756. doi:10.1172/jci195756. PMID: 40924476. PMC12618065.
35. OMIM® Online Mendelian Inheritance in Man. *620000 TRANSMEMBRANE PROTEIN 167A; TMEM167A. Baltimore (MD): Johns Hopkins University; last edited 11/08/2022. Available from: https://omim.org/entry/620000. Accessed 31 October 2026.
36. 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. p. 214–217.
37. OMIM® Online Mendelian Inheritance in Man. #226980 EPIPHYSEAL DYSPLASIA, MULTIPLE, WITH EARLY-ONSET DIABETES MELLITUS. Baltimore (MD): Johns Hopkins University; last edited 29/07/2025. Available from: https://omim.org/entry/226980. Accessed 31 October 2026.
38. Orphanet. Wolcott-Rallison syndrome (ORPHA:1667). https://www.orpha.net/en/disease/detail/1667
39. London S, De Franco E, Elias-Assad G, Barhoum MN, Felszer C, Paniakov M, et al. Case Report: Neonatal Diabetes Mellitus Caused by a Novel GLIS3 Mutation in Twins. Front Endocrinol (Lausanne). 2021;12:673755. doi:10.3389/fendo.2021.673755. PMID: 34093443. PMC8169976.
40. Allen HL, Flanagan SE, Shaw-Smith C, De Franco E, Akerman I, Caswell R, et al. GATA6 haploinsufficiency causes pancreatic agenesis in humans. Nat Genet. 2011;44(1):20-22. doi:10.1038/ng.1035. PMID: 22158542. PMC4062962.
41. GeneGuidance: gene records for IER3IP1, YIPF5, RFX6, GATA6, GATA4, GLIS3, NEUROG3, PTF1A and EIF2AK3. Accessed 31 October 2026.
42. Sansbury FH, Kirel B, Caswell R, Allen HL, Flanagan SE, Hattersley AT, et al. Biallelic RFX6 mutations can cause childhood as well as neonatal onset diabetes mellitus. Eur J Hum Genet. 2015;23(12):1744-1748. doi:10.1038/ejhg.2015.161. PMID: 26264437. PMC4795203.
43. Solorzano-Vargas RS, Bjerknes M, Wang J, Wu SV, Garcia-Careaga MG, Pitukcheewanont P, et al. Null mutations of NEUROG3 are associated with delayed-onset diabetes mellitus. JCI Insight. 2020;5(1):127657. doi:10.1172/jci.insight.127657. PMID: 31805014. PMC7030873.
44. Russ-Silsby J, Colclough K, Johnson MB, Wakeling MN, Owens NDL, Amaratunga SA, et al. Loss of function variants in the primate-specific gene ZNF808 cause neonatal, transient and adult-onset diabetes. EBioMedicine. 2026;124:106113. doi:10.1016/j.ebiom.2025.106113. PMID: 41500078. PMC12808904.
45. van Trotsenburg P, Stoupa A, Léger J, Rohrer T, Peters C, Fugazzola L, et al. Congenital Hypothyroidism: A 2020-2021 Consensus Guidelines Update-An ENDO-European Reference Network Initiative Endorsed by the European Society for Pediatric Endocrinology and the European Society for Endocrinology. Thyroid. 2021;31(3):387-419. doi:10.1089/thy.2020.0333. PMID: 33272083. PMC8001676.
46. European Reference Network on Rare Endocrine Conditions (Endo-ERN). Main Thematic Group 3: Genetic disorders of glucose & insulin homeostasis. Available from: https://endo-ern.eu/
47. ERN ERNICA — European Reference Network for rare inherited and congenital digestive disorders. Available from: https://ern-ernica.eu/
48. ERN RARE-LIVER — European Reference Network on Rare Hepatological Diseases. Available from: https://rare-liver.eu/
49. EpiCARE — European Reference Network for rare and complex epilepsies. Available from: https://www.epi-care.eu/
50. ERN ITHACA — European Reference Network on Rare Congenital Malformations and Rare Intellectual Disability. Available from: https://ern-ithaca.eu/
51. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 390–393.
52. OMIM® Online Mendelian Inheritance in Man. %601346 MARTINEZ-FRIAS SYNDROME. Baltimore (MD): Johns Hopkins University; last edited 10/04/2014. Available from: https://omim.org/entry/601346. Accessed 31 October 2026.
53. Smith SB, Qu HQ, Taleb N, Kishimoto NY, Scheel DW, Lu Y, et al. Rfx6 directs islet formation and insulin production in mice and humans. Nature. 2010;463(7282):775-780. doi:10.1038/nature08748. PMID: 20148032. PMC2896718.
54. Wang J, Cortina G, Wu SV, Tran R, Cho JH, Tsai MJ, et al. Mutant neurogenin-3 in congenital malabsorptive diarrhea. N Engl J Med. 2006;355(3):270-280. doi:10.1056/nejmoa054288. PMID: 16855267.
55. Weedon MN, Cebola I, Patch AM, Flanagan SE, De Franco E, Caswell R, et al. Recessive mutations in a distal PTF1A enhancer cause isolated pancreatic agenesis. Nat Genet. 2014;46(1):61-64. doi:10.1038/ng.2826. PMID: 24212882. PMC4131753.
56. De Franco E, Watson RA, Weninger WJ, Wong CC, Flanagan SE, Caswell R, et al. A Specific CNOT1 Mutation Results in a Novel Syndrome of Pancreatic Agenesis and Holoprosencephaly through Impaired Pancreatic and Neurological Development. Am J Hum Genet. 2019;104(5):985-989. doi:10.1016/j.ajhg.2019.03.018. PMID: 31006513. PMC6506862.
57. Cruz L, Schnur RE, Post EM, Bodagala H, Ahmed R, Smith C, et al. Clinical and genetic complexity of Mitchell-Riley/Martinez-Frias syndrome. J Perinatol. 2014;34(12):948-950. doi:10.1038/jp.2014.162. PMID: 25421130.
58. Barbetti F, Deeb A, Suzuki S. Neonatal diabetes mellitus around the world: Update 2024. J Diabetes Investig. 2024;15(12):1711-1724. doi:10.1111/jdi.14312. PMID: 39344692. PMC11615689.
59. Ahmed S, Lewis V, Russ-Silsby J, Wakeling MN, Thunander M, Vivanco M, et al. Identification of the ACTB p.Ser348Leu de novo variant in individuals with syndromic neonatal diabetes. EBioMedicine. 2026;128:106286. doi:10.1016/j.ebiom.2026.106286. PMID: 42107907. PMC13191097.
60. Philippi A, Heller S, Costa IG, Senée V, Breunig M, Li Z, et al. Mutations and variants of ONECUT1 in diabetes. Nat Med. 2021;27(11):1928-1940. doi:10.1038/s41591-021-01502-7. PMID: 34663987. PMC9356324.
61. Russ-Silsby J, Patel KA, Laver TW, Hawkes G, Johnson MB, Wakeling MN, et al. The Role of ONECUT1 Variants in Monogenic and Type 2 Diabetes Mellitus. Diabetes. 2023;72(11):1729-1734. doi:10.2337/db23-0498. PMID: 37639628. PMC7619148.
62. Merz S, Senée V, Philippi A, Oswald F, Shaigan M, Führer M, et al. A ONECUT1 regulatory, non-coding region in pancreatic development and diabetes. Cell Rep. 2024;43(11):114853. doi:10.1016/j.celrep.2024.114853. PMID: 39427318.
63. Patel KA, Kettunen J, Laakso M, Stančáková A, Laver TW, Colclough K, et al. Heterozygous RFX6 protein truncating variants are associated with MODY with reduced penetrance. Nat Commun. 2017;8(1):888. doi:10.1038/s41467-017-00895-9. PMID: 29026101. PMC5638866.
64. Montaser H, Leppänen S, Vähäkangas E, Bäck N, Grace A, Eurola S, et al. IER3IP1 Mutations Cause Neonatal Diabetes Due to Impaired Proinsulin Trafficking. Diabetes. 2025;74(4):514-527. doi:10.2337/db24-0119. PMID: 39441964. PMC11926274.
65. Giannakopoulos A, Chrysis D. Illness stress-induced transient hyperglycemia in a patient with a novel YIPF5 homozygous missense variant: expanding the phenotype. Hormones (Athens). 2024;23(4):831-834. doi:10.1007/s42000-024-00552-z. PMID: 38632213.
66. OMIM® Online Mendelian Inheritance in Man. *601656 GATA-BINDING PROTEIN 6; GATA6. Baltimore (MD): Johns Hopkins University; last edited 21/10/2025. Available from: https://omim.org/entry/601656. Accessed 31 October 2026.
67. Verstraeten A, Roos-Hesselink J, Loeys B. Bicuspid Aortic Valve. In: Baars HF, van der Smagt JJ, Doevendans PAFM, van Tintelen JP, editors. Clinical Cardiogenetics. 3rd ed. Cham: Springer; 2020. p. 345–361. doi:10.1007/978-3-030-45457-9_20
68. Kishore S, De Franco E, Cardenas-Diaz FL, Letourneau-Freiberg LR, Sanyoura M, Osorio-Quintero C, et al. A Non-Coding Disease Modifier of Pancreatic Agenesis Identified by Genetic Correction in a Patient-Derived iPSC Line. Cell Stem Cell. 2020;27(1):137-146.e6. doi:10.1016/j.stem.2020.05.001. PMID: 32442395. PMC7335348.
69. OMIM® Online Mendelian Inheritance in Man. *600733 PANCREAS/DUODENUM HOMEOBOX PROTEIN 1; PDX1. Baltimore (MD): Johns Hopkins University; last edited 17/06/2025. Available from: https://omim.org/entry/600733. Accessed 31 October 2026.
70. OMIM® Online Mendelian Inheritance in Man. *607194 PANCREAS TRANSCRIPTION FACTOR 1, ALPHA SUBUNIT; PTF1A. Baltimore (MD): Johns Hopkins University; last edited 03/11/2023. Available from: https://omim.org/entry/607194. Accessed 31 October 2026.
71. Aramaki M, Doi H, Kato-Fukui Y, Koga N, Kashimada K, Fukami M. Compound Heterozygosity of PTF1A Exonic and Enhancer Variants in a Japanese Boy With Pancreatic Hypoplasia. Clin Genet. 2026;109(6):1088-1089. doi:10.1111/cge.70134. PMID: 41536248.
72. Freeman PJ, Hart RK, Gretton LJ, Brookes AJ, Dalgleish R. VariantValidator: Accurate validation, mapping, and formatting of sequence variation descriptions. Hum Mutat. 2018;39(1):61-68. doi:10.1002/humu.23348. PMID: 28967166. Tool: https://variantvalidator.org/ (accessed 2026-10-31).
73. OMIM® Online Mendelian Inheritance in Man. *612659 REGULATORY FACTOR X, 6; RFX6. Baltimore (MD): Johns Hopkins University; last edited 21/04/2025. Available from: https://omim.org/entry/612659. Accessed 31 October 2026.
74. Rankin SA, Biesiada J, McCracken KW, Patel RK, Rana M, Luedeke DM, et al. A retinoic acid-Rfx6 cascade modulates Wnt/Bmp-dependent gut tube patterning: insight into the molecular basis of Mitchell-Riley Syndrome. Development. 2026;153(18):dev205552. doi:10.1242/dev.205552. PMID: 42677697. PMC13677009.
75. Ibrahim H, Balboa D, Saarimäki-Vire J, Montaser H, Dyachok O, Lund PE, et al. RFX6 haploinsufficiency predisposes to diabetes through impaired beta cell function. Diabetologia. 2024;67(8):1642-1662. doi:10.1007/s00125-024-06163-y. PMID: 38743124. PMC11343796.
76. Skopkova M, Ciljakova M, Havlicekova Z, Vojtkova J, Valentinova L, Danis D, et al. Two novel RFX6 variants in siblings with Mitchell-Riley syndrome with later diabetes onset and heterotopic gastric mucosa. Eur J Med Genet. 2016;59(9):429-435. doi:10.1016/j.ejmg.2016.08.005. PMID: 27523286.
77. OMIM® Online Mendelian Inheritance in Man. *604882 NEUROGENIN 3; NEUROG3. Baltimore (MD): Johns Hopkins University; last edited 10/03/2025. Available from: https://omim.org/entry/604882. Accessed 31 October 2026.
78. Zhang X, McGrath PS, Salomone J, Rahal M, McCauley HA, Schweitzer J, et al. A Comprehensive Structure-Function Study of Neurogenin3 Disease-Causing Alleles during Human Pancreas and Intestinal Organoid Development. Dev Cell. 2019;50(3):367-380.e7. doi:10.1016/j.devcel.2019.05.017. PMID: 31178402. PMC7082840.
79. OMIM® Online Mendelian Inheritance in Man. *610192 GLIS FAMILY ZINC FINGER PROTEIN 3; GLIS3. Baltimore (MD): Johns Hopkins University; last edited 16/04/2026. Available from: https://omim.org/entry/610192. Accessed 31 October 2026.
80. Meulebrouck S, Scherrer V, Boutry R, Toussaint B, Vaillant E, Dechaume A, et al. Pathogenic monoallelic variants in GLIS3 increase type 2 diabetes risk and identify a subgroup of patients sensitive to sulfonylureas. Diabetologia. 2024;67(2):327-332. doi:10.1007/s00125-023-06035-x. PMID: 38051360. PMC10789827.
81. OMIM® Online Mendelian Inheritance in Man. *609382 IMMEDIATE-EARLY RESPONSE 3-INTERACTING PROTEIN 1; IER3IP1. Baltimore (MD): Johns Hopkins University; last edited 16/04/2021. Available from: https://omim.org/entry/609382. Accessed 31 October 2026.
82. OMIM® Online Mendelian Inheritance in Man. *611483 YIP1 DOMAIN FAMILY, MEMBER 5; YIPF5. Baltimore (MD): Johns Hopkins University; last edited 17/06/2022. Available from: https://omim.org/entry/611483. Accessed 31 October 2026.
83. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 44, Disorders of Cellular Trafficking, 44.1.3 Other Types of Cellular Trafficking.
84. Bruno F, Anitei M, Di Fraia D, Durso W, Dau T, Cirri E, et al. The microcephaly-associated protein YIPF5 differentially regulates ER export. iScience. 2026;29(2):114791. doi:10.1016/j.isci.2026.114791. PMID: 41717013. PMC12915289.
85. OMIM® Online Mendelian Inheritance in Man. *604032 EUKARYOTIC TRANSLATION INITIATION FACTOR 2-ALPHA KINASE 3; EIF2AK3. Baltimore (MD): Johns Hopkins University; last edited 05/10/2017. Available from: https://omim.org/entry/604032. Accessed 31 October 2026.
86. Zegre Amorim M, Houghton JA, Carmo S, Salva I, Pita A, Pereira-da-Silva L, et al. Mitchell-Riley Syndrome: A Novel Mutation in RFX6 Gene. Case Rep Genet. 2015;2015:937201. doi:10.1155/2015/937201. PMID: 26770845. PMC4681791.
87. Nóbrega S, Monteiro MP, Pereira-da-Silva L, Pereira SS, Hartmann B, Holst JJ, et al. Congenital Glucagon-like Peptide-1 Deficiency in the Pathogenesis of Protracted Diarrhea in Mitchell-Riley Syndrome. J Clin Endocrinol Metab. 2021;106(4):e1084-e1090. doi:10.1210/clinem/dgaa916. PMID: 33382423.
88. Mutlu B. Neonatal diabetes due to a truncating RFX6 mutation in an extremely low birth weight preterm infant with Mitchell-Riley syndrome. J Pediatr Endocrinol Metab. 2026;39(9):933-938. doi:10.1515/jpem-2026-0162. PMID: 42332910.
89. Denson AM, Rodriguez KE, Letourneau-Freiberg LR, Kandasamy B, Naylor RN, Abdullatif HD, et al. Children With Pancreatic Hypoplasia Experience Poor Weight Gain and Labile Diabetes but Low Incidence of DKA. J Clin Endocrinol Metab. 2026;111(4):e1164-e1173. doi:10.1210/clinem/dgaf473. PMID: 41006196. PMC13017217.
90. Shalev SA, Tenenbaum-Rakover Y, Horovitz Y, Paz VP, Ye H, Carmody D, et al. Microcephaly, epilepsy, and neonatal diabetes due to compound heterozygous mutations in IER3IP1: insights into the natural history of a rare disorder. Pediatr Diabetes. 2014;15(3):252-256. doi:10.1111/pedi.12086. PMID: 24138066. PMC3994177.
91. Jelti H, Khabbache K, Bouressas S, Hacht F, Oulmaati A, Lamzouri A. Microcephaly, Epilepsy, and Diabetes Syndrome 1: A Moroccan Case Report of Novel Compound Heterozygous IER3IP1 Mutations and Literature Review. Balkan J Med Genet. 2025;28(1):20250006. doi:10.2478/bjmg-2025-0006. PMID: 41200581. PMC12588062.
92. Pinney SE, Oliver-Krasinski J, Ernst L, Hughes N, Patel P, Stoffers DA, et al. Neonatal diabetes and congenital malabsorptive diarrhea attributable to a novel mutation in the human neurogenin-3 gene coding sequence. J Clin Endocrinol Metab. 2011;96(7):1960-1965. doi:10.1210/jc.2011-0029. PMID: 21490072. PMC3135203.
93. Gaibee Z, Warner N, Bugda Gwilt K, Li W, Guan R, Yourshaw M, et al. The Genetic Architecture of Congenital Diarrhea and Enteropathy. N Engl J Med. 2025;392(13):1297-1309. doi:10.1056/nejmoa2405333. PMID: 40174224. PMC11968080.
94. Chao CS, McKnight KD, Cox KL, Chang AL, Kim SK, Feldman BJ. Novel GATA6 mutations in patients with pancreatic agenesis and congenital heart malformations. PLoS One. 2015;10(2):e0118449. doi:10.1371/journal.pone.0118449. PMID: 25706805. PMC4338276.
95. Alquraishi AS, Saad MM, Rayees S. Neonatal Diabetes Mellitus Due to Homozygous ZNF808 Mutation Associated With Skeletal Anomalies: A Novel Presentation of Pancreatic Agenesis-3. Cureus. 2026;18(2):e103126. doi:10.7759/cureus.103126. PMID: 41809312. PMC12968077.
96. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 267–268.
97. Estefanía-Fernández K, Andrés A, Alcolea A, Velayos-López M, Pastrían LG, Ramírez-Amorós C, et al. First multivisceral transplantation in Mitchell-Riley/Martinez-Frias syndrome. Pediatr Transplant. 2022;26(5):e14270. doi:10.1111/petr.14270. PMID: 35307919.
98. Sá B, Girbal I, Espírito-Santo AR, Gil J, Macedo C, Dupont J, et al. Dual genetic diagnosis of Mitchell-Riley syndrome and Temple syndrome in a neonate. Clin Dysmorphol. 2025;34(3):99-101. doi:10.1097/mcd.0000000000000527. PMID: 40445129.
99. Perdas E, Gadzalska K, Hrytsiuk I, Borowiec M, Fendler W, Młynarski W. Case report: Neonatal diabetes mellitus with congenital hypothyroidism as a result of biallelic heterozygous mutations in GLIS3 gene. Pediatr Diabetes. 2022;23(6):668-674. doi:10.1111/pedi.13341. PMID: 35394098.
100. Sarıkaya E, Kendirci M, Demir M, Dündar M. Neonatal Diabetes, Congenital Hypothyroidism, and Congenital Glaucoma Coexistence: A Case of GLIS3 Mutation. J Clin Res Pediatr Endocrinol. 2023;15(4):426-430. doi:10.4274/jcrpe.galenos.2022.2021-12-19. PMID: 35410112. PMC10683540.
101. Couper MR, Brown RM, Gupte G, Perera MTPR, Kelgeri C. Liver Disease in GLIS3 Mutations: Transplant Considerations and Bile Duct Paucity on Explant Histology. J Pediatr Gastroenterol Nutr. 2023;77(1):110-114. doi:10.1097/mpg.0000000000003773. PMID: 36917836.
102. Wejaphikul K, Srilanchakon K, Kamolvisit W, Jantasuwan S, Santawong K, Tongkobpetch S, et al. Novel Variants and Phenotypes in NEUROG3-Associated Syndrome. J Clin Endocrinol Metab. 2022;108(1):52-58. doi:10.1210/clinem/dgac554. PMID: 36149814.
103. Francis KL, Verma A, Pacheco MC, Wendel D, Vue PM, Hu SJ, et al. Neurogenin-3 Enteric Endocrinopathy: A Rare Case of Pediatric Congenital Diarrhea and Diabetes Mellitus. JPGN Rep. 2022;3(1):e173. doi:10.1097/pg9.0000000000000173. PMID: 37168762. PMC10158324.
104. Valenzuela I, Boronat S, Martínez-Sáez E, Clemente M, Sánchez-Montañez Á, Munell F, et al. Microcephaly with simplified gyral pattern, epilepsy and permanent neonatal diabetes syndrome (MEDS). A new patient and review of the literature. Eur J Med Genet. 2017;60(10):517-520. doi:10.1016/j.ejmg.2017.07.007. PMID: 28711742.
105. Söbü E, Kaya Özçora GD, Yılmaz Güleç E, Şahinoğlu B, Tahmiscioğlu Bucak F. A New Variant of the IER3IP1 Gene: The First Case of Microcephaly, Epilepsy, and Diabetes Syndrome 1 from Turkey. J Clin Res Pediatr Endocrinol. 2024;16(3):344-350. doi:10.4274/jcrpe.galenos.2022.2022-8-12. PMID: 36416459. PMC11590771.
106. Zegarra WA, Gallentine WB, Ruzhnikov MR, McAndrews CA, Gloyn AL, Addala A. Safe use of the ketogenic diet in an infant with microcephaly, epilepsy, and diabetes syndrome: a case report. BMC Pediatr. 2023;23(1):453. doi:10.1186/s12887-023-04272-y. PMID: 37689631. PMC10492392.
107. Turck D, Braegger CP, Colombo C, Declercq D, Morton A, Pancheva R, et al. ESPEN-ESPGHAN-ECFS guidelines on nutrition care for infants, children, and adults with cystic fibrosis. Clin Nutr. 2016;35(3):557-577. doi:10.1016/j.clnu.2016.03.004. PMID: 27068495.
108. Cengiz E, Danne T, Ahmad T, Ayyavoo A, Beran D, Ehtisham S, et al. ISPAD Clinical Practice Consensus Guidelines 2022: Insulin treatment in children and adolescents with diabetes. Pediatr Diabetes. 2022;23(8):1277-1296. doi:10.1111/pedi.13442. PMID: 36537533.
109. de Bock M, Codner E, Craig ME, Huynh T, Maahs DM, Mahmud FH, et al. ISPAD Clinical Practice Consensus Guidelines 2022: Glycemic targets and glucose monitoring for children, adolescents, and young people with diabetes. Pediatr Diabetes. 2022;23(8):1270-1276. doi:10.1111/pedi.13455. PMID: 36537523. PMC10107615.
110. European Medicines Agency. Lantus (insulin glargine): EPAR — Product Information (SmPC). Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/lantus
111. Agência Nacional de Vigilância Sanitária (ANVISA). Open data — registered medicines (DADOS_ABERTOS_MEDICAMENTOS.csv). Available from: https://dados.anvisa.gov.br/dados/DADOS_ABERTOS_MEDICAMENTOS.csv
112. European Medicines Agency. Levemir (insulin detemir): EPAR — Product Information (SmPC). Marketing authorisation 01/06/2004. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/levemir
113. European Medicines Agency. Tresiba (insulin degludec): EPAR — Product Information (SmPC). Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/tresiba
114. European Medicines Agency. NovoRapid (insulin aspart): EPAR — Product Information (SmPC). Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/novorapid
115. AbbVie Inc. CREON (pancrelipase) delayed-release capsules, for oral use. Prescribing information. BLA 020725. Revised 2024 Feb. Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=073201aa-556d-4a70-918e-84e9616fd88d
116. INFARMED. Kreon 10 000 e Kreon 25 000 (pancreatina) cápsulas gastrorresistentes — Resumo das Características do Medicamento. Viatris Healthcare, Lda. Infomed. Available from: https://extranet.infarmed.pt/INFOMED-fo/
117. Novo Nordisk. VICTOZA (liraglutide) injection, for subcutaneous use. Prescribing information. NDA 022341. Revised 2023 Feb. Available from: https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=BasicSearch.process (search: Victoza)
118. European Medicines Agency. Victoza (liraglutide): EPAR — Overview and Product Information. Marketing authorisation 30/06/2009. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/victoza
119. ClinicalTrials.gov. Understanding Beta Cell Disorders Through the Study of Rare Genotypes (ENDURE). NCT06478121. University of Exeter. Available from: https://clinicaltrials.gov/study/NCT06478121
120. Hassan SS, Musa SA, De Franco E, Myers R, Van Heugten R, Babiker OO, et al. Characterization of monogenic diabetes among Sudanese children: a multi-center experience from a population with high consanguinity. J Pediatr Endocrinol Metab. 2026;39(1):65-75. doi:10.1515/jpem-2025-0316. PMID: 41275391. PMC12780942.
121. Mohsin F, Azad K, Hassan Z, Islam N, Mahbuba S, Tayyeb S, et al. Clinical and Molecular Characterization of Children with Neonatal Diabetes: Experience from Bangladesh. Mymensingh Med J. 2026;35(4):1090-1098. PMID: 42811839.
122. Yakine F, Bouarab I, Alaoui-Inboui FZ, Ech-Charafi M, Benmohamed H, Jennane F, et al. Monogenic Neonatal Diabetes: Clinical Presentations, Genetic Findings, and Response to Therapy in a Retrospective Case Series. Cureus. 2026;18(1):e102578. doi:10.7759/cureus.102578. PMID: 41769619. PMC12949652.
123. Plengvidhya N, Tangjarusritaratorn T, Teerawattanapong N, Narkdontri T, Innang S, Songlilitchuwong S, et al. Targeted analysis of whole exome sequencing in Thai patients with neonatal diabetes. Hum Genet. 2026;145(1):11. doi:10.1007/s00439-025-02815-0. PMID: 41498801. PMC12779698.
124. University of Exeter. Diabetes Genes — Information for patients and professionals on genetic types of diabetes. Available from: https://www.diabetesgenes.org/
125. University of Chicago. Monogenic Diabetes Registry. Available from: https://monogenicdiabetes.uchicago.edu/
126. APDP — Associação Protectora dos Diabéticos de Portugal. Portal da Diabetes. Available from: https://apdp.pt/
127. Sociedade Brasileira de Diabetes. Available from: https://diabetes.org.br/