Monogenic Neonatal Diabetes (Including Wolcott-Rallison Syndrome)

A class of monogenic disorders in which diabetes is diagnosed in the first six months of life, either permanent or transient, isolated or part of multisystem syndromes; early molecular diagnosis changes treatment in about half of cases.

Literature and regulatory information updated: 26 October 2026.

0. Initial Reference Table

The following table brings together the main entities of the class. The proportions refer to case series from referral centres and are not equivalent to the clinical sensitivity of a test or to population prevalence [1].

In the table, DEND denotes the syndrome of developmental delay, epilepsy and neonatal diabetes, DMR the differentially methylated region, MS-MLPA methylation-specific multiplex ligation-dependent probe amplification, UPD6 uniparental disomy of chromosome 6 and IPEX the X-linked immunodysregulation, polyendocrinopathy and enteropathy syndrome. The sources of the rows are the GeneReviews on PNDM and 6q24-TNDM, the 2022 ISPAD (International Society for Pediatric and Adolescent Diabetes) guideline, the 2024 worldwide review and the original series cited in sections 2 and 3; the OMIM numbers, inheritance and cytobands were checked in the entries for each phenotype and gene [1, 2, 4, 13, 20, 22, 25, 27–55].

1. Introduction

1.1. Historical overview and milestones of molecular discovery

For decades, diabetes diagnosed in the first months of life was treated as an early form of type 1 diabetes, with lifelong insulin and no search for a cause. The first syndromic entity to be delineated was described in 1972 by Wolcott and Rallison, who reported, under the title "Infancy-onset diabetes mellitus and multiple epiphyseal dysplasia", the association of diabetes with onset in infancy and multiple epiphyseal dysplasia, now called Wolcott-Rallison syndrome (WRS) [22, 56]. In 2000, Delépine and colleagues identified biallelic variants in EIF2AK3, the gene for the endoplasmic reticulum (ER) kinase PERK, as the cause of this syndrome [57].

The transient form was the first to have a genetic explanation for isolated diabetes. Duplications of the 6q24 region visible on the karyotype and cases with paternal uniparental disomy of chromosome 6 pointed to a paternally expressed imprinted gene; the series of Temple and colleagues, published in 2000, consolidated the relationship between overexpression of the imprinted 6q24 locus and TNDM and described the pattern of remission and relapse [2, 13, 58]. The minimal region was later refined to the PLAGL1 and HYMAI genes, and in 2008 Mackay and colleagues showed that biallelic variants in ZFP57 cause hypomethylation of multiple imprinted loci in patients with TNDM [13, 59].

The milestone that transformed treatment came in 2004, when Gloyn and colleagues sequenced KCNJ11 in 29 patients with PNDM and found six novel heterozygous missense variants in 10 of them, eight spontaneous and two familial. The patients did not secrete insulin in response to glucose or glucagon, but responded to tolbutamide; four had severe developmental delay and muscle weakness, three of whom had epilepsy, a phenotype that would come to be called DEND syndrome. The most common variant, expressed with SUR1 in Xenopus oocytes, markedly reduced channel block by ATP [60]. In 2006, Pearson and colleagues treated 49 consecutive patients with Kir6.2 variants with sulfonylureas: 44 (90%) stopped insulin, and glycated haemoglobin (HbA1c) fell from 8.1% to 6.4% at 12 weeks, a value maintained at one year [61]. In the same year, Babenko and colleagues described the activating variants of ABCC8 (SUR1) [62].

These were followed by, among others, the genes PTF1A (pancreatic and cerebellar agenesis, 2004), GLIS3 (neonatal diabetes with congenital hypothyroidism, 2006), dominant INS (Støy and colleagues, 2007) and recessive INS (Garin and colleagues, 2010), GATA6 (2011), NKX2-2 and MNX1 (2014), the distal PTF1A enhancer (Weedon and colleagues, 2014) and activating STAT3 variants (Flanagan and colleagues, 2014) [63–70]. In 2015, an international cohort study showed the effect of early comprehensive genomic testing on clinical care in ND, the basis of the current recommendation to test all infants at diagnosis [2, 71]. In 2018, the European Medicines Agency (EMA) authorised the first paediatric formulation of glibenclamide, Amglidia® oral suspension [24]. Between 2018 and early 2024, six new ND genes were described (CNOT1, FICD, ONECUT1, PDIA6, YIPF5 and ZNF808) and three genes of other disorders were recognised as causes of ND (EIF2B1, NARS2 and KCNMA1) [4]. In 2025 and 2026 these were joined by recessive TMEM167A, in a syndrome of microcephaly, epilepsy and diabetes, and by the de novo ACTB variant p.(Ser348Leu) in a syndromic form [28, 29].

1.2. Nomenclature and reclassifications

The 6-month cut-off reflects the fact that autoimmune diabetes is exceptional before that age: only about 4% of ND cases correspond to T1D, and age under 6 months is the only feature considered pathognomonic of monogenic diabetes [2]. Because some patients are only diagnosed between 3 and 6 months, the terms "diabetes mellitus of infancy" and "congenital diabetes" were proposed to replace "neonatal diabetes", without displacing the current term [1]. Massa and colleagues (2005) proposed "permanent diabetes mellitus of infancy" (PDMI), with the gene in the name (for example, KCNJ11-PDMI) to avoid confusion with early autoimmune diabetes, and Colombo and colleagues (2008), faced with patients with INS variants diagnosed well after 6 months, proposed "monogenic diabetes of infancy" (MDI), a term covering transient and permanent forms with onset in the first years [6, 8]. The Oxford Desk Reference recommends referral for genetic testing of all patients diagnosed before 9 months, and ISPAD accepts testing between 6 and 12 months when there are no autoantibodies or other suggestive features are present [2, 3].

ISPAD organises monogenic diabetes into four groups: neonatal or early-infancy diabetes, MODY (maturity-onset diabetes of the young), diabetes with extrapancreatic features and monogenic insulin resistance syndromes. Within ND, Table 1 of the guideline distinguishes three mechanisms: abnormal pancreatic development, abnormal β-cell function and β-cell destruction [2]. The clinical distinction between PNDM and TNDM can only be made over time, because the two forms are difficult to separate at diagnosis; genetics anticipates the course [1]. Orphanet maintains codes for the group (ORPHA:224), for isolated PNDM (ORPHA:99885), for TNDM (ORPHA:99886) and for DEND and intermediate DEND syndromes, and classifies the syndromic forms under their own codes; the code for β-cell agenesis with ND (ORPHA:28455) is obsolete [14–16, 20].

1.3. Definition of the entities within scope

Permanent neonatal diabetes is diabetes diagnosed in the first 6 months that does not remit; the most frequent isolated causes are activating variants of KCNJ11 and ABCC8, INS variants and, in consanguineous families, complete glucokinase deficiency due to biallelic GCK [1, 2]. Transient neonatal diabetes remits within weeks or months and frequently relapses in adolescence or adulthood; about two thirds of cases are due to abnormalities of the 6q24 locus and most of the remainder to activating variants of the KATP channel genes, with smaller contributions from HNF1B, INS and SLC2A2 [2, 13]. 6q24-TNDM is defined by ND with relative hypomethylation of the PLAGL1:alt-TSS-DMR, resulting from paternal UPD6, paternal duplication of 6q24 or hypomethylation of the maternal allele, isolated or in the context of multi-locus imprinting disturbance (MLID) [13]. DEND and intermediate DEND (iDEND) syndromes are the KATP channel forms of ND with developmental delay and early-onset epilepsy (DEND) or with less severe delay and no epilepsy (iDEND) [2]. Wolcott-Rallison syndrome is the AR association of early-onset non-autoimmune diabetes, spondyloepiphyseal dysplasia and recurrent hepatic and/or renal dysfunction, caused by biallelic variants in EIF2AK3 [2, 72]. The remaining syndromic forms are grouped by mechanism: pancreatic agenesis or hypoplasia (PDX1, PTF1A, GATA6, GATA4, CNOT1, ONECUT1, ZNF808), defects of endocrine transcription factors with malformations (RFX6, NEUROG3, NEUROD1, NKX2-2, MNX1, GLIS3, PAX6), ER stress and intracellular trafficking (IER3IP1, YIPF5, TMEM167A, EIF2B1, WFS1, PDIA6, FICD), monogenic autoimmunity (FOXP3, STAT3, LRBA, IL2RA, CTLA4, ITCH) and transporter or metabolic defects (SLC2A2, SLC19A2, NARS2, mitochondrial DNA) [1, 2, 4].

Monogenic diabetes of the MODY type, including the KCNJ11, ABCC8, INS and heterozygous GCK MODY forms, IPEX syndrome and the Tregopathies, and Wolfram syndrome, in which diabetes usually appears after 6 months, belong to the differential diagnosis. These entities are described below only when they cause ND or enter the differential diagnosis.

2. Molecular Genetics and Pathophysiology

2.1. Genetic architecture and classification of types

Insulin secretion by the β cell depends on three conditions: a formed endocrine pancreas, a β cell able to couple glucose metabolism to exocytosis, and survival of that cell after birth. The monogenic causes of ND are distributed across these three stages, which underpins the ISPAD classification by mechanism. More than 30 subtypes had been described by 2022 and the list exceeded 40 genes in 2024 [2, 4]. The cause remains unknown in up to 20% of PNDM, and genetic testing provides a specific diagnosis in more than 80% of children diagnosed before 6 months [1, 2, 71].

The table brings together the types described in the sources consulted, ordered according to the ISPAD classification, with the additions of the 2024 worldwide review and the 2025–2026 publications. The cytobands follow GeneReviews and GeneGuidance where they differ from those of ISPAD, which uses older karyotypic coordinates.

Table sources: ISPAD 2022 (Table 1), GeneReviews on PNDM (Tables 1, 2 and 4), GeneGuidance and the review by Barbetti and colleagues (2024); TMEM167A and ACTB according to the original publications; numbers, inheritance and cytobands checked in OMIM [1, 2, 4, 12, 20, 28, 29, 39, 41, 43, 47, 49, 51, 73–78]. ISPAD also includes AIRE, STAT1 and STAT5B among the monogenic autoimmunity genes with very early-onset diabetes, without data on ND proper [2]. Some forms of severe congenital insulin resistance (INSR, AGPAT2, BSCL2, PIK3R1) also present as ND and are discussed in the differential [4]. SHORT syndrome (PIK3R1) enters the differential diagnosis, together with activated PI3Kδ syndrome (APDS1 and APDS2).

In OMIM numbering, the four PNDM types were assigned in order of description and not of frequency: PNDM1 corresponds to biallelic GCK (606176, AR), PNDM2 to KCNJ11 (618856, AD), PNDM3 to ABCC8 (618857, AD or AR) and PNDM4 to INS (618858, AD or AR); in TNDM, TNDM1 corresponds to the 6q24 locus (601410), TNDM2 to ABCC8 (610374) and TNDM3 to KCNJ11 (610582) [5–11]. TNDM1 is recorded with AD and AR inheritance, because it combines the paternal duplication, transmitted dominantly through the paternal line, and MLID due to biallelic ZFP57 [9, 36]. Several genes in the table have no phenotype entry of their own for ND in OMIM, and the number cited is that of the gene: NKX2-2, biallelic MNX1, biallelic NEUROD1, GATA4, ONECUT1, heterozygous EIF2B1, PDIA6 and FICD. In the entries for these genes, the phenotypes recorded are others (Currarino syndrome for MNX1, MODY6 for NEUROD1, congenital heart defects for GATA4, leukoencephalopathy with vanishing white matter for EIF2B1 and AR spastic paraplegia type 92, OMIM 620911, for FICD) or no human phenotype is recorded (NKX2-2, ONECUT1 and PDIA6; in the PDIA6 entry, only mouse models with diabetes and loss of β-cell identity) [45, 46, 79–84]. By contrast, ZNF808 pancreatic agenesis has its own entry in OMIM (pancreatic agenesis type 3, PAGEN3), and CNOT1 ND is included in holoprosencephaly type 12 with or without pancreatic agenesis (HPE12, AD), in which, in the original series, 3 unrelated patients had the same de novo variant p.Arg535Cys [41, 85]. The cytobands of IER3IP1 (18q21.1) and NEUROG3 (10q22.1) follow OMIM, which differs from the cytoband 10q21.3 given for NEUROG3 in other sources consulted [54, 86].

The relative distribution of causes depends mainly on consanguinity. In non-consanguineous populations, KATP channel and INS variants are the most frequent known causes of PNDM; when the parents are consanguineous, WRS and homozygous GCK are the most common [2]. In the University of Chicago Monogenic Diabetes Registry, the main causes of ND in the United States were KCNJ11 (35%) and INS (16%) [4]. The GeneReviews proportions (Table 1) derive from referral series and the HGMD database, and those for GCK differ between sources: about 5% in GeneReviews and 2–3% of all PNDM according to ISPAD, with higher values in consanguineous regions [1, 2].

2.2. KATP channel: KCNJ11 and ABCC8

The β-cell ATP-sensitive potassium (KATP) channel is a hetero-octamer formed by four Kir6.2 subunits, encoded by KCNJ11, which make up the pore, coupled to four regulatory SUR1 subunits, encoded by ABCC8. The channel couples the energy state of the cell to the membrane potential: glucose enters the β cell, is phosphorylated by glucokinase (hexokinase-4), its metabolism raises the ATP/ADP ratio, the channel closes, the membrane depolarises, voltage-dependent calcium channels open and the rise in cytosolic calcium triggers insulin exocytosis [1, 2]. Activating variants reduce the ability of ATP to inhibit the channel, which remains open at physiological ATP concentrations; the KATP current increases, the β cell becomes hyperpolarised, calcium influx is suppressed and glucose-stimulated insulin secretion disappears [1, 87]. A loss-of-function nonsense variant in ABCC8 has been described that, paradoxically, produces channel gain of function [2].

Sulfonylureas bind to SUR1 and close the channel by an ATP-independent pathway. In KATP PNDM, the sulfonylurea depolarises the hyperpolarised β cell and allows it to respond to incretins, such as glucagon-like peptide-1 (GLP-1), more than to glucose; secretion is therefore stimulated more by oral glucose or a mixed meal than by intravenous glucose, and exogenous glucagon only increases secretion in the presence of the sulfonylurea [61, 88]. In vitro block by tolbutamide mirrored the clinical response of each variant [61].

The KATP channel is also expressed in neurons and muscle, which explains the neurological manifestations [2]. In vascular smooth muscle, the KATP channel is formed by Kir6.1 (KCNJ8) and SUR2 (ABCC9) subunits, and activating variants of these genes cause Cantú syndrome, with congenital hypertrichosis, cardiomegaly and osteochondrodysplasia [89]. The location of the variant partly predicts severity, with exceptions: variants at residues of the ATP-binding site (Arg50, Ile192, Leu164, Arg201, Phe333) or at the interfaces between Kir6.2 subunits (Phe35, Cys42, Glu332) or between Kir6.2 and SUR1 (Gly53) are associated with isolated diabetes, whereas variants at residues distant from the ATP site (Gln52, Gly53, Val59, Cys166, Ile296) are associated with neurological features; GeneReviews includes Gly53 in both groups [1]. Of 24 patients with variants at residue Arg201, all but three had isolated PNDM; p.Val59Met is associated with intermediate DEND; p.Gln52Arg, p.Val59Gly, p.Cys166Phe, p.Ile296Val, p.Gly334Asp, p.Ile167Leu, p.Gly53Asp, p.Cys166Tyr and p.Ile296Leu are associated with DEND syndrome and have not been seen in less severely affected patients [1, 90–92]. The p.Gly334Asp variant, in the ATP-binding site, causes a sulfonylurea-insensitive form of DEND [91]. Some KCNJ11 variants cause TNDM and others PNDM, and p.Val252Ala and p.Arg201His have been described in both forms, because the KATP currents of TNDM and PNDM variants overlap [1].

About 90% of KCNJ11 variants cause PNDM and about 10% TNDM, whereas in ABCC8 about 66% cause TNDM [2, 62, 93]. ABCC8 correlations are less clear-cut: children with a dominant variant may have a parent with the same variant and type 2 diabetes (T2D), reflecting variable severity and age at onset [1, 62]. In 43% of ABCC8 PNDM, inheritance is recessive, from unaffected heterozygous parents [1, 94]. In a cohort of 59 patients with PNDM diagnosed before 6 months and without a KCNJ11 variant, ABCC8 variants were found in 16: 8 heterozygous de novo and 8 homozygous, compound heterozygous or mosaic (Ellard and colleagues, 2007) [7, 31]. Reduced penetrance has been described in ABCC8 and KCNJ11 forms [1, 93]. TNDM variants of the KATP channel may first manifest at around 30 years (5–50), with no neonatal phase, as if the disease began directly with relapse, and these patients may be classified as MODY because inheritance is dominant [93, 95].

2.3. Imprinted 6q24 locus and ZFP57

The 6q24.2 locus contains PLAGL1 (also called ZAC), which encodes a zinc finger protein, and HYMAI, a non-coding transcript without an open reading frame, which share a promoter with a DMR (PLAGL1:alt-TSS-DMR), the imprinting centre of the locus. Normally the maternal allele is methylated and silenced and only the paternal allele is expressed. All 6q24-TNDM mechanisms lead to a double dose of PLAGL1/HYMAI [13, 96]: (1) complete or partial paternal uniparental disomy of chromosome 6, which includes the locus and accounts for about half of sporadic TNDM; (2) unbalanced paternal duplication of 6q24, usually submicroscopic (from 200 kb to several Mb), responsible for most familial cases; (3) hypomethylation of the maternal allele, isolated or in the context of MLID [2, 13]. In the GeneReviews series, the proportions are about 41% for UPD6, 29% for duplication and 30% for imprinting centre defects [13].

MLID may result from biallelic ZFP57 variants in the proband, which account for almost half of TNDM with MLID, from biallelic variants in the mother in a maternal-effect gene such as NLRP7 (described in a patient with a combined phenotype of TNDM and Beckwith-Wiedemann syndrome) or from unknown mechanisms [13, 59]. TNDM is the most frequent phenotype of ZFP57 MLID, which may also present as Silver-Russell syndrome, Beckwith-Wiedemann syndrome or mixed pictures [13]. In the series that identified the gene (7 families with hypomethylation of several imprinted loci), all carriers of ZFP57 variants had, in addition to hypomethylation of the 6q24 DMR, hypomethylation of the PEG3 and GRB10 DMRs, whereas hypomethylation of KCNQ1OT1 and PEG1 occurred with or without a variant; 6 of 9 patients had additional features, mainly developmental delay, and stable mosaicism of the hypomethylation points to a failure of imprint maintenance in the first embryonic divisions [9, 36]. The mechanisms of the hyperglycaemia are not fully understood. Overexpression of PLAGL1 causes cell-cycle arrest and apoptosis, regulates the PACAP1 receptor, which stimulates insulin secretion, and binds to the KCNQ1OT1 CpG island, a negative regulator of CDKN1C, as part of a network of imprinted fetal growth genes. The defect is specific to glucose-stimulated secretion, with the response to glucagon via the stimulatory G protein preserved, and a transgenic mouse carrying the human locus reproduces the altered glucose homeostasis [13]. In adult tissues, PLAGL1 is also expressed biallelically from an alternative non-imprinted promoter located 55 kb upstream [13].

Congenital anomalies are significantly more frequent in paternal UPD6 and MLID than in duplication or isolated hypomethylation; there is no difference in the severity, duration or relapse of diabetes between the aetiological subgroups [13, 58, 97]. UPD6 is mostly isodisomic and can unmask recessive disorders of chromosome 6, the most frequent being HFE haemochromatosis, as well as isolated methylmalonic acidaemia and congenital adrenal hyperplasia due to 21-hydroxylase deficiency [13]. Cytogenetically visible duplications are associated with learning difficulties due to other genes included, whereas submicroscopic ones usually occur with normal intelligence [13]. Orphanet describes a paternal UPD6 syndrome (ORPHA:96191) with macroglossia and minor dysmorphic features [98]. Relative hypomethylation of the 6q24 DMR has been described in one patient with PNDM, with severe IUGR and no ketoacidosis [1, 13].

2.4. Insulin gene (INS)

The A and B chains of insulin, linked by two disulfide bonds, are formed from proinsulin by removal of C-peptide. Dominant heterozygous variants, the most common, disrupt proinsulin folding or disulfide bond formation; misfolded proinsulin is retained in the ER, induces stress and β-cell apoptosis and acts in a dominant-negative manner on biosynthesis from the normal allele [1, 2]. Studies with induced pluripotent stem cells (iPSCs) carrying p.Cys96Arg, p.Cys109Tyr, p.Leu39_Tyr40delinsHis (historical designation in B-chain numbering: LB15YB16delins) and p.Tyr50Cys (YB26C) also showed β-cell dedifferentiation (preclinical studies) [4]. Recessive biallelic variants are loss-of-function (promoter, initiation codon, premature termination codons): they abolish insulin biosynthesis before and after birth, which translates into much lower birth weight and earlier presentation [2, 66]. CRISPR correction of the c.3G>A variant restored proinsulin and normoglycaemia in mice (preclinical) [4].

GeneReviews associates c.3G>A, c.3G>T, c.184C>T, c.*59A>G and a 646 bp deletion with recessive PNDM, and the promoter variants "c.-218A>C", "c.-331C>A" and "c.-331C>G" with PNDM, TNDM and a monogenic childhood-onset diabetes without autoimmunity, with low C-peptide and insulin dependence [1]. On normalisation with VariantValidator against the MANE Select transcript NM_000207.3, c.-218A>C corresponds to NC_000011.10:g.2161368T>G; at the position designated c.-331, the genome reference base is not concordant with the "C>A/G" notation of the literature, which derives from another historical reference, so these variants should be reported in genomic notation. The GeneReviews notation "c.-366_343del" cannot be validated (inconsistent interval), and the 646 bp deletion has no published coordinates in the sources consulted [1, 99]. About 73% of INS PNDM cases are de novo, and about 20% of probands have a family history of dominant ND according to a screen of 1,044 patients [1, 2]. In the same screen, the proportion of probands with an INS variant decreased with age at diagnosis: 33 of 141 (23%) before 6 months, 2 of 86 (2%) between 6 and 12 months and none of 58 between 12 and 24 months; only 1 of 296 probands with MODY had a variant (MODY10, OMIM 613370) [32]. In this series, the median age at diagnosis was 11 weeks, presentation was with DKA in 59% and median birth weight was 2.7 kg (6th centile), whereas in an Italian series 9 of 11 patients had near-normal birth weight, suggesting that loss of β-cell mass due to misfolded proinsulin is mainly postnatal [8].

2.5. Glucokinase (GCK)

Glucokinase is the glucose sensor of the β cell (with similar functions in enteroendocrine cells, hepatocytes and hypothalamic neurons) and controls the rate-limiting step of glucose-stimulated secretion. Pathogenic missense variants increase the glucose concentration at half-saturation and the Km for ATP, reducing phosphorylation potential; splicing variants produce inactive protein [1]. Heterozygosity causes mild, stable fasting hyperglycaemia (GCK-MODY); complete deficiency, through homozygosity or compound heterozygosity, prevents any secretion in response to hyperglycaemia and causes PNDM with severe IUGR, diagnosis in the first days and need for insulin, without extrapancreatic features [2]. The p.Thr168Ala variant (NM_000162.5:c.502A>G; VCV002574157) has been described in a patient with a mild beneficial effect of sulfonylurea [1, 100]. In that 4-year-old homozygous Turkish patient, 3 months of oral glibenclamide increased basal and stimulated insulin secretion about 12-fold, lowered HbA1c from 9.4% to 8.1% and the insulin dose from 0.85 to 0.60 U/kg/day and reduced fasting glucose, with no change in postprandial glucose; the authors attributed the partial response to insufficient ATP production when glucose rises, unlike in KCNJ11 PNDM, and allowed for a better response with less severe variants [5]. In the families described by Njølstad and colleagues, the mothers had had gestational diabetes and then fasting hyperglycaemia or diet-treated diabetes, and the fathers glucose intolerance or mild diabetes, which reinforces the value of the parents' fasting glucose; a European series concluded that complete glucokinase deficiency is a rare cause of PNDM [5, 33]. The ClinGen Monogenic Diabetes Expert Panel has published specifications of the ACMG/AMP (American College of Medical Genetics and Genomics/Association for Molecular Pathology) rules for GCK, version 3.1.0 [1].

2.6. EIF2AK3 and Wolcott-Rallison syndrome

EIF2AK3 encodes PERK, a transmembrane ER kinase that detects misfolded proteins and phosphorylates the initiation factor eIF2α, shutting down general translation and activating the integrated stress response [2, 20]. In the absence of PERK, pancreatic development is essentially normal, but the accumulation of misfolded proteins after birth, when insulin synthesis increases, leads to β-cell apoptosis [2]. ER secretory dysfunction affects other tissues with high secretory activity, which explains the skeletal dysplasia and the hepatic, renal and haematological involvement [20, 72]. There is no specific biochemical marker [20]. In OMIM, the entry for the syndrome (226980) is titled "multiple epiphyseal dysplasia with early-onset diabetes mellitus", and the three siblings in the original description also had bone demineralisation with multiple fractures [12].

In the systematic review by Aldrian and colleagues (189 patients; transcript NM_004836.7), genotype was classified as PERK-1 (no truncating variant, 26 patients), PERK-2 (one truncating, 5) and PERK-3 (two truncating, 151). Overall survival was better with missense variants on one or both alleles than with two loss-of-function variants (p = 0.013), but there was no correlation between genotype and manifestations, age at onset or episodes of liver failure. The hypothesis of a milder phenotype with missense variants in the first kinase domain was not confirmed, except for p.Asn655Lys, with residual activity and onset at 30 months; the previously proposed correlation between neutropenia and specific variants was also not confirmed (p = 0.2434) [22]. Siblings with the same variant may have different manifestations, and the variability has been attributed to residual protein activity [22, 72]. About 3% of pathogenic variants are deletions or duplications detectable only by targeted methods [1].

2.7. Transcription factors of pancreatic and endocrine development

PDX1 (IPF1) is the master regulator of pancreatic development and β-cell differentiation and, in the mature cell, regulates insulin, glucokinase and GLUT2; biallelic variants cause pancreatic agenesis or hypoplasia with PNDM and exocrine insufficiency, milder with hypomorphic variants, and heterozygotes, including the parents, may have MODY4 [1, 20, 101]. PTF1A encodes a subunit of a pancreatic and cerebellar transcription factor: biallelic coding null variants cause pancreatic and cerebellar agenesis (OMIM 609069), whereas recessive variants in a distal enhancer of about 400 bp, located about 25 kb downstream of the gene, cause isolated pancreatic agenesis (OMIM 615935); Weedon and colleagues described 6 enhancer variants in 10 families, and enhancer variants account for more than 60% of PNDM variants in the gene [1, 27, 63, 69]. The enhancer, a conserved region of about 400 bp located within the long non-coding RNA C10orf115, was identified by homozygosity mapping in consanguineous families and genome sequencing; in addition to point variants, 2 siblings of Arab origin were homozygous for a 7.6 kb deletion removing the entire enhancer, and all variants tested abolished enhancer activity in human pancreatic progenitors [102]. The mouse homozygous for the enhancer deletion reproduces the pancreatic hypoplasia and low weight, without cerebellar agenesis [4].

GATA6 causes pancreatic agenesis through haploinsufficiency, with a phenotype ranging from PNDM lethal in the neonatal period to adult-onset diabetes without exocrine insufficiency [67, 103, 104]. ClinGen considers a single entity, "GATA6-related congenital heart disease with or without pancreatic agenesis or ND", with definitive validity [20]. A non-coding 3′ variant in trans, rs12953985, modifies expression: the A/A genotype with a coding variant is associated with pancreatic agenesis and ND with heart disease, and the G/G genotype with adult-onset diabetes, as in the father of a patient in the Exeter series (p.Gly323Ter with G/G and adult-onset diabetes) [4]. GATA4 causes ND or childhood-onset diabetes with pancreatic agenesis and heart disease [20, 105].

RFX6 is essential for formation of the endocrine pancreas, including the β cell: biallelic variants cause Mitchell-Riley syndrome, and also childhood-onset diabetes without the full malformative picture, whereas heterozygous truncating variants cause MODY-type diabetes with reduced penetrance, without malformations, so a heterozygote should not be reported merely as a Mitchell-Riley carrier [1, 4, 20, 106]. NEUROG3 is required for the formation of all enteroendocrine and pancreatic endocrine cells; null variants (p.Glu28Ter, p.Glu123Ter, p.Leu135Pro) and hypomorphic variants (p.Arg93Leu, p.Arg107Ser, p.Ser171fsTer68) always cause malabsorptive diarrhoea, with variable endocrine impact, from ND to diabetes in childhood [4, 20]. Homozygous NEUROD1 causes PNDM with a normal-sized pancreas and neurological anomalies, and heterozygous NEUROD1 low-penetrance MODY6 [20, 107]. NKX2-2 and MNX1 were established in 2014 by homozygosity analysis in 37 consanguineous patients with PNDM, consistent with knockout mice [68]. GLIS3, an activating and repressing zinc finger transcription factor, causes the syndrome of ND with congenital hypothyroidism; about 50% of variants are exonic deletions, and residual transcript expression explains incomplete phenotypes [1, 20, 64]. ONECUT1 (formerly HNF6) causes recessive PNDM with pancreatic hypoplasia; the association of monoallelic variants with diabetes is not supported by the Exeter series, with no enrichment in 484 undiagnosed suspected MODY cases [4, 20, 108]. CNOT1 causes pancreatic agenesis with holoprosencephaly in patients with the same de novo variant p.Arg535Cys [4]. ZNF808, a primate-specific gene, causes aberrant activation of hepatic genes during pancreatic differentiation [4]. In HNF1B, heterozygous loss of function mainly causes developmental kidney disease with diabetes in adolescence or young adulthood; about 50% of cases are de novo and about 50% result from large deletions, and the first two cases of TNDM due to whole-gene deletion were described in 2026 [2, 3, 109].

2.8. Endoplasmic reticulum stress and intracellular trafficking

Several syndromic forms converge on overload of the β-cell ER. IER3IP1, YIPF5 and TMEM167A take part in ER-to-Golgi transport and cause microcephaly, epilepsy and diabetes syndromes (MEDS) [28, 110]. YIPF5 p.Ile98Ser and TMEM167A p.Val59Glu, studied in iPSC-derived β cells, impair proinsulin trafficking and sensitise the cell to ER stress, with apoptosis (functional studies) [4, 28, 111]. TMEM167A encodes a small 72-residue transmembrane protein located in the Golgi apparatus, ubiquitously expressed but with high levels in the developing pancreas and brain, especially in neural progenitors. In iPSC-derived β cells homozygous for p.Val59Glu, differentiation into α, β and δ cells was preserved, but insulin content fell to about half, proinsulin arrival at the Golgi was delayed (colocalisation of 40% versus more than 60% at 10 minutes) and secretion stimulated by glucose and exendin-4 decreased after prolonged maturation; transplanted into immunodeficient mice, they produced almost no human C-peptide. The IER3IP1 protein forms a complex with TMEM167A, which supports the hypothesis of a mechanism common to the three forms of MEDS (preclinical studies) [28]. EIF2B1 encodes a subunit of the eIF2B factor: heterozygous de novo variants at the binding surface for phosphorylated eIF2α cause severe ER stress and PNDM with transient hepatic dysfunction, a feature shared in much more severe form with WRS, whereas biallelic variants cause leukoencephalopathy with vanishing white matter (OMIM 603896), an allelic disorder with different mechanism and inheritance [4, 20, 112]. PDIA6 and FICD, also recessive, were described between 2018 and 2024 in a few probands each (between 1 and 6) [4]. In WFS1, biallelic variants cause Wolfram syndrome, with diabetes rarely neonatal, and some dominant variants cause ND or diabetes of infancy with congenital cataracts and deafness [2].

2.9. Monogenic autoimmunity

In infants under 6 months, autoimmune diabetes is more often due to monogenic defects of immune tolerance than to polygenic T1D: FOXP3, STAT3 and LRBA are at least as likely as T1D at that age [2]. FOXP3 causes X-linked IPEX syndrome [2, 113]. In STAT3 (OMIM gene 102582), heterozygous de novo activating germline variants cause infantile-onset multisystem autoimmune disease type 1 (ADMIO1; OMIM 615952, AD inheritance), including neonatal diabetes, whereas dominant-negative inactivating variants cause autosomal dominant hyper-IgE syndrome with recurrent infections (HIES1; OMIM 147060), a distinct allelic disorder with chronic eczema, recurrent staphylococcal infections, raised serum IgE and eosinophilia [20, 70, 114–116]. Activating variants increase basal transcriptional activity and the response to cytokines, secondarily affect STAT5 and STAT1 phosphorylation, raise SOCS3 expression and reduce regulatory T cells; ADMIO is genetically heterogeneous, with types 2 to 5 caused by ZAP70, CBLB, PDCD1 and CD274 [115]. The study by Saarimäki-Vire and colleagues showed that an activating STAT3 variant induces early activation of NEUROG3 and premature endocrine differentiation, contributing to pancreatic hypoplasia [1, 117]. Table 4 of the GeneReviews on PNDM classifies STAT3 as AR; ISPAD (spontaneous, de novo variant), MedGen/ORDO (AD) and the original article contradict this classification, and no recessive form of STAT3 with ND was found in the sources consulted, so AD inheritance by gain of function is adopted [2, 70, 114]. LRBA (AR) causes LRBA deficiency with enteropathy and autoimmune β-cell destruction; IL2RA (AR) causes CD25 deficiency, with an IPEX-like picture; heterozygous CTLA4 causes autoimmune lymphoproliferation; ITCH (AR) causes multisystem autoimmunity with facial dysmorphism [2, 20, 118].

2.10. Transporters, mitochondrial metabolism and other mechanisms

SLC2A2 encodes GLUT2, the high-Km glucose and galactose transporter of hepatocytes, the proximal tubule and enterocytes. In human β cells the main transporter is GLUT1 and the rate-limiting step is glucokinase, so the contribution of the secretory defect to hyperglycaemia is controversial; the description of ND due to SLC2A2 nevertheless suggests a role for GLUT2 in human secretion [119, 120]. SLC19A2 encodes the high-affinity thiamine transporter THTR1; loss of function reduces intracellular thiamine and impairs ribose synthesis, causing megaloblastosis, whereas in pancreatic endocrine cells SLC19A2 expression is much higher than that of SLC19A3, which compensates only for intestinal absorption [21]. The variants act by reduced expression, ER retention or dysfunction of the protein once inserted in the membrane [21]. NARS2 (11q14.1) encodes mitochondrial asparaginyl-tRNA synthetase; biallelic variants cause ND with deafness and epilepsy, as well as AR deafness (OMIM 618434) and combined oxidative phosphorylation deficiency type 24 (OMIM 616239) [4]. AR deafness type 94 (DFNB94) has been described in a single consanguineous Pakistani family, with profound prelingual sensorineural deafness in 5 members homozygous for p.Val213Phe [121]. Combined oxidative phosphorylation deficiency type 24 ranges from severe infantile forms, with refractory seizures, hypotonia, auditory neuropathy (sometimes detected on newborn hearing screening) and Leigh- or Alpers-type imaging, sometimes with renal or hepatic involvement and death in childhood, to isolated myopathy in adults; lactate may be normal. The OMIM entries for these two disorders do not mention diabetes [122]. The large mitochondrial DNA deletions of Pearson syndrome can cause ND [123, 124]. In KCNMA1, the de novo variant p.Gly375Arg, also described in Liang-Wang syndrome, was found in a patient with diabetes on the third day and severe congenital anomalies [4]. The de novo variant p.(Ser348Leu) of ACTB, the β-actin gene, was identified by trio genome sequencing in 2 of 38 unexplained probands and found in 7 patients from the literature; none of the 96 carriers of other pathogenic ACTB variants recorded in HGMD had diabetes, which points to a variant-specific mechanism [29]. The variant (NM_001101.5:c.1043C>T, in exon 6) is absent from gnomAD v4.1, had already been published as a cause of Baraitser-Winter syndrome type 1 in 7 patients and is classified as pathogenic or likely pathogenic in ClinVar; ACTB sequencing in a replication cohort of 288 probands with unexplained ND found no other carriers. ND and hyperglycaemia were significantly more frequent with p.(Ser348Leu) than with the other variants of the gene (5 of 9 versus 0 of 96), as was gastrointestinal atresia or stenosis (7 of 9 versus 3 of 96; 3 patients with apple-peel intestinal atresia), whereas the thrombocytopenia typical of exon 5 and 6 variants did not occur. Very low birth weight, early diagnosis and exocrine insufficiency in one patient suggest a defect of pancreatic development; structural modelling predicted disruption of actin binding to the acetyltransferase NAA80, a hypothesis still without functional confirmation. With 3 unrelated probands with confirmed ND, the minimum required by ClinGen and Genomics England to consider a gene-disease association is reached, and the authors recommend including ACTB in ND panels [29]. A pig with PAX4 inactivation by base editing developed severe hyperglycaemia in the first 24 hours, with almost complete loss of β and δ cells (preclinical, with no human counterpart described) [125].

2.11. Allelic disorders, genes with two inheritance patterns and penetrance

Several ND genes cause opposite phenotypes or phenotypes of different severity depending on dose and the direction of the functional effect. Activating variants of KCNJ11 and ABCC8 cause ND, MODY and T2D, and inactivating variants cause congenital hyperinsulinism; GCK causes PNDM in biallelic dose, MODY in heterozygosity and hyperinsulinism with activating variants [1, 126]. ABCC8 accounts for 40–45% of non-syndromic genetic hyperinsulinism, with founder variants in Ashkenazi Jews and Finns [20]. Patients cured of diffuse hyperinsulinism due to these genes need lifelong glycaemic surveillance, and subtotal pancreatectomy results in insulin-dependent diabetes in 91% after 14 years [126]. HNF4A is the reverse differential: it increases birth weight (by 800 g on average) and causes neonatal hypoglycaemia in about 10% of carriers [3]. In ABCC8, leucine-sensitive hypoglycaemia of infancy (AD) is a form of hyperinsulinism in which protein meals or leucine cause symptomatic hypoglycaemia; in the family studied molecularly, 5 members heterozygous for p.Arg1353His had protein-induced hypoglycaemia, and the proband responded well to diazoxide, unlike most hyperinsulinism due to recessive ABCC8 variants [127]. In INS, familial hyperproinsulinaemia (AD) results from structural variants of proinsulin or of the cleavage sites that prevent its conversion to insulin, or from mutant insulins with low biological activity; most carriers are asymptomatic or have only mild glucose intolerance, with hyperinsulinaemia and a normal response to exogenous insulin, a phenotype distinct from INS PNDM [128]. In ABCC8, the MODY form has its own entry (MODY12, OMIM 621196): by November 2020 about 2,100 variants of the gene had been published, of which about 1,100 caused ND and 55 MODY12, most described only once, and some recurrent MODY12 variants, such as p.Arg1380His, had previously been described in children with TNDM [31]. In KCNJ11, the MODY form (MODY13, OMIM 616329) corresponds to AD diabetes with onset usually before 25 years, without obesity and responsive to sulfonylureas, and the same p.Cys42Arg variant caused, in a four-generation Japanese family, TNDM, diabetes in childhood, gestational diabetes and adult-onset diabetes [11, 30]. In Currarino syndrome (heterozygous MNX1, AD), the classic triad combines partial sacral agenesis with a sickle-shaped hemisacrum, a presacral mass (teratoma, anterior meningocele) and anorectal malformation, with frequent chronic constipation and perianal sepsis; expressivity is highly variable and up to 33% of carriers are asymptomatic, with reduced penetrance described in several families and prenatal ultrasound diagnosis possible [129]. Biallelic EIF2B1 variants cause leukoencephalopathy with vanishing white matter type 1, a progressive and episodic neurological disorder with cerebellar ataxia and spasticity, aggravated by infections or minor head trauma, with ovarian failure in women; EIF2B1 accounts for about 1% of patients with eIF2B gene disease [130].

Sources: GeneReviews on PNDM (Table 3), GeneReviews on 6q24-TNDM, GeneGuidance and OMIM entries for the allelic disorders and genes [1, 13, 20, 30–33, 42, 44–46, 78, 84, 85, 116, 121, 122, 127–131].

ISPAD and the 2024 review identify genes with two inheritance patterns, both verified in the sources: ABCC8 (dominant or de novo and recessive), INS (dominant proteotoxic heterozygous and biallelic loss of function), WFS1 (recessive and dominant), GCK (heterozygous with MODY and biallelic with PNDM), EIF2B1 (de novo heterozygous with ND and biallelic with leukoencephalopathy), RFX6 (biallelic with Mitchell-Riley and heterozygous truncating with MODY-type diabetes), GATA6 (heterozygous with a modifier in trans) and the 6q24 locus (sporadic UPD6, duplication with dominant paternal transmission, AR ZFP57) [2, 4, 109]. A small fraction of non-obese, non-autoimmune early-onset diabetes without a history of ND has 6q24 abnormalities, and the clue is a history of IUGR or of being small for gestational age at birth [13, 132]. The overall penetrance of ND is high, about 95% according to the Oxford Desk Reference, but it is reduced in ABCC8 and KCNJ11 forms and poorly known for the rare genes; in 6q24-TNDM, siblings with a paternal duplication or biallelic ZFP57 may not have ND [1, 3, 13].

2.12. Variant nomenclature

Variants from the sources were normalised in VariantValidator (GRCh38, MANE Select transcripts) and the c./p. correspondence confirmed in ClinVar. The classifications returned by the ClinVar summary for several well-established DEND variants (p.Gln52Arg, p.Gly334Asp, p.Val59Gly) corresponded to a single "benign" submission associated with TNDM, inconsistent with the functional and clinical literature, and were not used; classification should be consulted in the full record or in the curations of the ClinGen Monogenic Diabetes panel [2, 99, 100].

The MANE Select transcripts of the remaining genes of the class, confirmed in VariantValidator, are: SLC19A2 NM_006996.3, STAT3 NM_139276.3, FOXP3 NM_014009.4, PTF1A NM_178161.3, GATA6 NM_005257.6, PDX1 NM_000209.4, RFX6 NM_173560.4, NEUROD1 NM_002500.5, SLC2A2 NM_000340.2, ZFP57 NM_001109809.5, PLAGL1 NM_001317162.2, GATA4 NM_001308093.3, GLIS3 NM_001042413.2, NEUROG3 NM_020999.4, MNX1 NM_005515.4, IER3IP1 NM_016097.5, YIPF5 NM_030799.9, EIF2B1 NM_001414.4, CNOT1 NM_016284.5, ONECUT1 NM_004498.4, LRBA NM_001364905.1, WFS1 NM_006005.3, HNF1B NM_000458.4, TMEM167A NM_174909.5, ACTB NM_001101.5 and ZNF808 NM_001039886.4 (the 2024 review uses NM_001321425.2, which is not MANE) [4, 99]. RefSeq accessions given without version in secondary databases do not replace these transcripts in the report.

3. Clinical Spectrum and Manifestations by System

3.1. Features common to the class

Fetal insulin is a prenatal growth factor, so many infants with ND are small for gestational age; reduced birth weight reflects the deficit of insulin secretion in utero [1, 2]. In the Oxford Desk Reference, birth weight in ND is typically between 1.8 and 2.5 kg [3]. At diagnosis there is hyperglycaemia, glycosuria, osmotic polyuria, severe dehydration and poor weight gain; insulin deficiency is partial or complete, and insulin therapy corrects the hyperglycaemia with marked catch-up growth [1]. In PNDM, the mean age at diagnosis is 7 weeks (from birth to 26 weeks) [1, 60]. In the series from India published in 2025 (32 patients from Chandigarh, median age at diagnosis 13.2 weeks), DKA was present in 24 [137]; in the Chicago Registry, DKA was present at presentation in 66% of 88 patients with ND, but in none of the 10 with 6q24 [138].

The table summarises the diabetes phenotype and associated features for the genes in GeneReviews and the series cited.

Sources: GeneReviews on PNDM (Table 2) and on 6q24-TNDM, ISPAD 2022, Aldrian 2024, Demirbilek 2020 and Habeb 2018 [1, 2, 13, 22, 25, 27, 88].

3.2. KATP channel neonatal diabetes

KATP ND is the most common cause of PNDM and the second most common of TNDM [2]. It typically presents between 4 and 8 weeks (range 0 to 26), with severe hyperglycaemia and frequent ketoacidosis; C-peptide is low or undetectable [2, 88]. There are no significant differences in IUGR or age at diagnosis between the permanent and transient KATP forms; compared with 6q24-TNDM, IUGR is milder and diagnosis somewhat later, and in KATP TNDM remission is later and relapse earlier [2]. In KATP TNDM, diagnosis occurs at a median of 4 weeks (0–16), remission at 8 months (0.5–48) and relapse at around 15 years (2–45); presentation with ketoacidosis does not preclude remission without treatment [95].

About 20% of children with KCNJ11 variants have neurological features [2]. The most damaging variants cause DEND syndrome (severe developmental delay, early-onset epilepsy and ND), and intermediate DEND, which is more frequent, involves less severe delay and no epilepsy [1, 2]. With p.Val59Met, the variant most often associated with developmental problems, most patients have moderate to severe difficulties, with variability between patients; the spectrum ranges from normal schooling to inability to walk or talk, and epilepsy appears in the first year in DEND and later (3–10 years) in p.Val59Met iDEND, usually as absences that do not require prolonged treatment [133]. Even the so-called mild variants have neurodevelopmental consequences: in studies with siblings as controls, patients with KCNJ11 had deficits in intelligence quotient, academic performance and executive function, and many met criteria for developmental coordination disorder (visuospatial dyspraxia), ADHD, anxiety or autism spectrum disorder (ASD), or had behavioural or sleep difficulties [2, 139, 140]. In the Chicago Registry, 12 patients with KCNJ11 compared with 12 unaffected siblings differed on the Social Responsiveness Scale (SRS-2; p < 0.001), with a score suggestive of ASD in 7 of 10 affected individuals and in none of 11 siblings; anxiety, reported in 7 of 12 affected individuals and 4 of 10 siblings, and perseverance ("grit") did not differ significantly [141]. The median age at assessment was 13.8 years, 7 of the 12 patients had p.Val59Met and the others p.Arg201His (2), p.Arg201Cys, p.Ala174Gly and p.Gly53Asp, and sulfonylurea had been started at a median age of 1.4 years. The mean total SRS-2 score was 66.2 in patients and 45.9 in siblings, with all subscales worse; of the 7 patients scoring in the range associated with ASD, 2 were in the mild range, 3 in the moderate and 2 in the severe (both with p.Val59Met), with no clear relationship between variant and severity, and one patient with p.Arg201His, a variant considered mild, was also in the mild range. One patient with p.Val59Met exposed to sulfonylurea in utero was in the mild range. Anxiety (mean score 25.3 versus 22.7; p = 0.69) was high in both groups, above the lifetime prevalence of 31.9% in US adolescents, so the authors propose assessing siblings as well. The study is small, relies on self-report or parent-report questionnaires, and the SRS-2 is not a diagnostic instrument for ASD [141]. In the French glibenclamide suspension cohort, neuropsychological assessment was abnormal in 13 of 27 patients (48.1%), predominantly dyslexia and dysphasia [142]. A syndrome of infantile epileptic spasms associated with KCNJ11 and ND has been described [143].

Neurological features are less frequent with ABCC8, which more often causes TNDM, but ABCC8 PNDM has a range of difficulties similar to that of KCNJ11; in the international ABCC8 series, 62% had neurological features [2, 4, 144]. In India, among 181 patients with PNDM, 39 (22%) had KCNJ11 (20) or ABCC8 (19) variants; three of the former had DEND and 7 of the 19 with ABCC8 had developmental delay [4].

The course of the disease treated with sulfonylureas differs from the natural history on insulin. In the international long-term KCNJ11 study (81 of 90 eligible patients, median follow-up 10.2 years), 93% remained on sulfonylurea alone, median HbA1c fell from 8.0% before transfer to 5.9% at one year and was 6.4% at 10 years, with no severe hypoglycaemia; there were microvascular complications in 7 of 81 (9%) and no macrovascular ones, and patients with complications had been transferred later (median age 20.5 versus 4.1 years; p = 0.0005) [24, 145]. In the ABCC8 series, 86% remained insulin-independent after about 10 years, with HbA1c going from 7.2% to 6.5% [4, 144]. The German and Austrian registries and other cohorts show mean HbA1c of 5.9–6.0% with minimal hypoglycaemia after 5–10 years, and microvascular complications vary between cohorts (microalbuminuria 4–19%, retinopathy 6%); there is no strong evidence of loss of efficacy over time, but later initiation of sulfonylurea is associated with the need for combined therapy with insulin [1, 146]. In 15 patients with KATP or INS PNDM, there were no severe ocular complications after a median of 24 years of diabetes [2].

3.3. 6q24 transient neonatal diabetes

6q24-TNDM is the most frequent isolated cause of ND and always results in transient diabetes [2]. IUGR may be noted in the third trimester; in the series of 30 infants by Temple and colleagues, mean birth weight was 1,930 g at 39 weeks, a value consistent with the 2024 cohort of McCullough and colleagues [13, 58, 147]. In Docherty's international cohort of 155 patients, the mean age at presentation was 8 days and the mode 1 day [97, 138]. Diabetes usually appears in the first week, may be discovered incidentally in a sick or dehydrated infant and progresses with rapid dehydration; plasma insulin is usually detectable but low for the glucose level, ketonuria is usually absent, islet antibodies are negative and the pancreas is present [13]. About one third have macroglossia and umbilical hernia is rarer [2]. Insulin requirements decrease with weight gain: most no longer need treatment at a median age of 12–14 weeks, with remission close to 100%; the mean duration is 3 months and may exceed one year, with remission up to 18 months [2, 13]. In the Polish series of 22 patients with diabetes up to the first year, 5 (22.7%) had 6q24-TNDM, with median remission at 4 months [148]. Some infants remit without insulin or sulfonylurea, and others respond to oral sulfonylurea [2]. In a series of 97 patients with diabetes before 6 months and remission before 5 years, 70 (72%) had a 6q24 abnormality (paternal UPD6 in 39%, paternal duplication in 40% and methylation defect in 22% of these), 13 had ABCC8 variants, 12 KCNJ11 variants and 3 no identified cause, the latter with atypical phenotypes (prematurity, epilepsy, several relapses and remissions); patients with a 6q24 abnormality had lower birth weight and earlier diagnosis and remission than those with KATP channel variants, with no clinical differences between ABCC8 and KCNJ11 [9]. In a Japanese series of 16 patients with TNDM, the 11 with a 6q24 abnormality had earlier onset, less DKA and more macroglossia at presentation than carriers of KCNJ11 variants [9, 11].

During remission, episodes of hyperglycaemia may occur with intercurrent illness; in a study of 7 children in remission, insulin secretion was low in 4 and normal in 3, and an abnormal response to intravenous glucose suggests future relapse [13]. A small proportion develop clinically significant hypoglycaemia after remission, sometimes with prolonged treatment [2, 149]. Relapse occurs in at least 50–60%, usually around puberty, and has been described from 4 years of age; in a French prospective cohort followed to 18 years (Busiah and colleagues, 2017, according to ISPAD), relapse reached 85% [2]. In Temple's series, the mean age at relapse was 14 years, and in the French series of Metz 5 of 7 patients had diabetes again after 8 years of age [13]. Relapse resembles early T2D with loss of first-phase insulin secretion and is treated with diet, oral antidiabetic agents or insulin [2, 13]. A Japanese patient followed with annual oral glucose tolerance tests (OGTT), from remission at 185 days to relapse at 14.5 years, showed a reduced early insulin response and fluctuations in β-cell function and insulin resistance, and a 10-year-old Japanese girl with relapse and marked insulin resistance improved with metformin [4]. In women, relapse may occur in pregnancy and present as gestational diabetes [13]. Long-term outcomes include lower educational attainment, and patients with diabetes have reduced secretory capacity [2]. Chronic complications have not been studied systematically; one patient with poor adherence and persistent hyperglycaemia from 14 to 28 years developed microangiopathy [13].

Intelligence and growth are usually normal, except in MLID, in which the spectrum includes IUGR, macroglossia, congenital cardiac anomalies, developmental delay, learning difficulties, epilepsy and, rarely, early death, with no correlation between severity and the degree of hypomethylation or the number of loci affected [13]. The recommended assessment in ZFP57 MLID includes screening for pseudohypoparathyroidism (calcium, phosphate and parathyroid hormone) and for digit anomalies [13]. It is prudent to screen for congenital hypothyroidism [13].

3.4. INS and biallelic GCK neonatal diabetes

INS variants are the second most common cause of PNDM after KATP; patients have no extrapancreatic features and are insulin-dependent [2]. Heterozygotes have IUGR similar to that of the KATP forms, but somewhat later presentation, mostly before 6 months, although diabetes may appear up to the first year or later, which justifies testing INS in early-onset diabetes with negative autoantibodies and a MODY-type phenotype [2]. Recessive forms have much lower birth weight and earlier presentation, and may run a course as PNDM or TNDM [1, 2]. In India, 8 of 189 patients with PNDM (4.2%) had INS variants, with mean birth weight of 2,600 g and onset at 4 months, all insulin-dependent [4].

In complete glucokinase deficiency, IUGR is severe, diagnosis occurs in the first days and exogenous insulin is required, without relevant extrapancreatic features [2]. The heterozygous parents have mild, asymptomatic fasting hyperglycaemia, often unrecognised [1, 2].

3.5. Wolcott-Rallison syndrome

In WRS, diabetes is the initial and universal manifestation. In the systematic review and follow-up by Aldrian and colleagues (189 patients: 159 from 62 published studies and 30 unpublished; median follow-up 48 months, maximum 35 years, total 945.75 patient-years), 51.3% were female and the parents were consanguineous in 138. Diabetes appeared at a median of 2.5 months, with genetic diagnosis also at a median of 2.5 months; 8 patients (4%) had onset after 6 months (6.4–30 months), with no correlation between age at onset and genotype. DKA was present at first presentation in 51 of 84 patients with data (60.7%), and median initial HbA1c was 8% in 23 patients. There was no autoimmunity. All but two received insulin from the outset; the remaining two received empirical glibenclamide for suspected KATP ND, without response [22]. In the German and Austrian series of the DPV registry, diabetes was diagnosed at a median of 2.4 months, with frequent DKA [1, 150].

The frequency of manifestations, coded in the HPO ontology, was as follows: hepatic involvement 73% (138/189), growth abnormality 72% (136/189), skeletal anomalies 59.8% (113/189), nervous system 37.6% (71/189), kidney 35.4% (67/189), insufficient haematopoiesis 34.4% (65/189), hypothyroidism 14.8% (28/189) and exocrine pancreatic insufficiency 10.6% (20/189) [22].

Hepatic involvement manifests as episodes of acute liver injury (without coagulopathy) or acute liver failure according to paediatric criteria (PALF), with hepatomegaly and raised transaminases and bilirubin, often triggered by mild intercurrent infections, mainly of the upper airways, which may progress to jaundice, hypoglycaemia, encephalopathy and multiorgan failure or remit spontaneously [20, 22]. In the Aldrian series, there was a median of one episode of acute liver failure and two of acute liver injury per patient, recurrent in 71 patients; the first episode occurred at a median of 10 months, within the first two years [22]. In the bone dysplasia atlas, 11 of 17 patients had Reye-like acute liver disease and 3 of these signs of chronic liver disease; 8 of 17 had acute kidney injury, with chronic renal dysfunction in a further 3 [72]. Unlike the other genetic causes of recurrent paediatric acute liver failure (NBAS, RINT1, SCYL1, LARS1), in which episodes decrease with age, in WRS the hepatic risk persists [151].

The skeletal dysplasia becomes apparent in the first and second years, with short-trunk short stature, gait difficulty, pain and joint contractures [72]. In the Aldrian series, there was dysplasia of long bones, pelvis and vertebrae in 32.3%, short trunk in 27.5%, spondyloepiphyseal dysplasia in 21.7%, osteopenia or osteoporosis in 15.9%, genu valgum and medullary dysplasia in 4.8% each, joint laxity in 2.6% and kyphosis in 2.1%, with rare cases of depressed nasal bridge, craniocervical instability, brachydactyly, pectus carinatum and cleft palate; the bone changes caused spinal cord compression in some patients [22]. Radiology shows platyspondyly, sometimes with cervical instability, generalised epiphyseal dysplasia, shallow acetabula, coxa valga, lateral displacement of the femoral heads and generalised osteopenia, which may be severe and early, with multiple fractures [72]. The figure legends of the atlas describe, in a patient aged 12 years, a disproportionately short and broad trunk with accentuated lumbar lordosis, slightly flattened cervical vertebral bodies with a hypoplastic odontoid process, hypoplastic ilia, small and laterally placed femoral epiphyses with progressive hip subluxation into adulthood, platyspondyly with posterior scalloping, knees with flattened epiphyses and stippled ossification, and hands with small epiphyses and small, irregular carpal bones [72].

In the nervous system, the Aldrian series recorded unspecified developmental delay in 18.5%, isolated neuromotor delay in 10 patients, mild intellectual disability (ID) in 10.6% and severe ID in only 3, seizures in 9.5%, microcephaly in 9 patients and, rarely, ataxia, nystagmus, hydrocephalus, cerebellar atrophy, leukodystrophy and lissencephaly [22]. In the kidney, renal dysfunction in 15.3%, renal failure in 11.6%, renal tubular acidosis in 3.2% and isolated cases of multicystic dysplastic kidneys, enlarged kidneys, ectopic kidney and acute tubular necrosis. In the blood, neutropenia in 22.2%, anaemia in 14.2% and pancytopenia in 2.6%. Hypothyroidism in 14.8%, small-for-gestational-age newborns in 14.8% and exocrine insufficiency in 10.6%; rare associations have been described with Wolff-Parkinson-White syndrome, hypoaldosteronism, glucose-6-phosphate dehydrogenase (G6PD) deficiency, atrial septal defect and double-outlet right ventricle [22]. Recurrent hypoglycaemia, attributed to impaired hepatic gluconeogenesis, complicates glycaemic control [22]. General anaesthesia may precipitate deterioration [72].

3.6. Pancreatic agenesis and hypoplasia

In the series of 30 patients with PTF1A enhancer variants, the largest published, median birth weight was −3.4 standard deviations (SD) and diabetes was diagnosed at a median age of 5 days; all had exocrine insufficiency, postnatal growth was impaired in about 69%, transaminases were raised in 25 of 30 and there was anaemia in 25 of 30, with transiently raised ferritin and cholestasis attributed to malabsorption; only 5 had developmental delay, mild in 4 [4, 20, 27]. All received insulin, and all but 2 had PNDM; unlike the KATP, INS and GCK forms, there was no postnatal catch-up growth [4]. In Weedon's original series, with 14 patients from 10 families, insulin-dependent diabetes was diagnosed at birth or in the first year in most, but in 3 only at 8, 10 and 22 years; all had exocrine insufficiency, pancreatic hypoplasia or agenesis was documented by imaging in the 9 studied, there were no neurological features apart from mild delay in one patient, and another patient died of cholestatic liver failure [102]. Coding PTF1A variants cause, in addition to pancreatic agenesis, cerebellar hypoplasia or agenesis, central respiratory dysfunction, dysmorphism and optic atrophy, with severe IUGR and very little subcutaneous tissue [2, 20]. In the families described (Pakistani, northern European, Turkish and Saudi, all consanguineous), the dysmorphism included an aged triangular face, beaked nose, low-set dysplastic ears and micrognathia, with flexion contractures, bilateral clubfoot, optic nerve hypoplasia or atrophy and meconium ileus; in one patient at autopsy there was no pancreatic tissue, and two infants died at 6 weeks and 4 months [152].

In GATA6, congenital heart disease is the most frequent manifestation (83% of probands in the largest cohort: persistent truncus arteriosus, tetralogy of Fallot, atrial septal defect, atrioventricular septal defect and other conotruncal anomalies), and the pancreas ranges from complete agenesis to normal appearance [20]. In the Exeter cohort, most had pancreatic agenesis with insulin-dependent ND and exocrine insufficiency, 2 had PNDM without need for enzymes, 1 had TNDM and 4 heterozygous parents had diabetes diagnosed between 12 and 46 years; some carriers never develop diabetes [20, 103]. In PDX1, insulin deficiency is more severe than in the ABCC8, GCK or KCNJ11 forms, with low weight and earlier diagnosis [1]. CNOT1 p.Arg535Cys has been described in 4 patients with pancreatic agenesis and holoprosencephaly, and other variants of the gene cause Vissers-Bodmer syndrome (about 45 patients, dysmorphism in 93%, delay or ID in 92%, endocrine manifestations in 56%) [4, 20]. ONECUT1 causes syndromic PNDM with IUGR, pancreatic hypoplasia and gallbladder agenesis or hypoplasia, described in 1 to 6 probands [20, 108]. ZNF808 has been described in 18 patients from 14 families with pancreatic agenesis [4].

Exocrine insufficiency manifests as poor weight gain and loose, foul-smelling stools [1].

3.7. Syndromes with malformations due to transcription factors

In Mitchell-Riley syndrome (RFX6), ND is prominent and is associated with malrotation, duodenal or jejunal atresia, annular or hypoplastic pancreas, gallbladder agenesis and extra- and intrahepatic biliary atresia [1, 153]. In the series that identified the gene (biallelic variants in 5 of 6 probands), the 3 patients with complete loss-of-function variants died in the first months, whereas the 2 with residual function were alive at 4.5 and 9 years; Martinez-Frias syndrome (OMIM 601346), without ND and sometimes with tracheo-oesophageal fistula, shares the other malformations [47, 48]. A preterm infant of 31 weeks weighing 880 g, homozygous for p.Arg525Ter, had duodenal atresia, hyperglycaemia from day 4, chronic diarrhoea, exocrine insufficiency, cholestasis, gallbladder aplasia and pancreatic hypoplasia, and died on day 195 of metabolic instability and progressive hepatic dysfunction [136]. In enteric anendocrinosis (NEUROG3), severe watery diarrhoea from the neonatal period persists when feeds are withheld and requires parenteral nutrition; villi are normal but enteroendocrine cells are absent; most develop insulin-dependent diabetes in the neonatal period or early childhood, with short stature and delayed puberty, and pituitary hypoplasia with combined hormone deficiency, hypogonadotropic hypogonadism, subclinical hypothyroidism with a hypoplastic thyroid and proximal tubulopathy have been described [20, 86]. In the 3 boys of the original description, the diarrhoea stopped on fasting and recurred with oral glucose or amino acid solutions, exocrine function was normal and the 2 survivors only developed diabetes at 8 years, whereas a compound heterozygous Chilean child (p.Glu28Ter and p.Leu135Pro) had PNDM and malabsorptive diarrhoea with normal exocrine function [86]. In biallelic NEUROD1, PNDM appears before 2 months with a normal pancreas and no exocrine insufficiency, associated with severe cerebellar hypoplasia, learning difficulties, profound sensorineural deafness and retinal dystrophy [20, 107]. In NKX2-2, in addition to delay, hypotonia, short stature, deafness and constipation, the fourth infant described, homozygous for c.356del, had very low weight, severe diabetes with undetectable C-peptide and severe obesity at one year due to paradoxical glucose-induced ghrelin secretion, controlled with three daily doses of a long-acting analogue [20, 135]. In MNX1, biallelic PNDM (about 1% of PNDM) is associated with delay, sacral agenesis and imperforate anus, whereas heterozygotes have Currarino syndrome (sickle-shaped sacrum, anorectal malformation and presacral mass, with tethered cord or conus lipoma in most of a series of 29) [1, 20].

In the syndrome of ND with congenital hypothyroidism (GLIS3), newborns are small for gestational age and have non-autoimmune insulin-dependent diabetes in the first days or weeks and congenital hypothyroidism, often detected by newborn screening; the complete syndrome includes congenital glaucoma (with high hypermetropia and short axial length), cholestatic liver disease with fibrosis and paucity of bile ducts, sometimes progressing to biliary cirrhosis, and polycystic kidneys, as well as exocrine insufficiency, delay, dysmorphism, sensorineural deafness and osteopenia; about 22 patients from 16 families in 11 countries had been described by 2022 [20, 64]. In biallelic PAX6, ND is associated with microphthalmia, microcephaly and brain malformations [1, 2]. In Liang-Wang syndrome due to KCNMA1, not all patients have ND, and delay, intestinal and cardiac malformations, bone dysplasia and dysmorphism coexist [2]. In the series that defined the syndrome (8 patients with heterozygous de novo missense variants), all variants abolished or reduced the potassium current of the BK channel, and the 3 carriers of p.Gly375Arg had the most severe picture, with coarse dysmorphism, intestinal atresia, dilatation or malrotation, patent ductus arteriosus, arterial tortuosity and bone changes; all patients have some degree of neurological dysfunction, and about half have cerebral or cerebellar atrophy or a thin corpus callosum on imaging [77]. With ACTB p.(Ser348Leu), the two patients from 2026 had onset at 1 and 8 days, with birth weights of −3.22 and −3.98 SD, deafness and delay in one and intestinal atresia in the other; among the 7 carriers from the literature there was deafness in 3, neurodevelopmental abnormalities in 4 and gastrointestinal atresia in 6, with ND or hyperglycaemia in 5 of the 9 carriers in total [29].

3.8. Microcephaly, epilepsy and diabetes syndromes and other ER stress forms

In IER3IP1 (MEDS1), primary microcephaly with a simplified gyral pattern, severe infantile-onset epileptic encephalopathy with multifocal seizures and refractory infantile spasms, and permanent insulin-dependent diabetes in the first months coexist, with severe delay, cortical blindness, partial agenesis of the corpus callosum and recurrent respiratory infections; about 12 patients had been described by 2025, one of whom died at 8 years [20, 154]. In the original YIPF5 series (MEDS2; 6 patients from 5 consanguineous families, homozygous), median head circumference was −6.2 SD, epilepsy began between the neonatal period and 7 months and insulin-dependent diabetes between 4 weeks and 20 months; one proband died at 1.3 years, and another patient had only brief, reversible severe hyperglycaemia triggered by intercurrent illness [20, 111]. The 6 patients with TMEM167A described in 2025, identified by genome sequencing in 2 probands and then in 4 of 284 patients with diabetes before 6 months, all had ND, diagnosed at a median age of 34 days (1 day to 22 weeks) and treated with insulin from diagnosis (median dose 1.13 U/kg/day; 0.6 to 1.7), and severe microcephaly (median head circumference −8.9 SD; −4.17 to −10.3); 5 had epilepsy diagnosed in the neonatal period, and the sixth died at 59 days without confirmed epilepsy [28]. Three had IUGR, and the 2 patients with MRI available had lissencephaly; 5 of the 6 had died between 2 months and 14 years. The variants, all rare or absent from gnomAD and classified as likely pathogenic, included missense variants in homozygosity or compound heterozygosity (p.Arg52Trp in 2 patients, p.Val59Glu, p.Ile18Lys with p.Cys46Arg), a variant of the last nucleotide of exon 2 causing exon skipping and two frameshift changes with complete loss of function [28]. In EIF2B1, five patients with PNDM due to heterozygous de novo variants had transient hepatic dysfunction [4, 112]. In NARS2, three patients had ND, deafness and epilepsy [4]. In WFS1, recessive Wolfram syndrome causes non-autoimmune insulin-dependent diabetes at a mean age of 6 years, rarely neonatal, followed by optic atrophy about 4 years later; the dominant forms may cause ND or diabetes of infancy with congenital cataracts and deafness [2].

3.9. Autoimmune forms

True autoimmune diabetes accounts for about 4% of ND, and trisomy 21 is much more frequent than expected in ND, with an autoimmune form distinct from common T1D, not associated with the HLA system [2]. IPEX syndrome ranges from severe intrauterine forms to moderate forms and should be considered in male infants with diarrhoea, eczema, autoimmune diabetes, immunodeficiency or severe infection [2]. STAT3 multisystem autoimmune disease includes autoimmune enteropathy or coeliac disease, haematological autoimmunity with cytopenias, short stature, dermatitis and autoimmune diabetes, often neonatal, with hypothyroidism, arthritis, delayed puberty and recurrent infections as variable manifestations; in an international cohort of 191 patients, the median age at onset was 2.3 years, with frequent lymphoproliferation, autoimmune cytopenias and enteropathy [20, 70, 114]. In the original series of 5 patients, 3 had insulin-dependent diabetes diagnosed in the first weeks of life and a fourth at 43 weeks, with autoimmune enteropathy or coeliac disease in the first two years in 3, and later interstitial lung disease, arthritis, hypothyroidism, abnormal dentition and, in one patient, T-cell large granular lymphocytic leukaemia; in a series of 13 patients from 10 families, only 2 had diabetes, almost all had lymphadenopathy with hepatosplenomegaly, autoimmune cytopenias and recurrent infections (including fungal) with hypogammaglobulinaemia predominated, and 2 families showed incomplete penetrance [115]. In a review of 109 patients with LRBA deficiency (93 homozygous and 16 compound heterozygous), there was autoimmunity in 82%, enteropathy in 63%, splenomegaly in 57% and pneumonia in 49%; in a prospective cohort of 22 patients, recurrent infections in 86.4%, immune dysregulation in 81.8% and lymphoproliferation in 72.7% [20, 118]. In CTLA4 haploinsufficiency (AD), the estimated clinical penetrance in 133 patients from 54 families was about 67%, with recurrent infections and hypogammaglobulinaemia, autoimmune cytopenias, enteropathy and lymphocytic infiltration of the lung, brain and gastrointestinal tract [76]. CD25 deficiency (IL2RA) causes an early-onset IPEX-like disease, with autoimmune enteropathy, eczema, infections, CD8+ lymphocyte lymphoproliferation, congenital hypothyroidism and sepsis [1, 20].

3.10. Forms with anaemia, tubulopathy or mitochondrial disease

In TRMA, the triad consists of megaloblastic anaemia with ring sideroblasts and normal vitamin B12 and folate, progressive sensorineural deafness, sometimes congenital and irreversible, and non-type 1 diabetes, with insulin secretion present but defective [21]. Among more than 183 patients from more than 138 families, megaloblastic anaemia was present in 60–70% (anaemia in more than 90%), deafness and diabetes in more than 90%, ophthalmological manifestations such as optic atrophy and retinal dystrophy in 20–30%, cardiovascular manifestations (sudden death, stroke, high-output heart failure, arrhythmias, atrial standstill, atrial and ventricular septal defects) in 20–30%, neurological manifestations (stroke, epilepsy, mood disorder, delay, ID) in 20–40% and thrombocytopenia in 10–30% [21]. In the international Exeter cohort (32 patients from 27 families), the first symptom appeared between birth and 4 years (median 6 months), diabetes was present in all, with onset between 6 weeks and 8 years (median 43 weeks) and before 6 months in 9, with DKA in 3; anaemia, present in all but 2, was the first symptom in 18 [25]. DKA occurred in 15% in another cohort, and most patients initially respond to oral antidiabetic agents but become insulin-dependent [21]. A family described as DIDMOAD (diabetes insipidus, diabetes mellitus, optic atrophy and deafness) had TRMA [21].

In Fanconi-Bickel syndrome, presentation occurs at 3–10 months with hepatomegaly, Fanconi-type nephropathy with severe glycosuria, fasting hypoglycaemia and postprandial glucose and galactose intolerance; some present with ND, others with hypergalactosaemia on newborn screening or early cataracts, and later moon facies, a protuberant abdomen, short stature, hypophosphataemic rickets and delayed growth and puberty appear, with normal mental development [119].

Pearson syndrome, caused by large deletions or rearrangements of mitochondrial DNA, has a frequency of about 1 in 1,000,000 live births and combines macrocytic sideroblastic anaemia, thrombocytopenia, neutropenia, exocrine pancreatic insufficiency, renal tubulopathy and, on the endocrine side, insulin-dependent diabetes and ND; more than 50% die before 4 years, and survivors usually develop Kearns-Sayre syndrome at around 20 years [123, 124]. Among 80 patients with large mitochondrial DNA deletions, about 25% had diabetes, with onset between 0.1 and 57.4 years, without the number with ND being reported [4].

3.11. Natural history and treated disease

In the KATP forms, sulfonylurea controls glycaemia better than insulin and partly attenuates the neurological manifestations; early treatment with glibenclamide modifies glycaemic control and, in part, neurodevelopment, so data from treated cohorts do not describe the untreated course [2, 145]. In 6q24-TNDM, remission is spontaneous and does not depend on the treatment chosen [13]. In WRS, the course is determined by the hepatic episodes, and transplantation alters survival in transplanted patients (section 11) [22]. In TRMA, thiamine corrects the anaemia and improves glycaemic control in some patients, with no demonstrated effect on deafness [25].

4. Clinical Recognition and Red Flags

4.1. When to suspect

Any diabetes diagnosed before 6 months should be considered monogenic until proven otherwise, because age is the only pathognomonic feature and T1D is exceptional at that age [2]. GeneReviews proposes suspecting PNDM in the presence of persistent hyperglycaemia (plasma glucose above 250 mg/dL, about 13.9 mmol/L) in an infant under 6 months lasting more than 7–10 days, with the typical features of diabetes (glycosuria, ketonuria, hyperketonaemia) and insulin and C-peptide low or undetectable relative to glucose [1]. HbA1c is not useful for diagnosis before 6 months, because of the high proportion of fetal haemoglobin [1].

Between 6 and 12 months, the great majority of infants with diabetes have T1D, but ND may be diagnosed at this age; ISPAD recommends considering testing when autoantibodies are negative, there are extrapancreatic features (gastrointestinal anomalies, congenital defects), an unusual family history or several autoimmune diseases at an early age [2]. In WRS, 4% of patients develop diabetes after 6 months, up to 30 months, and ISPAD mentions possible onset at 3–4 years; in TRMA, diabetes often begins after the first year [2, 22, 25]. The Oxford Desk Reference recommendation to test up to 9 months and the ADA recommendation to test anyone diagnosed before 6 months regardless of current age extend the search to adults with diabetes diagnosed in infancy and treated for decades as T1D [3, 23].

4.2. Signs by context

Birth weight and gestational age guide suspicion: marked IUGR is typical, and in 6q24-TNDM the newborn is term or slightly preterm and very small for gestational age [3, 13]. Normal glucose values in the first 24–48 hours do not exclude ND in a small-for-gestational-age newborn, and the absence of DKA does not exclude 6q24 [138]. The table links red flags to the entities that should be sought.

Sources: ISPAD 2022, GeneReviews on PNDM and 6q24-TNDM, GeneGuidance, Aldrian 2024 and the series cited in section 3 [1, 2, 13, 20, 22, 119, 123].

4.3. Urgent signs

Rapid dehydration and DKA are the emergencies at presentation and require immediate rehydration and intravenous insulin [1]. In a newborn with 6q24-TNDM and paternal UPD6, DKA presented at 19 days as respiratory distress interpreted as bronchiolitis, with shock, glucose of 1,260 mg/dL, venous pH of 6.92 and multiorgan failure (section 10) [138]. In WRS, signs of acute liver disease (jaundice, hepatomegaly, hypoglycaemia, altered consciousness) during an intercurrent infection require hospital assessment, because liver failure is the main cause of death and families should be taught to recognise them early [20, 22]. In forms treated with glibenclamide, food refusal or reduced intake predispose to severe hypoglycaemia [142]. In TRMA, one patient had hyperglycaemia during intercurrent illness in a period without thiamine, and another died at 7 years of DKA after a viral illness [25].

4.4. Newborn screening

There is no newborn screening for ND. Some entities of the class are detected by newborn screening for other conditions: the congenital hypothyroidism of the GLIS3 syndrome and the hypergalactosaemia of Fanconi-Bickel syndrome [20, 119]. Diabetes may be discovered incidentally on a glucose measurement in a small-for-gestational-age or sick newborn [13].

5. Diagnosis: Criteria, Tests and Algorithms

5.1. Criteria and recommendations

ND is defined by a diagnosis of diabetes in the first 6 months of life; PNDM is the form that does not remit, and TNDM the form that remits, even if it relapses later [1, 2]. The diagnosis of each entity is established by finding pathogenic or likely pathogenic variants according to ACMG/AMP criteria (clinically synonymous), or, in 6q24-TNDM, by relative hypomethylation of the 6q24 DMR; a variant of uncertain significance (VUS) neither establishes nor excludes the diagnosis [1, 13]. In TRMA, the diagnosis rests on megaloblastic anaemia with normal B12 and folate that responds to oral thiamine, with or without diabetes or deafness, and/or on biallelic SLC19A2 variants [21].

Sources: ISPAD 2022 and ADA Standards of Care in Diabetes—2026, section 2; ADA section 14 (children and adolescents) refers monogenic diabetes to section 2 [2, 23, 155].

5.2. Laboratory and functional assessment

Plasma glucose, insulin and C-peptide confirm the secretory deficit; in 6q24-TNDM insulin is usually detectable but low for the glucose level [1, 13]. Islet autoantibodies support the monogenic autoimmune forms, but all infants under 6 months should be tested genetically whatever the result [2]. Pancreatic imaging is unreliable in the newborn (on ultrasound in one 6q24 case the pancreas was only partially visualised), so pancreatic aplasia is assessed by exocrine function tests, faecal elastase and faecal fat, supplemented by serum fat-soluble vitamins [1, 2, 156]. In HNF1B disease, faecal elastase is always abnormal and absence of the body and tail of the pancreas is highly suggestive [2]. Fasting glucose in both parents should be measured in all children with ND, even without consanguinity or family history, because mild hyperglycaemia in both points to biallelic GCK [2]. When TRMA is suspected, a full blood count and bone marrow aspirate are requested (megaloblasts and ring sideroblasts), and serum thiamine is normal without supplementation [21]. In Fanconi-Bickel syndrome, glycosuria is disproportionate with very reduced or absent tubular glucose reabsorption, and raised serum biotinidase serves as a screen for the hepatic glycogen storage diseases [119]. HbA1c is unreliable in TRMA because of altered red cell survival [25].

5.3. Genetic testing: order and methods

Genetic testing should be carried out at diagnosis, without waiting for remission or the appearance of other features; the major laboratories offer comprehensive testing of all subtypes and very rapid testing of the subtypes that change treatment [2]. Panel NGS has replaced gene-by-gene Sanger sequencing and is preferable to phenotype-directed testing, especially in ND; in low-resource settings, Sanger sequencing of the genes most relevant to treatment may be the practical option, and Sanger is appropriate for cascade testing of a known variant [2]. GeneReviews describes two options: a multigene ND panel including the genes of the class and those of the differential, with methods that may include sequencing, deletion/duplication analysis and non-sequencing tests such as methylation analysis, or the exome (more widely used) or the genome [1]. ISPAD warns that some panels include genes without robust evidence, which may cause diagnostic errors [2]. The 2024 review recommends routinely including INSR, AGPAT2, BSCL2 and PIK3R1, and investigating mitochondrial DNA in ND undiagnosed by panels [4]. In the 2026 study that identified ACTB p.(Ser348Leu), the genome was sequenced in 38 unexplained probands and their parents, with replication in 288 unsolved cases [29].

In 6q24-TNDM, no single test resolves all mechanisms. Methylation analysis is the central test, because it detects hypomethylation whatever the mechanism; MS-MLPA, used in practice, assesses methylation and copy number simultaneously, and the methylation component allows the parental origin of a duplication (paternal and pathogenic, or maternal) to be inferred from the proband's sample alone. Loss of maternal methylation with normal copy number corresponds to paternal UPD6 or an isolated epigenetic defect, which MS-MLPA does not distinguish; the distinction is made by parent-of-origin analysis with STR or SNP markers in the proband and both parents [13, 148]. In isolated PLAGL1 hypomethylation, MLID should be assessed, with ZFP57 sequencing in the proband and of maternal-effect genes such as NLRP7 in the mother [13]. Ratiometric methylation-specific PCR of the 6q24 DMR, applied to 45 TNDM cases (12 with paternal UPD6, 11 with an isolated methylation defect, 16 with duplication and 6 of unknown cause) and 29 controls, correctly classified all samples [9].

Source: GeneReviews on 6q24-TNDM (Table 2) [13, 97]. In the 2025 Polish series, MS-MLPA identified paternal UPD6 or isolated hypomethylation in 3 patients and paternal duplication in 2, but the lack of trio samples prevented UPD6 from being separated from epimutation, and the authors recommend including methylation and trio analysis by SNP or STR early in the algorithm [148].

5.4. Limitations of the tests

The proportions of variants detectable by each method (GeneReviews, Table 1) are not equivalent to the clinical sensitivity of the test. In KCNJ11, sequencing detects all known variants, which are activating, and no deletions are expected; in EIF2AK3, about 3% are deletions or duplications; in GATA6, about 10%; in GLIS3, about 50%; in INS, less than 5%; in PTF1A, less than 10%, but more than 60% of PNDM variants lie in the distal enhancer, outside the coding regions captured by many exomes [1, 20]. Detection of deletions and duplications requires targeted methods (quantitative PCR, long-range PCR, MLPA, targeted microarray), and the exome and genome detect them with lower sensitivity [1]. Large deletions or duplications of PNDM genes have only been described in patients with additional phenotypes, not in isolated PNDM [1]. In SLC2A2, a 26 bp insertion and a complex rearrangement are detectable by sequencing [1]. INS promoters and non-coding regions need explicit coverage, and the notation of promoter variants should be genomic (section 2.4) [99]. Deep intronic variants escape the exome and panels centred on coding regions: in an infant with PNDM, the child of first cousins from Southeast Asia and negative on exome and on Sanger sequencing of 36 monogenic diabetes genes, the cause was the INS variant NM_000207.3:c.187+241G>A (NC_000011.10:g.2160544C>T) in homozygosity, found on reviewing low-coverage intronic regions; the variant creates a splicing donor site, no normal INS transcript was detected and several heterozygotes in the family had gestational diabetes or insulin-treated adult-onset diabetes [8, 32, 99].

Gonadal mosaicism has been described in KCNJ11 and is not detected in the parents' leucocyte DNA [1, 157]. The cause remains unknown in less than 20% of PNDM, so a negative result does not exclude a monogenic cause, especially if the panel does not include the recent genes, the PTF1A enhancer, mitochondrial DNA and 6q24 methylation [1, 4, 71]. Unvalidated targeted tests may lead to therapeutic errors: in a 2026 Indian series, three of four undiagnosed infants received glibenclamide on the basis of an amplification refractory mutation system PCR (ARMS-PCR) suggestive of ABCC8, later reclassified as benign on exome sequencing, and the sulfonylurea was stopped [158].

5.5. Interpretation and content of the report

The report should state the method, the limitations of the test, the gene, the transcript with version (MANE Select), the variant in HGVS, the zygosity, the classification (pathogenic, likely pathogenic or VUS, with the evidence) and the probability of transmission [1, 2]. Classification should follow ACMG/AMP standards with the specifications of the international ClinGen Monogenic Diabetes Variant Curation Expert Panel, which has curated hundreds of variants, freely accessible and recognised by the FDA (Food and Drug Administration) [2]. A VUS, or predictive testing of asymptomatic relatives, justifies consultation at a specialist centre; family segregation and parental studies help to classify de novo or recessive variants [2]. For 6q24, the report should distinguish isolated hypomethylation, UPD6, duplication (with parental origin) and MLID, because recurrence risk depends on the mechanism (section 8) [13]. For heterozygous truncating RFX6 variants, the report should mention the risk of MODY-type diabetes and not merely carrier status [20].

5.6. Practical algorithm

5.7. Prenatal and preimplantation diagnosis

Prenatal diagnosis (PND) and preimplantation genetic testing (PGT) are possible once the variant has been identified in the family [1]. The Oxford Desk Reference notes that PND is requested mainly for the severe syndromic forms, such as WRS and IPEX syndrome, which may be fatal in childhood [3]. For 6q24, PND and PGT are possible for biallelic ZFP57 or paternal duplication, but the phenotype of ZFP57 MLID cannot be predicted by PND; in UPD6 associated with a structural anomaly, the risk depends on the rearrangement [13]. A fetus affected by 6q24 identified at PND should be expected to show growth restriction in the third trimester [13]. In PTF1A pancreatic and cerebellar agenesis, absence of the cerebellum on fetal ultrasound or MRI, with severe IUGR, allowed prenatal diagnosis in an already affected family, before the gene was identified [152].

6. Differential Diagnosis

The differential of ND has two levels: distinguishing monogenic diabetes from other causes of hyperglycaemia in the infant, and, within the class, distinguishing the entities from one another, because treatment depends on the subtype. The table compares the entities and conditions most often discussed.

Table sources: ISPAD 2022, GeneReviews on PNDM, 6q24-TNDM and TRMA, the 2024 review, the bone dysplasia atlas, the Oxford Desk Reference and Saudubray [1–4, 13, 20, 21, 72, 119, 123, 151, 153, 159].

6.1. Autoimmune diabetes and Tregopathies

The small fraction of ND with autoimmune T1D (about 4%) does not remove the need for genetic testing, because in infants under 6 months the immune tolerance genes are at least as likely as T1D [2]. ISPAD lists at least nine genes of autoimmune syndromes with neonatal or infancy-onset diabetes with autoantibodies (AIRE, CTLA4, FOXP3, IL2RA, ITCH, LRBA, STAT1, STAT3, STAT5B). IPEX syndrome should be considered in a male infant with diarrhoea, eczema and autoimmune diabetes, and treatment is based on immunosuppression or haematopoietic stem cell transplantation [2]. The Tregopathies present with autoimmune diabetes and autoantibodies, whereas in 6q24-TNDM islet antibodies are absent and in WRS there is no autoimmunity [2, 13, 22].

6.2. Transient versus permanent

At neonatal diagnosis it is difficult to distinguish TNDM from PNDM [1]. 6q24-TNDM has more severe IUGR, earlier presentation, absence of ketosis and faster remission; in KATP TNDM, remission is later and relapse earlier, and DKA may occur [2, 95]. Because some KCNJ11 variants cause both forms, genotype does not always predict the course, and KATP channel TNDM variants may first manifest in adulthood as dominantly inherited diabetes [1, 95].

6.3. Wolcott-Rallison syndrome and other disorders with hepatic or bone involvement

WRS should be considered in every infant with apparently isolated PNDM, especially with consanguinity, because the diabetes may precede the skeletal and hepatic changes [1, 2]. The bone dysplasia atlas considers it the first-line diagnosis in a child with early-onset diabetes and consanguineous parents, even before the bone changes; multiple epiphyseal dysplasia is distinguished by less vertebral involvement and the absence of early-onset diabetes [72]. The absence of an EIF2AK3 variant in a child with early-onset diabetes and epiphyseal dysplasia without the other complications suggests genetic heterogeneity [72]. In recurrent acute liver failure with fever, the panel should include NBAS, RINT1, SCYL1, LARS1 and EIF2AK3 [151]. In EIF2B1, the hepatic dysfunction is transient and less severe, and there is no skeletal dysplasia [4].

6.4. Diabetes with deafness, anaemia or optic atrophy

TRMA is distinguished from Wolfram syndrome (WFS1), Wolfram syndrome type 2 (CISD2, with high-frequency deafness only and gastrointestinal ulcers) and mitochondrial disorders with diabetes and deafness by macrocytic anaemia with megaloblastic marrow and by the response to thiamine; megaloblasts with ring sideroblasts distinguish it from metabolic and nutritional anaemias, and the acquired differential is myelodysplastic syndrome [21]. In Wolfram syndrome, diabetes with progressive optic atrophy before 16 years defines the diagnosis; diabetes appears at a mean age of 6 years and visual loss about 4 years later, and may be confused with retinopathy [2]. Pearson syndrome combines ND, sideroblastic anaemia, cytopenias and exocrine pancreatic insufficiency [123, 124]. In mitochondrial diabetes due to m.3243A>G, rare in childhood (less than 1%), about 20% present acutely, including with ketoacidosis, and transmission is maternal [2].

6.5. Inborn errors of metabolism and insulin resistance

Ketolysis defects can mimic ND, through ketoacidosis with hyperglycaemia [159]. Neonatal hyperglycaemia is a sign of metabolic presentation described in Saudubray, which associates IER3IP1, PTF1A and WFS1 with diabetes with microcephaly and brain malformations [159]. In Italy, among 104 patients with ND in 2024, 7 had severe congenital insulin resistance (6 with biallelic INSR and 1 with BSCL2); in Sudan, 2 of 37 had generalised lipodystrophy due to AGPAT2; and a patient with SHORT syndrome due to PIK3R1 p.Arg649Trp had diabetes at 5 days with remission at 6 weeks [4].

6.6. Differential with later-onset diabetes

After 6 months, monogenic diabetes enters the differential of T1D through the ISPAD criteria (family history in a parent, absence of autoantibodies, detectable C-peptide during prolonged remission) and includes MODY, HNF1B diabetes, Wolfram syndrome, mitochondrial diabetes, the exocrine forms (CEL, CFTR, PRSS1, SPINK1) and the TRMT10A and DNAJC3 syndromes [2]. HNF1B disease does not respond well to sulfonylureas, because of hepatic insulin resistance, and requires early insulin [2].

7. Treatments and Risk Management Strategies

7.1. General principles and initial assessment

GeneReviews notes that there are no clinical practice guidelines devoted solely to PNDM and refers to the 2022 ISPAD guideline, which is the reference for testing, choice of sulfonylurea and its dose; the Exeter laboratory's protocols for transfer to glibenclamide, cited by GeneReviews and ISPAD as a reference resource, serve as the operational algorithm [1, 2, 88]. Except for the KATP channel forms and some SLC19A2 forms, all causes of ND are treated with subcutaneous insulin [2]. After diagnosis, GeneReviews recommends follow-up by paediatric endocrinology, renal ultrasound in HNF1B forms, developmental assessment (with neurology and electroencephalogram if seizures are suspected) in KCNJ11, MNX1, NEUROD1 and NKX2-2 forms, pancreatic imaging and exocrine assessment (faecal elastase, serum fat-soluble vitamins) in HNF1B, PDX1, PTF1A and RFX6 forms, genetic counselling and family support, including social services and home nursing [1]. In 6q24-TNDM, the initial assessment includes weight, length and head circumference, glucose and C-peptide, assessment of macroglossia and its impact on feeding, of umbilical hernia and of speech and, in 6q duplication or ZFP57 MLID, developmental assessment, neurological examination, electroencephalogram if seizures are suspected, brain MRI if there is delay or hypotonia, echocardiogram, audiology, ophthalmology, renal ultrasound and measurement of calcium, phosphate and parathyroid hormone [13].

7.2. Acute treatment and insulin therapy in the infant

At presentation, especially with ketoacidosis, immediate rehydration and continuous intravenous insulin infusion constitute the initial treatment; subcutaneous insulin begins when the infant is stable and tolerating oral feeds [1]. In 6q24-TNDM, intravenous insulin is switched to subcutaneous as early as possible, often within two weeks, frequently diluted or by pump; some infants need only rehydration, and insulin requirements are no higher than in the common forms of diabetes [13]. Ketoacidosis contraindicates glibenclamide, and its reference treatment is continuous intravenous insulin with normal saline [24]. In infants who are small for gestational age or have exocrine insufficiency, caloric intake should be high [1, 13].

Data on insulin preparations in infants are scarce. Long-acting peakless insulins (glargine, detemir) may work better in small infants, and some centres prefer continuous subcutaneous infusion by pump, as safer, more physiological and more precise. Rapid-acting analogues (lispro, aspart) and regular insulin should be avoided as boluses, except in continuous intravenous or subcutaneous infusion, because of the risk of severe hypoglycaemia; NPH insulin has a shorter duration in the infant. With very low requirements, insulin diluted to 5 or 10 U/mL is used, with extreme caution to avoid dosing errors. The target proposed by GeneReviews is glucose of 70–180 mg/dL for more than 70% of the day and HbA1c below 7.0%, weighed against the risk of prolonged hypoglycaemia, especially before 6 years, with individualisation in children who do not express symptoms of hypoglycaemia [1]. Insulin doses are titrated to glucose and have no guideline-fixed value; as an observational reference, patients in the French cohort received a median of 0.6 IU/kg/day (interquartile range 0.34 to 0.8) before switching to glibenclamide, and in KATP channel TNDM good control with less than 0.5 U/kg/day indicates significant endogenous secretion [95, 142].

The insulin pump allows the precise dosing that small infants need; sensor-augmented systems with predictive suspension of delivery for hypoglycaemia have been described in three patients, and hybrid closed-loop systems with diluted insulin in TNDM with very low daily requirements, without significant hypoglycaemia (case reports) [4]. In 2026 the implementation of a hybrid closed-loop system at 29 days of life was published (section 10), and a Portuguese team from the Unidade Local de Saúde de Santa Maria, in Lisbon, described the use of continuous glucose monitoring and a sensor-integrated pump in early-onset TNDM [156, 160]. In a resource-limited setting, basal NPH insulin alone, without boluses, controlled a newborn with IUGR (1.5 kg; glucose 640 mg/dL) within 7 days and allowed exclusive breastfeeding, whereas severe malnutrition with metabolic acidosis delayed control to 40 days, against 5 to 7 in the others [158].

No insulin is authorised for newborns, so its use in ND before 1 year is off-label; dilution has label support only for insulin lispro in the United States.

Sources: EMA SmPCs and FDA labels; the clinical scheme follows GeneReviews and ISPAD [1, 2, 161, 162].

7.3. Sulfonylureas in KCNJ11 and ABCC8 neonatal diabetes

ISPAD recommends treatment with sulfonylurea, especially glibenclamide (also called glyburide), in KCNJ11 and ABCC8 ND (grade B), and notes that glibenclamide significantly improved the neurological and neuropsychological abnormalities of these patients, with greater benefit when started earlier (grade B) [2]. About 90% of patients with activating KATP channel variants switch from insulin to oral sulfonylurea, an off-label use when tablets are used; the Exeter laboratory reports more than 2,500 patients successfully transferred worldwide between 2004 and 2023, with rates of 85 to 90% in KATP channel PNDM, depending on the variant and age, and higher in children than in adults [2, 61, 88]. ADA states that 30 to 50% of people with KATP channel ND improve their glucose with high-dose sulfonylureas, a figure that differs from those of ISPAD, the Amglidia® SmPC and the Pearson trial, all close to 90%; ISPAD is adopted as the reference and the divergence is recorded [2, 23, 24, 61].

Glibenclamide closes the KATP channel by binding to SUR1, by an ATP-independent pathway, and allows the β cell to respond to incretins (section 2.2). Published experience is almost entirely with glibenclamide, chosen because it is non-selective, acting in all tissues that express the channel, and more potent; gliclazide and glimepiride have been used, may be less effective and have no described neurological benefit, and gliclazide binds only to the SUR1 of the pancreas and neurons, whereas glibenclamide also binds to SUR2A in heart and muscle. To convert to another sulfonylurea, Exeter compares the maximum adult dose of that sulfonylurea with 10 mg of glibenclamide [88, 95]. Sulfonylurea crosses the blood-brain barrier, but the concentrations reached in cerebrospinal fluid may limit the neurodevelopmental benefit; in rodents, systemic glibenclamide does not reach therapeutic levels in the brain, and other agents have been proposed for this purpose [2].

The reference dose is that of the guideline: ISPAD states that the glibenclamide dose required, calculated by weight, is higher than that used in adults with T2D, typically about 0.5 mg/kg/day, with values up to 2.3 mg/kg/day occasionally described, and depends mainly on age at onset and the variant; many patients progressively reduce the dose after the transition while maintaining good control [2, 163, 164]. GeneReviews gives "0.4–1.0 mg/kg/day of glibenclamide or 0.5–2.5 mg/kg/day of glyburide", two ranges for the same substance from different series, not two distinct doses [1, 2]. In the long-term series, the median dose fell from 0.30 to 0.23 mg/kg/day in patients with KCNJ11, between the first and tenth years, and from 0.37 to 0.25 mg/kg/day in patients with ABCC8 [4, 144, 145]. At Exeter, the median maintenance dose is about 0.3 mg/kg/day (0.1 to 1.2), against about 0.07 mg/kg/day in adult T2D, and doses up to 2.8 mg/kg/day have been described; in variants with neurological features a minimum of 0.5 mg/kg/day is recommended, with many patients needing 1 mg/kg/day [88, 133]. The half-life is long, levels rise even at a stable dose, the dose is not adjusted to the meal or to pre-prandial glucose, unlike insulin, and with growth the dose increases to maintain the mg/kg/day [88].

Transfer takes place as soon as the variant is identified, ideally in the first 6 months of life, a period associated with a better neurological outcome [2, 133]. The Amglidia® SmPC and the Exeter protocols describe equivalent schemes, with the same starting dose, the same increment and the same pre-dose glucose threshold, applied daily as an inpatient or weekly as an outpatient, and withdraw insulin as glucose falls, starting with the long-acting insulin (box 7.3) [24, 88, 165]. Exeter prefers the outpatient route before 12 months, because the glycaemic response is slow and altered renal clearance prolongs the half-life, which leads the inpatient scheme to raise the dose faster than necessary; when the dose rises slowly, each step reaches its full effect by the end of the week and insulin reduction becomes easier to manage [165]. The target dose and the decision to reduce insulin depend on the transfer rate described for the variant, which is obtained from the diagnostic laboratory or from the reference series [88]. The dated Exeter protocol, of 9 June 2023, includes the appendix on effects on the brain: about 20% of patients with KCNJ11 or ABCC8 ND have developmental delay, and in these a minimum of 0.5 mg/kg/day is recommended, with many patients needing 1 mg/kg/day and some taking up to 2 mg/kg/day; in patients with established difficulties, the improvements are small, in areas such as concentration and speech, and the best outcome is obtained when sulfonylurea is started in the first 6 months of life, so the appendix recommends identifying high-risk variants and treating early with about 1 mg/kg/day [165]. The same document notes that severe hypoglycaemia is extremely rare despite the high doses, with no cases in 700 patient-years of follow-up, that about half of patients with neurological features improve on starting sulfonylurea and that neurological improvement is more marked in those treated as infants; it also reminds the reader that these are clinical guidance tested in a relatively small number of patients, whose application is the responsibility of the treating physician [165].

Box 7.3. Transfer from insulin to glibenclamide in KCNJ11 and ABCC8 PNDM: steps, doses, criteria for reducing insulin, monitoring and formulation, according to ISPAD 2022, the Exeter protocol of 9 June 2023 and the Amglidia® SmPC. Doses are total daily doses, in two doses before meals or feeds, unless otherwise stated; 7 mmol/L corresponds to 126 mg/dL.

In KATP channel TNDM, endogenous secretion is significant and patients are very sensitive to sulfonylurea: before remission, in infants and children under 2 years, Exeter recommends starting with 0.05 mg/kg/day divided into two doses, in a solution prepared with water for injections, with slow increases, noting that hypoglycaemia is rare even at this dose and that, weeks to months later, remission requires reduction and withdrawal; at relapse or late presentation between 10 and 40 years, 0.05 mg/kg/day is used in children and, in adults, glibenclamide 0.625 to 1.25 mg (a quarter to half a tablet) or gliclazide 20 to 40 mg, as in HNF1A diabetes, increasing up to the usual adult dose (no paediatric maximum dose defined in the source, which notes that the doses needed remain below the weight-adjusted adult maximum dose) and then adding basal insulin or other oral antidiabetic agents if the response diminishes [95]. Before the sulfonylurea trial, postprandial C-peptide is confirmed, typically above 200 pmol/L [95].

In 6q24-TNDM, the systematic review of precision treatment identified 7 patients without response to sulfonylurea, who remained on insulin, and 9 who stopped insulin; the French cohort with oral suspension demonstrated safety in newborns, but adequate glycaemic control is essential even if insulin is continued [13, 142, 166]. In biallelic GCK, a mild beneficial effect of sulfonylureas has been described in some patients, not all [1]. In WRS, sulfonylurea does not apply: two patients received empirical glibenclamide for suspected KATP channel ND, without response [20, 22].

Few adverse effects have been described in this population. ISPAD cites only transient diarrhoea and tooth discolouration [2]. In a follow-up of more than 10 years of 81 patients with KCNJ11 PNDM, 11 (14%) had adverse effects and none stopped the drug: 9 with gastrointestinal upset (transient diarrhoea or nausea at the start or on increasing the dose), 1 with initial hepatic steatosis, 1 with tooth discolouration and 2 with mild transient leucopenia; as the doses exceed the recommended adult maximum, effects not yet recognised cannot be excluded [88, 145]. The Exeter page on TNDM reports skin allergy in 1 to 5% and gastrointestinal disturbances in 1 to 2%, with very rare haematological changes [95]. Coeliac disease can cause secondary sulfonylurea failure not explained by poor adherence [2]. On sulfonylurea, insulin secretion depends mainly on food intake via KATP channel-independent pathways, and meals of carbohydrate alone or of protein and fat alone produce similar responses, so including carbohydrate in most meals is recommended to avoid postprandial hypoglycaemia [2]. Occasional unexplained glucose values above 20 mmol/L (360 mg/dL), which then stabilise, have been described; in intercurrent illness without vomiting sulfonylurea is continued, and with vomiting and hyperglycaemia rapid-acting insulin is added [88].

In adolescence there may be an apparent loss of efficacy: a Japanese patient with KCNJ11 ND regained control with a substantial dose increase, rather than reintroduction of insulin, which justifies reassessing the dose at this age [4]. Cases of adding incretin-based agents have been described: sitagliptin allowed insulin to be stopped in a 24-year-old Japanese patient with long-standing KCNJ11 ND, and in a woman with KCNJ11 PNDM in whom sulfonylurea had been started late, sitagliptin allowed insulin to be stopped and semaglutide, 0.25 mg weekly over 5 years of follow-up, kept HbA1c between 6.1 and 6.5% with a 60% reduction in glibenclamide dose and no severe hypoglycaemia, with a transient and reversible rise in HbA1c when semaglutide was interrupted (isolated cases, off-label use) [4, 167]. Exeter suggests considering the addition of sitagliptin when control is not optimal [88].

7.4. Glibenclamide formulations and regulatory status

The paediatric oral suspension Amglidia®, in strengths of 0.6 mg/mL and 6 mg/mL, received orphan designation in the EU on 15 January 2016 and marketing authorisation through the centralised procedure on 24 May 2018, as a hybrid medicine with Daonil® (tablets) as reference; the product information was updated on 6 December 2023 and the risk management plan was published by the EMA on 11 May 2026 [24]. The approved indication is the treatment of neonatal diabetes in newborns, infants and children, with the note that sulfonylureas are effective in patients with variants of the KATP channel genes and in 6q24-related TNDM [24]. The SmPC gives an expected mean maintenance dose of 0.2 to 0.5 mg/kg/day, allows doses up to 2.8 mg/kg/day without described adverse reactions, suggests that three to four daily doses may improve control and requires discontinuation when there is no improvement in insulin therapy, glycaemic control and the neurological picture [24]. Because of the benzoate content (E211), 1 mL/kg/day should not be exceeded, so the 0.6 mg/mL strength is not suitable for doses above 0.6 mg/kg/day; the SmPC includes a table with the maximum weight for each dosage at that strength (up to 10 kg, 0.6 mg/kg/day; 11–12 kg, 0.5; 13–15 kg, 0.4; 16–20 kg, 0.3) and requires prescription in mL stating the strength, with 1 mL (0.05 mL graduation) and 5 mL (0.1 mL) syringes [24]. It is given directly into the mouth with the syringe, 15 minutes before the feed, without shaking and avoiding the feeding tube; shelf life is 3 years and 30 days after opening [24].

The suspension is not bioequivalent to tablets with the same amount of glibenclamide, even crushed: the maximum concentration is about 40% higher than that of the crushed Daonil® 5 mg tablet (201.71 and 206.93 versus 148.34 ng/mL), relative bioavailability is 121.6% for the 0.6 mg/mL strength and 114.1% for the 6 mg/mL strength, and time to maximum concentration is 2.5 versus 3 hours; the conversion rate between micronised tablets and suspension has not been established [24]. The EPAR summary for the public reports a similar level in the study in 18 healthy adults, a discrepancy in which the SmPC prevails. This difference explains why the maintenance doses observed with the suspension are lower than those with tablets.

In the United States, glyburide tablets are indicated only as an adjunct to diet and exercise in adults with T2D, and the label states that safety and efficacy in paediatric patients have not been established; there is no FDA-approved liquid formulation, so treatment of ND is off-label, with crushed tablets or pharmacy-compounded suspensions [168]. The US label dose for adults is 2.5 to 5 mg/day with breakfast at the start (1.25 mg/day in the most sensitive) and 1.25 to 20 mg/day for maintenance, with weekly increases of no more than 2.5 mg; 20 mg/day in a 70 kg adult corresponds to about 0.29 mg/kg/day, well below the doses used in ND, which Exeter describes as two to four times higher than those in T2D [88, 168]. The orphan designation of the suspension in the United States in 2016, mentioned by the French cohort, was not confirmed in the FDA database, and there is no marketing authorisation in that country [142].

In Portugal, a search of INFARMED's Infomed returned as authorised medicines containing glibenclamide alone the generic tablets Glibenclamida Generis 2.5 mg and 5 mg, as well as the combination Glucovance® with metformin; Amglidia® has a valid centralised authorisation in Portugal, but national marketing was not confirmed in this search [24, 169]. The national SmPC for the tablets, approved on 25 September 2023, indicates them for adult T2D, with a usual dose of 5 to 10 mg/day, not recommended above 15 mg/day, and contraindicates them in children and adolescents under 18 years, as well as in pregnancy, breastfeeding, T1D, ketoacidosis, severe renal or hepatic impairment and concomitant use of bosentan; their use in ND in Portugal is therefore not only off-label but contrary to a contraindication of the national label, and the alternative with a paediatric indication is Amglidia® [169]. In Brazil, ANVISA open data show active registrations of glibenclamide alone as tablets (Daonil® and several generics and similars) and no registration of a glibenclamide oral suspension or of Amglidia®; the text of the package leaflets was not consulted, so the Brazilian paediatric indication was not verified and use in ND is with crushed tablets or compounded preparations [170].

Sources: Amglidia® EPAR and SmPC, FDA label for glyburide, INFARMED SmPC and ANVISA open data [24, 168–170].

In the United Kingdom, Exeter describes difficulties in the supply of glibenclamide and the use of importation, with 2.5 and 5 mg tablets of Canadian origin or packs imported from France, and describes the hospital preparation of a 2.5 mg/mL suspension in xanthan gum diluent, with 90-day stability, also stable in water, as well as lower-strength capsules prepared in the pharmacy; the French commercial suspension is described as very expensive, and the published results were obtained mainly with prepared solutions or crushed tablets [88, 171]. The Exeter page on TNDM states that no sulfonylurea is licensed in children, a statement that predates the European authorisation of Amglidia® [95].

The contraindications and precautions in the European SmPC are hypersensitivity to glibenclamide, to other sulfonylureas or to sulfonamides, ketoacidosis, porphyria, use of bosentan and severe renal or hepatic impairment; it should not be used in autoimmune T1D [24]. Acute haemolytic anaemia has been described in G6PD deficiency, so an alternative is strongly recommended and, if that is not possible, monitoring for haemolysis. Hypoglycaemia may be severe and prolonged, requiring glucose for several days, and the dose should be reduced with glucose values below 4 mmol/L (72 mg/dL). Initial diarrhoea on increasing the dose resolves if treatment is continued, nausea does not require reintroduction of insulin and significant vomiting justifies rapid-acting insulin until it stops. Benzoate may worsen neonatal jaundice by displacing bilirubin from albumin [24]. In overdose with hypoglycaemic coma, the SmPC indicates intravenous glucose 0.5 g/kg as a 30% solution, followed by a 10% solution to keep glucose above 100 mg/dL (5.5 mmol/L), with monitoring for at least 48 hours; dialysis is of no benefit [24]. Glibenclamide is potentiated by oral anticoagulants, phenytoin, salicylates, non-steroidal anti-inflammatory drugs, angiotensin-converting enzyme inhibitors, beta-blockers (which also mask counter-regulation), clarithromycin, co-trimoxazole, fluoxetine, monoamine oxidase inhibitors, miconazole, quinolones, insulin and other antidiabetic agents, and attenuated by acetazolamide, sympathomimetics, barbiturates, calcium channel blockers, corticosteroids, diazoxide, diuretics, glucagon, isoniazid, oestrogens, progestogens, phenothiazines, rifampicin and thyroid hormones; it raises ciclosporin concentrations, and colesevelam reduces its absorption, so it should be given at least 4 hours before [24].

The pharmacokinetics explain several practical rules: absorption is rapid, the effect appears within 2.5 hours and lasts up to 15 hours, the half-life is 5 to 10 hours (about 8 hours with the suspension), albumin binding is 99%, metabolism is hepatic and complete via CYP2C9 into three inactive metabolites excreted in bile (60%) and urine (40%), with complete elimination in 45 to 72 hours; hepatic impairment prolongs elimination and renal function has no effect above a creatinine clearance of 30 mL/min [24]. The minimum active concentration is 30 to 50 ng/mL [24].

7.5. Trials and long-term efficacy data

Sources: ClinicalTrials.gov records, section 5.1 of the Amglidia® SmPC, ISPAD and the corresponding publications [2, 24, 61, 142, 144, 145, 172].

These studies are not comparable with one another: they differ in population, formulation, disease duration before transfer and outcome. The founding trial and the long-term studies are not randomised and have no insulin control group, NEOGLI is a single-centre, open-label, non-randomised phase II study with 10 patients, and the French cohort is retrospective, with a small sample, three quarters of European origin, non-standardised titration and variable follow-up, with two authors who declared being consultants in the development of Amglidia® and shareholders in the company [24, 142]. HbA1c values were already close to target before the change of formulation, so the absence of a significant difference in that study does not measure the efficacy of sulfonylurea, but rather clinical equivalence between formulations. The 90% response rates derive from selected patients with KATP channel variants and do not apply to the other causes of ND. No pharmacological intervention trials targeting WRS, 6q24-TNDM, TRMA or pancreatic agenesis were found, nor suspended or withdrawn trials specific to ND; a negative search does not demonstrate non-existence [172].

In the French oral suspension cohort, the median starting dose was 0.16 mg/kg/day (interquartile range 0.05 to 0.30; 0.19 in patients with KCNJ11, 0.10 in those with ABCC8 and 0.08 in the patient with 6q24), with three daily doses in KCNJ11 and two in the others, and the last dose was 0.07 mg/kg/day; the median absolute starting dose was 1.4 mg/day (0.9 to 3.5; minimum 0.03 and maximum 8.85). Thirteen patients (48%) had previously used crushed tablets, at a median of 0.35 mg/kg/day for 3 months, with HbA1c going from 6.75% to 6.36% on switching to the suspension (p = 0.41), but with individual variations of up to 2.3 percentage points. Height increased by 1.28 SD (p = 0.099) and weight by 0.71 SD (p = 0.072). In terms of safety, there were two episodes of severe hypoglycaemia, both during periods of reduced intake (a patient with intellectual disability who refused food and another with anorexia of psychiatric cause in adolescence), and diarrhoea in three, persistent in one, which led to discontinuation at 3 months, with subsequent tolerance of tablets; there was no hypersensitivity, renal or hepatic impairment, tooth discolouration or diabetes complications. The authors calculated 0.027 severe hypoglycaemic episodes per patient-year, a value they compared with 0.19 in a 2002 paediatric T1D series and 0.0559 in a 2025 closed-loop T1D series, an indirect comparison between different populations and eras [142]. The adverse reactions listed in the SmPC, observed with the suspension or with crushed tablets in ND, are very common for neutropenia, hypoglycaemia, transient diarrhoea, abdominal pain, vomiting, dyspepsia, transient rise in transaminases and rash, and common for blurred vision and tooth discolouration; in adults treated with other products, photosensitivity, severe skin reactions, anaphylaxis, cytopenias, agranulocytosis, haemolytic anaemia and bone marrow aplasia have been described [24].

7.6. Treatment of forms not responsive to sulfonylurea

In 6q24-TNDM, insulin is stopped when glucose stabilises, usually at 3 to 6 months, rarely at 12 to 18 months; at relapse, treatment may be dietary alone, with oral antidiabetic agents or with insulin, at a lower dose than in T1D because there is residual secretion, and sulfonylureas or diet suffice in several patients; when sulfonylurea is insufficient, other non-insulin agents are useful, and GLP-1 analogues may be effective [13]. Metformin improved control in a 10-year-old Japanese girl with relapse and marked insulin resistance [4]. Metformin tablets are indicated on the US label as an adjunct to diet and exercise in adults and in paediatric patients from 10 years with T2D, with a starting dose of 500 mg twice daily with meals and weekly increases of 500 mg up to a maximum of 2,000 mg daily in two doses, which provides the reference dosage for an adolescent with relapse of 6q24-TNDM, by analogy with T2D, declared as an extrapolation [173]. Two siblings with 6q24 duplication were treated with sulfonylurea into adulthood, at 40 and 43 years [142].

In WRS, insulin is lifelong from diagnosis, intravenous with rehydration and then subcutaneous; the pump is recommended, especially in small infants and in the presence of recurrent hypoglycaemia attributed to impaired hepatic gluconeogenesis, and glycaemic targets are deliberately less tight, because hypoglycaemia may precipitate acute deterioration [20, 22, 174]. No drug reverses liver failure. Liver transplantation, alone or combined with pancreas or kidney, can be life-saving and improve outcome: in the series of 189 patients, 9 (4.8%) were transplanted, at a median age of 42 months (6 liver only, 1 liver and pancreas, 2 liver, pancreas and kidney), all alive at a median post-transplant follow-up of 24 months and maximum of 14 years, and in patients with acute liver injury or failure survival was better with transplantation (p = 0.0057) [22]. The authors recommend pursuing liver transplantation, especially after previous hepatic crises, consider combined liver and pancreas transplantation an option for difficult diabetes, as it also corrects exocrine insufficiency, and consider the role of the kidney uncertain; in the absence of access to transplantation, they describe as a bridge the supportive measures for acute liver failure, including renal replacement techniques and plasmapheresis, and state that N-acetylcysteine "may be considered", without systematic data [22]. The 2025 systematic review on recurrent paediatric acute liver failure notes that trials of N-acetylcysteine in paediatric non-paracetamol acute liver failure showed no benefit and had worse transplant-free survival at one year [151]. It is also recommended to avoid general anaesthesia whenever possible and to limit drugs and vaccines that are not strictly necessary, because of the risk of secondary hepatic or renal failure, to teach families to recognise a hepatic episode early, to treat fractures and deformities with paediatric orthopaedics and to monitor thyroid, kidney, full blood count and exocrine function [20, 22, 72, 174].

In TRMA, oral thiamine is the specific treatment. GeneReviews recommends lifelong pharmacological doses of 50 to 100 mg/day, and the same source, like the international cohort, notes that there is no additional benefit above 150 mg/day; the metabolic disease handbook gives 25 to 50 mg/day as a dose that usually produces a good response in anaemia, diabetes and arrhythmias [21, 25, 175]. In the Exeter cohort, 15 patients were treated with maximum doses of 25 to 300 mg/day, started between 6 months and 14 years, for 2 to 17.5 years (median 4.7; total 103.5 patient-years), without adverse effects up to 300 mg/day; there was glycaemic benefit in 11 of 15 (7 with better HbA1c and less insulin, 4 off insulin), responders were younger at diagnosis (p = 0.042), at genetic testing (p = 0.01) and at the start of thiamine (median 14 versus 158 months; p = 0.007), and 10 of the 11 started before 3 years; haemoglobin normalised within 6 to 8 weeks in all anaemic patients, platelets normalised, none of those alive needed transfusions again and there was no effect on hearing or on the other extrapancreatic features [25]. Anaemia relapses if thiamine is stopped, and red cells remain macrocytic on treatment [21]. With prolonged treatment, the response may diminish, and previously controlled patients may become dependent on transfusions and insulin [175]. Diabetes is treated by the usual means in addition to thiamine, which may delay its onset and reduce insulin requirements [21]. There is no society guideline with a dose for TRMA, nor an oral thiamine label for this indication: in Portugal there is injectable thiamine and oral derivatives (benfotiamine, sulbutiamine), in Brazil oral thiamine hydrochloride has a specific registration (Benerva® and others) and in the United States oral thiamine is a supplement without a drug label; the reference is therefore cohort evidence, which supports a useful ceiling of about 150 mg/day [21, 25, 170]. In siblings not yet genotyped, GeneReviews recommends supplementation with 25 to 100 mg/day, against the recommended daily intake of 1.5 mg/day in the general population [21]. Deafness may not be prevented by thiamine, and cochlear implantation has been described [21].

In pancreatic agenesis and hypoplasia, ISPAD states that children need pancreatic enzyme supplements, without specifying the dose [2]. There is no published dose for this indication, so the age-based label dose for exocrine pancreatic insufficiency, an analogous condition, is used: the US Creon® (pancrelipase) label recommends, from 0 to 12 months, 3,000 lipase units (one capsule) per 120 mL of formula or per breastfeed; from 12 months to 4 years, starting with 1,000 lipase units/kg per meal; and from 4 years, 500 units/kg per meal, without exceeding 2,500 units/kg per meal, 10,000 units/kg per day or 4,000 units per gram of fat per day; high doses have been associated with fibrosing colonopathy, with colonic stricture in children under 12 years exposed to more than 6,000 units/kg per meal, and the capsule should be opened and the contents mixed with soft acidic food, without crushing [176]. The Portuguese SmPC for Kreon® gives the same age-based doses in cystic fibrosis and 25,000 to 80,000 units per meal in other causes of exocrine insufficiency, without a specific dose below 12 months [177]. The extrapolation, from cystic fibrosis and the age-based label, should be recorded in the clinical notes. Fat-soluble vitamins are monitored [1].

In Fanconi-Bickel syndrome, treatment is symptomatic: water and electrolyte replacement, alkali, phosphate and vitamin D for the tubulopathy, galactose restriction and adequate caloric intake in small, frequent meals with slowly absorbed carbohydrates, which may improve growth [49, 119]. Uncooked cornstarch had a beneficial effect in an early description, but the metabolic disease handbook considers continuous overnight tube feeding superior to uncooked starch; fructose, absorbed by GLUT5 and with preserved metabolism, is an alternative carbohydrate source, and insulin is reserved for diabetes [49, 119].

In the monogenic autoimmune forms, treatment includes immunosuppressants (sirolimus or corticosteroids) or allogeneic haematopoietic stem cell transplantation with reduced-intensity conditioning, recommended in IPEX syndrome: survival is similar with the two approaches, but disease-free survival and quality of life are better with transplantation [2]. In STAT3 gain-of-function disease, almost all patients in the series of 13 needed intensive immunosuppression, 2 received haematopoietic stem cell transplantation and IL-6 blockade improved arthritis in one patient; in a series of 6 patients with severe autoimmunity treated with JAK inhibitors (5 with ruxolitinib and 1 with tofacitinib), all also treated with tocilizumab before, during or after the inhibitor (when given before, alone, it had not been effective), autoimmune cytopenias improved in 4 and haemophagocytic lymphohistiocytosis went into remission, but 2 patients died (one with Pneumocystis jirovecii pneumonia and sepsis, the other with end-stage lung disease), and the authors recommended aciclovir prophylaxis; these are small series without a defined dose for ND, and the decision rests with immunology [115]. In LRBA deficiency, abatacept (Orencia®), a CTLA4-immunoglobulin fusion protein, improved interstitial lung disease and severe enteritis in a series of 9 patients, and in CTLA4 haploinsufficiency the options reviewed include CTLA4 fusion proteins, mTOR inhibitors and haematopoietic stem cell transplantation; here too there is no defined dose for ND, and the scheme follows immunology [75, 76]. In the GLIS3 syndrome, hypothyroidism is treated with levothyroxine, glaucoma with eye drops (timolol, latanoprost, dorzolamide) or surgery, and progressive biliary cirrhosis may require liver transplantation, with a difficult postoperative course because of labile diabetes and renal disease [20]. In forms with congenital hypothyroidism, the 2020–2021 European consensus recommends starting oral levothyroxine once daily at up to 15 µg/kg/day: 10 to 15 µg/kg/day in severe forms (free T4 <5 pmol/L with raised TSH) and about 10 µg/kg/day in mild forms, with 5 to 10 µg/kg/day to be considered when free T4 is normal for age [178]; the first reassessment takes place one to two weeks after starting [179]. The US label differentiates the starting dose by age: 10 to 15 µg/kg/day from 0 to 3 months, 8 to 10 from 3 to 6 months, 6 to 8 from 6 to 12 months, 5 to 6 from 1 to 5 years, 4 to 5 from 6 to 12 years, 2 to 3 above 12 years with incomplete growth and puberty and 1.6 µg/kg/day when these are complete, with titration every 2 weeks by TSH and free T4 [180]. In forms with epilepsy, antiseizure medications are used in addition to sulfonylurea when seizures persist in KCNJ11 DEND, and the ketogenic diet controlled spasms in an infant with IER3IP1, with a greater tendency to ketoacidosis during intercurrent illness [1, 20]. In infants with developmental delay, developmental support, physiotherapy, occupational therapy, speech therapy and augmentative and alternative communication are recommended, in addition to each country's early intervention programmes [1, 13]. In 6q24, macroglossia is assessed by ENT or a craniofacial team, because of the possibility of airway obstruction, which has never been described as requiring treatment, and umbilical hernia may require surgery [13].

7.7. Special situations, pregnancy and transition

Drugs to avoid are rapid-acting insulin analogues and regular insulin as boluses in the infant, except in continuous infusion, and metformin in mitochondrial diabetes [1, 2]. In forms treated with glibenclamide, major surgery requires switching to insulin, and stress (trauma, surgery, febrile infection) may require temporary insulin; in mild to moderate hepatic or renal impairment the dose is reduced, starting with the lowest [24]. In WRS, general anaesthesia should be avoided whenever possible [72, 174].

In women with permanent ND, pregnancy should be followed by endocrinology and maternal-fetal medicine, with the glycaemic control recommended for diabetes in pregnancy, to prevent maternal complications, macrosomia and congenital anomalies, and with high-resolution ultrasound and fetal echocardiography [1]. Insulin is traditionally the mainstay of treatment; there are reports of safety and efficacy of glyburide in pregnancy, but a meta-analysis associated it with a higher risk of neonatal hypoglycaemia, and the European SmPC requires switching to insulin in women planning pregnancy and not using glibenclamide in pregnancy, because placental transfer, although small, is variable; the US label recommends stopping the sulfonylurea at least two weeks before delivery, because of prolonged neonatal hypoglycaemia of 4 to 10 days described with these drugs [1, 24, 168]. In breastfeeding, in 11 treated mothers glibenclamide was not excreted in milk and there was no hypoglycaemia in the infants, and monitoring of the glucose of exclusively breastfed infants is recommended [24]. In women with a history of 6q24-TNDM, pregnancy is a risk factor for relapse and diabetes may present as gestational diabetes, so professionals should be informed of the susceptibility [13]. In TRMA, there are no studies in pregnancy and good glycaemic control before and during gestation is recommended [21].

In the transition to adulthood, surveillance for complications, reassessment of the sulfonylurea dose (which may need to rise in adolescence) and neuropsychological and educational follow-up should be maintained; in 6q24, education about the risk of relapse and its symptoms (thirst, polyuria, repeated bacterial infections) should be directed at the parents and later at the patient, with a recommendation to measure glucose in the presence of these symptoms, and excessive weight gain and other cardiovascular risk factors should be avoided [4, 13]. Adults over 30 years transfer less often and frequently need 1 mg/kg/day of glibenclamide, but successful transfers have been described up to 60 years, so a therapeutic trial is recommended at any age [88].

7.8. Developments in the last twelve months

In the window from October 2025 to October 2026, developments were mainly genetic and technological, with no new drugs. In September 2025 the identification of TMEM167A as a recessive cause of a syndrome of microcephaly, epilepsy and diabetes in 6 individuals was published, with functional demonstration in iPSC-derived β cells (human study with a preclinical component) [28]. In these models, ISRIB, imeglimin and the GLP-1 receptor agonist exendin-4 increased β-cell survival under ER stress, with and without the variant; there is no approved disease-modifying treatment for the ER stress forms of ND, and these results are preclinical and do not support clinical decisions [28]. In May 2026 the de novo ACTB variant p.(Ser348Leu) was described in syndromic ND (human study) [29]. In February 2026 the implementation of a hybrid closed-loop system with diluted insulin in a 29-day-old infant was published (case report) and, in August 2026, the Portuguese experience with continuous monitoring and a sensor-integrated pump in TNDM (case report) [156, 160]. In March 2026 the 5-year follow-up with semaglutide and sitagliptin in a patient with KCNJ11 PNDM was published (case report, off-label use) [167]. In January 2026 the comparison of neurodevelopment in patients with KCNJ11 and their siblings appeared (observational study) [141], and in October 2025 and February, April and June 2026 national series and cases were published (Poland, Ireland, India, Thailand, HNF1B, Mitchell-Riley), described in sections 3, 5 and 9 [109, 134, 136, 137, 148, 158, 181]. On the regulatory side, the only milestone in the period was the publication of the Amglidia® risk management plan by the EMA on 11 May 2026, a document not consulted in this review [24]. The Europe PMC search did not find an update of the ISPAD guideline on monogenic diabetes later than 2022, which does not demonstrate that none exists. In animal models, glyburide, and not other sulfonylureas, promotes the differentiation of NEUROG3-positive progenitors into insulin-producing cells, and brain organoids with the p.Val59Met variant show defects in the development of the cortical neuronal network partly corrected by tolbutamide; both findings are preclinical and do not support clinical decisions [2, 4, 142].

8. Reproductive Options and Family Counselling

8.1. Recurrence risks by mechanism

The recurrence risk depends on the gene and, in 6q24, on the exact mechanism, so formal genetic counselling is strongly recommended [2, 13].

Sources: GeneReviews on PNDM and 6q24-TNDM, ISPAD 2022 and GeneReviews on TRMA [1, 2, 13, 21].

8.2. Parental testing and mosaicism

When a case is isolated, the parents are tested for the variant and their glycaemia is assessed clinically (fasting glucose, HbA1c and, rarely, OGTT), because hyperglycaemia may be asymptomatic [1]. If the variant is absent in both, with confirmed parentage, it is a de novo variant or gonadal mosaicism, described in KCNJ11 and in other families, with unknown incidence and not detectable in leucocyte DNA (in a KCNJ11 series, the pedigree indicated germline mosaicism in 2 of 18 families with unaffected parents, and somatic mosaicism was detected in 1 of 12 parents tested) [6]; the risk to siblings is low but not negligible [1, 2, 157]. A carrier parent may be undiagnosed, because of reduced penetrance (ABCC8, KCNJ11) or a mild phenotype [1]. In STAT3 gain-of-function disease, usually de novo, families with incomplete penetrance have been described, so a healthy parent may be a carrier [115]. In recessive inheritance, the parents are presumed heterozygotes and may or may not have diabetes; in ABCC8 forms, heterozygotes may have normal glucose tolerance, and in GCK, INS or PDX1 forms they may have mild diabetes [1]. In 6q24, if the duplication is not present in the father with confirmed paternity, it is considered de novo or paternal gonadal mosaicism; in ZFP57 MLID, if only one parent is a carrier, a de novo variant or parental mosaicism is considered, and apparent homozygosity may correspond to an undetected exonic deletion or to uniparental isodisomy [13]. In highly consanguineous families, a parent may be homozygous for ZFP57 [13]. In TRMA, homozygosity may also result from uniparental isodisomy, and carriers are asymptomatic and not at risk [21].

8.3. Cascade testing and at-risk relatives

Asymptomatic relatives, younger and older, should be assessed: molecular testing if the variant is known and, otherwise, HbA1c or fasting and postprandial glucose, rarely OGTT; continuous glucose monitoring has been used in relatives of patients with monogenic diabetes [1]. In 6q24, clarifying the status of siblings at risk of paternal duplication or biallelic ZFP57 allows hyperglycaemia to be monitored in the neonatal period and later [13]. In TRMA, it is recommended to test siblings and supplement them with thiamine at pharmacological doses of 25 to 100 mg/day as early as possible, until the genotype is known [21]. In families with 6q24 duplication, the father is at risk of later diabetes or has a history of early-onset diabetes, and should be assessed with targeted analysis and karyotype, to exclude a balanced translocation or complex rearrangement [13]. ZFP57 heterozygotes have no known phenotype or increased risk of diabetes [13]. ISPAD recommends referring families to a specialist in monogenic diabetes or to a clinical genetics unit, to guide management and to test affected or presymptomatic relatives [2].

8.4. Prenatal diagnosis, preimplantation testing and planning

Once the familial variant has been identified, PND and PGT are technically possible; the optimal time for risk assessment is before pregnancy, and counselling of affected or at-risk young adults and referral to maternal-fetal medicine are recommended [1]. PND is requested mainly for the severe syndromic forms that are potentially fatal in childhood, such as WRS and IPEX syndrome [3]. In 6q24, PND and PGT apply to biallelic ZFP57, paternal duplication and UPD6 associated with a structural anomaly, with the caveat that the MLID phenotype cannot be predicted [13]. When there is no molecular diagnosis, or when the cause of hypomethylation remains unknown, DNA banking for future analysis is recommended [1, 13].

8.5. Content of counselling

In addition to recurrence risk, counselling should address the practical consequences of the diagnosis: the possibility of switching from insulin to sulfonylurea in the KATP channel forms, with effect also in adult relatives diagnosed late; the risk of later diabetes in asymptomatic carriers of TNDM variants and of 6q24 duplications; surveillance for HFE haemochromatosis and other recessive disorders in isodisomic UPD6; the risk of hypoglycaemia in relatives with inactivating variants of the KATP channel genes; and the progressive or multisystem nature of the syndromic forms [2, 13, 126]. In families with WRS and other AR forms, consanguinity and geographical origin are part of the assessment [2, 22]. In the Irish series of 19 patients, the genetic diagnosis informed family risk and changed the management of one parent [181].

9. Special Considerations: Population Diversity and Epidemiology

9.1. Prevalence and incidence: what each figure measures

Published estimates derive from national registries, referral rates to diagnostic laboratories and centre series, and are not directly comparable; some measure birth prevalence, others annual incidence, and the denominators range from one region to several countries.

Sources: Orphadata, GeneReviews, Oxford Desk Reference, 2024 review and French cohort [1, 3, 4, 13–17, 19, 142, 182–192]. These estimates cannot be converted into a national prevalence or added together: they measure different populations, periods and forms of detection, and those based on referrals for testing underestimate untested cases.

About half of ND cases are permanent and half transient; within TNDM, about 70% are due to 6q24 abnormalities according to GeneReviews, and about two thirds according to ISPAD [1, 2, 13]. Monogenic diabetes accounts for 2.5 to 6.5% of paediatric diabetes, and monogenic defects for less than 5% of all people with diabetes; among those diagnosed before 6 months, 80 to 85% have an identifiable monogenic cause [2, 23].

9.2. Distribution of causes by region

Consanguinity determines both incidence and aetiology: in non-consanguineous populations dominant or de novo variants of the KATP channel and INS predominate, and in consanguineous populations the recessive forms [2, 4].

Sources: 2024 worldwide review (Table 1) and 2025–2026 publications [4, 134, 137, 148, 181]. The series differ in inclusion criteria (some include onset up to 9 or 12 months) and in the panel used, which limits comparison.

In the three recent European and American series (Italy, Ukraine, Brazil), KCNJ11 is the main cause of PNDM (34.8 to 44.4%), followed by dominant INS (17.7 to 26% in Ukraine and Italy), with rare recessive genes and monogenic autoimmune PNDM in 6.6 to 12.5% [4]. In the Middle East and Africa, recessive genes predominate, with a probable founder effect: EIF2AK3 in 11.1 to 53.3% and the PTF1A enhancer in 14 to 54.5% of the series from Abu Dhabi, Qatar, Egypt and Kurdistan [4]. For the PTF1A enhancer, the founder effect has been demonstrated: in the original series, 9 patients from 6 Syrian, Lebanese, Kurdish and Turkish families were homozygous for the same point variant, on a shared 1.2 Mb haplotype [102]. In WRS, 2 apparently unrelated children, born in the same region of Quebec and with the same French surname, were homozygous for the same nonsense variant of EIF2AK3, with discordant phenotypes [12, 37]. Israel resembles Europe, with 66% heterozygous KCNJ11 and INS variants [4]. In south-eastern Turkey, 86.7% of the 15 PNDM cases were recessive; in two other Turkish series with 17 patients, 64.8% were biallelic [4]. In Asia, Pakistan and India had 73.3% recessive forms among 30 patients with PNDM, and in Vietnam consanguinity is lower, with a predominance of KCNJ11, ABCC8, INS and 6q24 [4]. In TNDM, 6q24 accounts for 25 to 44.4% of European and American series, 84.6% of the Japanese series and 68.7% of the Vietnamese [4].

9.3. Data from Portugal and Brazil

In the sources consulted, Portuguese data are limited to the description, by a team from the Unidade Local de Saúde de Santa Maria (Lisbon), of the safe use of continuous glucose monitoring and a sensor-integrated pump in early-onset TNDM, of which only the title and affiliation could be read; no Portuguese national series of ND was identified, which is a limitation of this review and does not demonstrate that no published data exist [160]. From Brazil, the series of Costa-Riquetto and colleagues, cited in the worldwide review, included 16 patients with PNDM (KCNJ11 43%, GCK 25%, FOXP3 12.5%, ABCC8, INS and PTF1A enhancer with one case each) and 4 with TNDM (ABCC8 3, 6q24 1) [4]. The GeneReviews on TRMA includes Brazil among the countries of origin of the families described with this syndrome [21].

9.4. Special populations

Consanguineous and endogamous populations, including those of the Middle East, North Africa and the Indian subcontinent, have an incidence several times higher and a distinct genetic spectrum, dominated by recessive syndromic forms, which changes the priority of genes to test and the probability of response to sulfonylurea [2, 4]. TRMA is described almost exclusively in consanguineous families or isolated populations, with about 130 kindreds: Israeli Arabs, Lebanese, an Alaskan family of Native and Russian origin, Brazil, Japan, Oman, Tunisia, Italy (Veneto and other regions), Iran, India, Pakistan, Kashmiri families in the United Kingdom, Kurds, northern Europeans and African Americans [21]. In the French oral suspension cohort, 73% of patients were of European origin, 19% from North Africa and the rest from Asia and sub-Saharan Africa, a composition that limits the generalisation of the results [142]. In resource-limited settings, severe malnutrition delays metabolic control, access to genetics is late and unvalidated targeted tests may be used, with a risk of wrong therapeutic decisions [158]. Diagnostic delay in WRS may reflect access to health care [22]. Children with trisomy 21 are over-represented among cases of autoimmune ND [2].

9.5. Reference networks and resources

The relevant European Reference Network is Endo-ERN (European Reference Network on Rare Endocrine Conditions), whose thematic group MTG3, "Genetic disorders of glucose & insulin homeostasis", explicitly includes patients with very unstable or difficult-to-treat diabetes, such as neonatal diabetes, as well as hyperinsulinism, hypoglycaemia, rare diabetes and lipodystrophy; in Portugal, APDP – Diabetes Portugal (Lisbon) is a reference centre for MTG3, and the Unidade Local de Saúde de Santa Maria and the Centro Hospitalar Universitário de São João belong to other thematic groups [26]. The French centre PRISIS, at the Hôpital Necker in Paris, is a member of the network [142]. For the Tregopathies, the relevant network is ERN RITA, and for the liver disease of WRS ERN RARE-LIVER, not consulted directly in this review [26].

The reference laboratories and programmes with genetic testing and specialist counselling indicated by ISPAD and GeneReviews are Exeter, in the United Kingdom (diabetesgenes.org, sometimes free of charge in a research setting), the University of Chicago Monogenic Diabetes Registry (active since 2008), the Norwegian programme (mody.no), the euro-wabb.org consortium for Wolfram, Alström and Bardet-Biedl syndromes, the Ospedale Bambino Gesù in Rome, the Hôpital Robert-Debré (AP-HP) in Paris, and the monogenic and atypical diabetes programme of the Children's Hospital of Philadelphia [1, 2, 88]. The Oxford Desk Reference notes that genetic testing is available free of charge through diabetesgenes.org for all those diagnosed before 9 months [3]. There is also a British registry of transient neonatal diabetes [13]. The patient associations and organisations mentioned by the sources are Diabetes UK, the American Diabetes Association, ISPAD and, in Portugal, APDP; no association specific to neonatal diabetes, nor a dedicated Brazilian association, was identified in the sources consulted, which remains a limitation [1, 13, 26].

10. Illustrative Clinical Case

The following case was published by Ghaben and colleagues in 2026, in open access, and is summarised here faithfully to the report [156].

A term newborn, born at 38 weeks by induction for oligohydramnios, with prenatal IUGR, weighed 2,630 g (Z-score −1.61) and measured 47.6 cm (Z-score −1.21). About ten hours after birth he had hypoglycaemia of 38 mg/dL, refractory to glucose gel, treated with 10% intravenous dextrose. At 18 hours of life he developed hyperglycaemia above 300 mg/dL, which persisted after dextrose was stopped, with breast milk as the only source of carbohydrate. On examination he had little subcutaneous tissue and no dysmorphic features. At 87 hours, glucose was 344 mg/dL, with insulin below 1.6 µU/mL, C-peptide of 0.18 ng/mL (0.06 nmol/L) and beta-hydroxybutyrate of 0.19 mmol/L. On two abdominal ultrasound scans, the pancreas was only partially visualised.

Rapid genome sequencing, requested on day four, was negative on day 14. MS-MLPA, requested that day, showed maternal hypomethylation of the 6q24 region, confirmed in the sixth week, after discharge. Before the genetic result, empirical glyburide was tried at a dose of 1 mg/kg/day, stopped after 48 hours for lack of efficacy; the authors cite a rate of about 40% of sulfonylurea responders in ND, a value lower than that of the ISPAD guideline for the KATP channel forms (section 7.3). From day 7, the infant received intravenous regular insulin by infusion, with technical difficulties due to the very low rate, as well as subcutaneous corrections with lispro, and a continuous glucose monitoring sensor was placed on the thigh, validated for 24 hours against serum glucose.

On day 29 a hybrid closed-loop system was started, with insulin lispro at the usual concentration, basal rate programmed at zero, maximum sensitivity factor and night mode, without carbohydrate boluses and with free feeding; this was use outside the terms of the device's authorisation, with parental consent. Mean glucose fell from 302 to 162 mg/dL and time in the range of 70 to 180 mg/dL rose from 6.6% to 56.9% (p < 0.001), with an effective basal delivery of 0.02 to 0.03 U/kg/h, but time below 70 mg/dL rose from 0.36% to 5.75%. The switch to insulin lispro diluted to 10 U/mL, prepared in the pharmacy and with 28 days' refrigerated stability, with an effective basal of 0.03 U/h and a higher target, led to a mean glucose of 180 mg/dL, 45.5% time in target, no hypoglycaemia and a coefficient of variation of 24.1%. The infant was discharged on the diluted insulin, had an HbA1c of 4.8% two months later and the diabetes remitted at around 4 months, a course typical of 6q24-TNDM.

The report includes two figures consulted in this review: Figure 1, with the clinical and diagnostic timeline (https://pmc.ncbi.nlm.nih.gov/articles/PMC12930193/figure/luaf338-F1/), and Figure 2, with the mean seven-day glucose profile obtained by continuous monitoring for each treatment modality and the frequency distribution of glucose values (https://pmc.ncbi.nlm.nih.gov/articles/PMC12930193/figure/luaf338-F2/); the legends were read, without visual analysis of the images. The teaching points highlighted by the authors are the adaptations needed in the infant (diluted insulin, omission of carbohydrate boluses, higher glycaemic target), the fact that hypoglycaemia is the most serious risk and the need to educate caregivers while still in hospital. The approaches described are not routine: they are the experience of one centre with a system used outside the terms of its authorisation, and the guideline keeps subcutaneous insulin, diluted or by pump, as the standard treatment for TNDM (section 7.2).

A second case illustrates the severe presentation of the same entity [138]. A term boy with IUGR, weighing 1,990 g (0.17th centile), 44 cm long and with a head circumference of 31 cm, the son of a woman whose previous pregnancy had also been complicated by IUGR, had ten capillary glucose values between 55 and 99 mg/dL in the first 25 hours and was discharged on the second day, with appointments on days 3, 6 and 14, at the last of which he had gained 15% of his weight. On day 19 he developed respiratory distress, interpreted as bronchiolitis, which progressed to shock and respiratory failure. Glucose was 1,260 mg/dL (70 mmol/L), venous pH 6.92, bicarbonate 10 mmol/L, lactate 9.6 mmol/L, with ketones on urinary organic acids, aspartate aminotransferase 518 U/L and alanine aminotransferase 102 U/L, ammonia 81.6 µmol/L, acute kidney injury, international normalised ratio 2.0 and fibrinogen 142 mg/dL; insulin was below 1 µU/mL and C-peptide 0.6 ng/mL. Cerebrospinal fluid showed leucocytosis with raised protein and lactate, and the infectious work-up was negative. MRI revealed bilateral thalamic hyperintensities, which raised the suspicion of mitochondrial disease and prompted the introduction of thiamine, riboflavin, a multivitamin and antiseizure medications. Treatment included ventilation, vasopressors, bicarbonate and insulin infusion, with resolution of the acidosis in about 12 hours; insulin was stopped on day 22 of admission, with glucose values of 150 to 200 mg/dL without treatment. Exome sequencing identified paternal UPD6 by isodisomy, establishing the diagnosis of 6q24-TNDM, and the patient remained off insulin. MRI at one year showed focal encephalomalacia with haemosiderin in the left thalamus, attributed to previous haemorrhage due to probable venous thrombosis associated with dehydration, and motor deficits persisted, improving with physiotherapy and occupational therapy. The published figure is the axial T2-weighted MRI from the second day of admission, with the thalamic hyperintensities marked (https://pmc.ncbi.nlm.nih.gov/articles/PMC12448923/figure/luaf145-F1/). The authors emphasise that, in the international cohort of 155 patients with 6q24-TNDM, the mean age at presentation was 8 days, with a mode of 1 day, that in the Chicago Registry there was ketoacidosis in 66% of 88 patients with ND but in none of the 10 with 6q24, so the absence of ketoacidosis does not exclude this aetiology, and that normal glucose values in the first 24 to 48 hours do not exclude ND in a small-for-gestational-age newborn.

11. Prognosis and Long-Term Follow-Up

11.1. Prognosis by entity

In the KATP channel forms transferred to sulfonylurea, the metabolic prognosis is good and durable: 93% of patients with KCNJ11 remained on sulfonylurea alone after a median of 10.2 years, with HbA1c of 6.4%, and 86% of patients with ABCC8 remained insulin-independent after about 10 years, with HbA1c of 6.5%; there was no severe hypoglycaemia and microvascular complications occurred in 9% of patients with KCNJ11, associated with later transfer [4, 24, 144, 145]. Across cohorts, microalbuminuria ranged from 4 to 19% and retinopathy was 6%, and in a group of 15 patients with KATP channel or INS PNDM there were no severe ocular complications after a median of 24 years of diabetes [1, 2]. The neurological prognosis depends on the variant and on how early treatment is started: about half of patients with neurological features improve on starting sulfonylurea, improvement is greater in those treated as infants, and among patients with KCNJ11 47% had initial improvement although 64% retained central nervous system features [4, 145, 165].

In 6q24-TNDM, remission is almost universal, but at least 50 to 60% relapse, and in a French cohort followed to 18 years relapse reached 85%; educational outcome is lower than expected and patients with diabetes have reduced secretory capacity [2, 13]. In WRS, patient survival was 89.4% at one year, 65.5% at 5 years and 53.1% at 10 years, with 29.6% deaths (56 of 189) and a median age at death of 36 months; liver failure was the main cause of death (17.9%), followed by unknown cause (6.3%), multiorgan failure (4.8%) and metabolic decompensation (3.2%). Overall survival was better with missense variants on one or both alleles than with two loss-of-function variants (p = 0.013) and worse in patients with acute liver injury (p = 0.014). Native liver survival was 99.4%, 96.1% and 89.5% at 1, 5 and 10 years; the 9 transplanted patients were all alive, with a median follow-up of 24 months and maximum of 14 years, and in patients with acute liver injury or failure transplantation was associated with better survival (p = 0.0057), in an observational comparison subject to candidate selection [22]. The bone dysplasia atlas states that most patients die young, with survival observed up to 35 years [72]. In TRMA, thiamine corrects the anaemia within 6 to 8 weeks and avoids transfusions, and 11 of 15 treated patients had glycaemic benefit, but deafness is irreversible and patients with diabetes usually become insulin-dependent; a stroke at 2 years on insulin and thiamine and a death at 7 years from ketoacidosis after a viral illness have been described [21, 25]. In the forms with pancreatic agenesis, there is no postnatal catch-up growth, unlike the KATP channel, INS and GCK forms [4, 27]. The MEDS forms have a guarded prognosis, with deaths in early childhood described [20, 111]. In 6q24 with MLID, long-term outcomes are unknown [13].

11.2. Surveillance table

The intervals coincide with those given in section 7: HbA1c every 3 months on glibenclamide according to the Amglidia® SmPC, HbA1c at least annually after remission in KATP channel TNDM according to Exeter, and surveillance for relapse for 6 to 24 months after remission in 6q24 according to GeneReviews [13, 24, 95].

11.3. Long-term prevention and education

In 6q24, the parents and, later, the patient should understand the high risk of relapse and recognise the symptoms, and excessive weight gain and other cardiovascular risk factors, which predispose to later diabetes, should be avoided [2, 13]. In patients on sulfonylurea, follow-up should record HbA1c, C-peptide, time in the range of 3.5 to 10 mmol/L and the doses and their distribution through the day, and repeat developmental testing; the dose should keep pace with weight gain [88, 165]. In WRS, families should be taught to recognise a hepatic episode early and care should be multidisciplinary, including paediatric orthopaedics [22]. In TRMA, thiamine is lifelong and stopping it causes the anaemia to relapse [21].

12. Summary and Final Considerations

Diabetes diagnosed in the first six months of life is monogenic in 80 to 85% of cases, and age at diagnosis is the only feature that suffices to indicate genetic testing. Half of the cases are transient, two thirds of these due to overexpression of the imprinted 6q24 locus, and half are permanent, with a predominance of activating variants of the KATP channel genes and INS variants in non-consanguineous populations and of recessive syndromic forms, above all Wolcott-Rallison syndrome, biallelic GCK and the PTF1A enhancer, in consanguineous populations, where the incidence is up to ten times higher. The list of genes exceeds forty and continues to grow, with TMEM167A and ACTB described in 2025 and 2026, but the cause remains unidentified in less than 20% of cases.

Molecular diagnosis is made by next-generation sequencing with a comprehensive panel or by genome, always with methylation analysis of the 6q24 region and a search for deletions and duplications, without waiting for remission or the appearance of other features. A panel that does not cover the distal PTF1A enhancer, the INS promoter regions, mitochondrial DNA and methylation misses causes that are treatable or have distinct family implications, and a negative result does not exclude monogenic disease. Classification follows ACMG/AMP standards with the specifications of the international panel dedicated to monogenic diabetes, and the report should state the transcript with version, the zygosity and the limitations of the method.

The impact of the diagnosis is therapeutic. About 90% of patients with KATP channel variants switch from insulin to oral glibenclamide, with sustained improvement in glycaemic control over ten years, rare severe hypoglycaemia and partial neurological benefit, greater the earlier treatment begins; the guideline reference dose is about 0.5 mg/kg/day, well above that used in type 2 diabetes, with values up to 2.3 to 2.8 mg/kg/day described, and the oral suspension authorised in the European Union since 2018 is not bioequivalent to crushed tablets. In the other forms, the treatment is insulin, with the adaptations required by the infant, to which are added thiamine in thiamine-responsive megaloblastic anaemia syndrome, pancreatic enzymes in pancreatic agenesis, immunosuppression or haematopoietic stem cell transplantation in the autoimmune forms and, in Wolcott-Rallison syndrome, liver transplantation in selected patients, in a disease in which survival at ten years is about half and liver failure is the main cause of death.

Follow-up extends well beyond infancy. In the transient 6q24 form, remission does not end the disease: at least half of patients relapse, usually around puberty, and women may relapse in pregnancy, which justifies glucose surveillance after remission and annual glycated haemoglobin from 8 to 10 years. In the permanent forms, renal function and the retina are monitored from ten years of age, as are development and exocrine function, and, in paternal uniparental disomy of chromosome 6, the risk of unmasked recessive disorders, notably haemochromatosis. Counselling depends on the exact mechanism, which distinguishes recurrence risks close to zero, of 1%, of 25% or of 50% within the same clinical presentation.

Finally, the diagnosis has family-wide and lifelong reach: adults treated for decades as having type 1 diabetes, but diagnosed before six months, should be tested at any age, because transfer to sulfonylurea remains possible, and a result in an infant may reclassify the diabetes of one of the parents. Referral to a centre with experience in monogenic diabetes, integrated into the relevant European network, is part of the treatment and not an accessory to it.

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/
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. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 390–393.
4. 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.
5. OMIM® Online Mendelian Inheritance in Man. #606176 DIABETES MELLITUS, PERMANENT NEONATAL, 1; PNDM1. Johns Hopkins University; last edited 01/12/2020. https://omim.org/entry/606176. Accessed 28 October 2026.
6. OMIM® Online Mendelian Inheritance in Man. #618856 DIABETES MELLITUS, PERMANENT NEONATAL, 2; PNDM2. Johns Hopkins University; last edited 01/12/2020. https://omim.org/entry/618856. Accessed 28 October 2026.
7. OMIM® Online Mendelian Inheritance in Man. #618857 DIABETES MELLITUS, PERMANENT NEONATAL, 3; PNDM3. Johns Hopkins University; last edited 01/12/2020. https://omim.org/entry/618857. Accessed 28 October 2026.
8. OMIM® Online Mendelian Inheritance in Man. #618858 DIABETES MELLITUS, PERMANENT NEONATAL, 4; PNDM4. Johns Hopkins University; last edited 01/12/2020. https://omim.org/entry/618858. Accessed 28 October 2026.
9. OMIM® Online Mendelian Inheritance in Man. #601410 DIABETES MELLITUS, TRANSIENT NEONATAL, 1; TNDM1. Johns Hopkins University; last edited 04/12/2024. https://omim.org/entry/601410. Accessed 28 October 2026.
10. OMIM® Online Mendelian Inheritance in Man. #610374 DIABETES MELLITUS, TRANSIENT NEONATAL, 2; TNDM2. Johns Hopkins University; last edited 01/12/2020. https://omim.org/entry/610374. Accessed 28 October 2026.
11. OMIM® Online Mendelian Inheritance in Man. #610582 DIABETES MELLITUS, TRANSIENT NEONATAL, 3; TNDM3. Johns Hopkins University; last edited 04/12/2024. https://omim.org/entry/610582. Accessed 28 October 2026.
12. OMIM® Online Mendelian Inheritance in Man. #226980 EPIPHYSEAL DYSPLASIA, MULTIPLE, WITH EARLY-ONSET DIABETES MELLITUS. Johns Hopkins University; last edited 29/07/2025. https://omim.org/entry/226980. Accessed 28 October 2026.
13. Greeley SAW, Del Gaudio D. 6q24-Related Transient Neonatal Diabetes Mellitus. 2005 Oct 10 [updated 2026 Aug 13]. 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/NBK1534/
14. Orphanet. Isolated permanent neonatal diabetes mellitus (ORPHA:99885) — epidemiology. Available from: https://www.orpha.net/en/disease/detail/99885
15. Orphanet. Neonatal diabetes mellitus (ORPHA:224). Available from: https://www.orpha.net/en/disease/detail/224
16. Orphanet. Transient neonatal diabetes mellitus (ORPHA:99886). Available from: https://www.orpha.net/en/disease/detail/99886
17. Orphanet. DEND syndrome (ORPHA:79134). Available from: https://www.orpha.net/en/disease/detail/79134
18. Orphanet. Intermediate DEND syndrome (ORPHA:99989). https://www.orpha.net/en/disease/detail/99989
19. Orphanet. Wolcott-Rallison syndrome (ORPHA:1667). https://www.orpha.net/en/disease/detail/1667
20. GeneGuidance: gene records KCNJ11, ABCC8, INS, GCK, PLAGL1, ZFP57, EIF2AK3, FOXP3, PTF1A, GATA6, GATA4, PDX1, NEUROD1, RFX6, SLC2A2, SLC19A2, GLIS3, NKX2-2, MNX1, IER3IP1, HNF1B, NEUROG3, YIPF5, EIF2B1, CNOT1, ONECUT1, STAT3, LRBA, IL2RA. Accessed 2026-10-26.
21. Sako S, Tsunogai T, Oishi K. Thiamine-Responsive Megaloblastic Anemia Syndrome. 2003 Oct 24 [updated 2022 Jul 28]. 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/NBK1282/
22. Aldrian D, Bochdansky C, Kavallar AM, Mayerhofer C, Deeb A, Habeb A, et al. Natural history of Wolcott-Rallison syndrome: A systematic review and follow-up study. Liver Int. 2024;44(3):811-822. doi:10.1111/liv.15834. PMID: 38230874.
23. American Diabetes Association Professional Practice Committee for Diabetes. 2. Diagnosis and Classification of Diabetes: Standards of Care in Diabetes—2026. Diabetes Care. 2026;49(Suppl 1):S27–S49. doi:10.2337/dc26-S002. PMID: 41358893. PMC12690183.
24. European Medicines Agency. Amglidia (glibenclamide) 0.6 mg/mL and 6 mg/mL oral suspension: EPAR – Product Information (Summary of Product Characteristics). AMMTeK; first authorisation 24 May 2018; last updated 06 Dec 2023. Available from: https://www.ema.europa.eu/en/documents/product-information/amglidia-epar-product-information_en.pdf ; medicine page: https://www.ema.europa.eu/en/medicines/human/EPAR/amglidia
25. Habeb AM, Flanagan SE, Zulali MA, Abdullah MA, Pomahačová R, Boyadzhiev V, et al. Pharmacogenomics in diabetes: outcomes of thiamine therapy in TRMA syndrome. Diabetologia. 2018;61(5):1027-1036. doi:10.1007/s00125-018-4554-x. PMID: 29450569. PMC6449001.
26. European Reference Network on Rare Endocrine Conditions (Endo-ERN). Genetic disorders of glucose & insulin homeostasis (MTG3); Reference Centres. Available from: https://endo-ern.eu/rare-genetic-disorders-of-glucose-insulin-homeostasis/ ; https://endo-ern.eu/about/reference-centres/
27. 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:e4351-9.
28. 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). doi:10.1172/JCI195756. PMID: 40924476. PMC12618065.
29. 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.
30. OMIM® Online Mendelian Inheritance in Man. *600937 POTASSIUM CHANNEL, INWARDLY RECTIFYING, SUBFAMILY J, MEMBER 11; KCNJ11. Johns Hopkins University; last edited 02/04/2025. https://omim.org/entry/600937. Accessed 28 October 2026.
31. OMIM® Online Mendelian Inheritance in Man. *600509 ATP-BINDING CASSETTE, SUBFAMILY C, MEMBER 8; ABCC8. Johns Hopkins University; last edited 09/05/2025. https://omim.org/entry/600509. Accessed 28 October 2026.
32. OMIM® Online Mendelian Inheritance in Man. *176730 INSULIN; INS. Johns Hopkins University; last edited 26/01/2024. https://omim.org/entry/176730. Accessed 28 October 2026.
33. OMIM® Online Mendelian Inheritance in Man. *138079 GLUCOKINASE; GCK. Johns Hopkins University; last edited 17/01/2023. https://omim.org/entry/138079. Accessed 28 October 2026.
34. OMIM® Online Mendelian Inheritance in Man. *603044 PLAG1-LIKE ZINC FINGER 1; PLAGL1. Johns Hopkins University; last edited 21/11/2024. https://omim.org/entry/603044. Accessed 28 October 2026.
35. OMIM® Online Mendelian Inheritance in Man. *606546 HYDATIDIFORM MOLE-ASSOCIATED AND IMPRINTED TRANSCRIPT; HYMAI. Johns Hopkins University; last edited 04/03/2019. https://omim.org/entry/606546. Accessed 28 October 2026.
36. OMIM® Online Mendelian Inheritance in Man. *612192 ZFP57 ZINC FINGER PROTEIN; ZFP57. Johns Hopkins University; last edited 27/09/2019. https://omim.org/entry/612192. Accessed 28 October 2026.
37. OMIM® Online Mendelian Inheritance in Man. *604032 EUKARYOTIC TRANSLATION INITIATION FACTOR 2-ALPHA KINASE 3; EIF2AK3. Johns Hopkins University; last edited 05/10/2017. https://omim.org/entry/604032. Accessed 28 October 2026.
38. OMIM® Online Mendelian Inheritance in Man. *300292 FORKHEAD BOX P3; FOXP3. Johns Hopkins University; last edited 07/08/2025. https://omim.org/entry/300292. Accessed 28 October 2026.
39. OMIM® Online Mendelian Inheritance in Man. #304790 IMMUNODYSREGULATION, POLYENDOCRINOPATHY, AND ENTEROPATHY, X-LINKED; IPEX. Johns Hopkins University; last edited 15/10/2025. https://omim.org/entry/304790. Accessed 28 October 2026.
40. OMIM® Online Mendelian Inheritance in Man. *607194 PANCREAS TRANSCRIPTION FACTOR 1, ALPHA SUBUNIT; PTF1A. Johns Hopkins University; last edited 03/11/2023. https://omim.org/entry/607194. Accessed 28 October 2026.
41. OMIM® Online Mendelian Inheritance in Man. #260370 PANCREATIC AGENESIS 1; PAGEN1. Johns Hopkins University; last edited 23/10/2024. https://omim.org/entry/260370. Accessed 28 October 2026.
42. OMIM® Online Mendelian Inheritance in Man. *600733 PANCREAS/DUODENUM HOMEOBOX PROTEIN 1; PDX1. Johns Hopkins University; last edited 17/06/2025. https://omim.org/entry/600733. Accessed 28 October 2026.
43. OMIM® Online Mendelian Inheritance in Man. #600001 HEART DEFECTS, CONGENITAL, AND OTHER CONGENITAL ANOMALIES; HDCA. Johns Hopkins University; last edited 29/01/2018. https://omim.org/entry/600001. Accessed 28 October 2026.
44. OMIM® Online Mendelian Inheritance in Man. *601656 GATA-BINDING PROTEIN 6; GATA6. Johns Hopkins University; last edited 21/10/2025. https://omim.org/entry/601656. Accessed 28 October 2026.
45. OMIM® Online Mendelian Inheritance in Man. *600576 GATA-BINDING PROTEIN 4; GATA4. Johns Hopkins University; last edited 01/12/2025. https://omim.org/entry/600576. Accessed 28 October 2026.
46. OMIM® Online Mendelian Inheritance in Man. *601724 NEUROGENIC DIFFERENTIATION 1; NEUROD1. Johns Hopkins University; last edited 21/04/2025. https://omim.org/entry/601724. Accessed 28 October 2026.
47. OMIM® Online Mendelian Inheritance in Man. #615710 MITCHELL-RILEY SYNDROME; MTCHRS. Johns Hopkins University; last edited 22/11/2024. https://omim.org/entry/615710. Accessed 28 October 2026.
48. OMIM® Online Mendelian Inheritance in Man. *612659 REGULATORY FACTOR X, 6; RFX6. Johns Hopkins University; last edited 21/04/2025. https://omim.org/entry/612659. Accessed 28 October 2026.
49. OMIM® Online Mendelian Inheritance in Man. #227810 FANCONI-BICKEL SYNDROME; FBS. Johns Hopkins University; last edited 01/10/2019. https://omim.org/entry/227810. Accessed 28 October 2026.
50. OMIM® Online Mendelian Inheritance in Man. *138160 SOLUTE CARRIER FAMILY 2 (FACILITATED GLUCOSE TRANSPORTER), MEMBER 2; SLC2A2. Johns Hopkins University; last edited 03/09/2020. https://omim.org/entry/138160. Accessed 28 October 2026.
51. OMIM® Online Mendelian Inheritance in Man. #249270 THIAMINE-RESPONSIVE MEGALOBLASTIC ANEMIA SYNDROME; TRMA. Johns Hopkins University; last edited 09/12/2022. https://omim.org/entry/249270. Accessed 28 October 2026.
52. OMIM® Online Mendelian Inheritance in Man. *603941 SOLUTE CARRIER FAMILY 19 (THIAMINE TRANSPORTER), MEMBER 2; SLC19A2. Johns Hopkins University; last edited 22/02/2022. https://omim.org/entry/603941. Accessed 28 October 2026.
53. OMIM® Online Mendelian Inheritance in Man. *102582 SIGNAL TRANSDUCER AND ACTIVATOR OF TRANSCRIPTION 3; STAT3. Johns Hopkins University; last edited 11/08/2025. https://omim.org/entry/102582. Accessed 28 October 2026.
54. OMIM® Online Mendelian Inheritance in Man. #614231 MICROCEPHALY, EPILEPSY, AND DIABETES SYNDROME 1; MEDS1. Johns Hopkins University; last edited 16/04/2021. https://omim.org/entry/614231. Accessed 28 October 2026.
55. OMIM® Online Mendelian Inheritance in Man. #619278 MICROCEPHALY, EPILEPSY, AND DIABETES SYNDROME 2; MEDS2. Johns Hopkins University; last edited 16/04/2021. https://omim.org/entry/619278. Accessed 28 October 2026.
56. Wolcott CD, Rallison ML. Infancy-onset diabetes mellitus and multiple epiphyseal dysplasia. J Pediatr. 1972;80(2):292-7. doi:10.1016/s0022-3476(72)80596-1. PMID: 5008828.
57. Delépine M, Nicolino M, Barrett T, Golamaully M, Lathrop GM, Julier C. EIF2AK3, encoding translation initiation factor 2-alpha kinase 3, is mutated in patients with Wolcott-Rallison syndrome. Nat Genet. 2000;25(4):406-9. doi:10.1038/78085. PMID: 10932183.
58. Temple IK, Gardner RJ, Mackay DJ, Barber JC, Robinson DO, Shield JP. Transient neonatal diabetes: widening the understanding of the etiopathogenesis of diabetes. Diabetes. 2000;49:1359-66.
59. Mackay DJ, Callaway JL, Marks SM, White HE, Acerini CL, Boonen SE, et al. Hypomethylation of multiple imprinted loci in individuals with transient neonatal diabetes is associated with mutations in ZFP57. Nat Genet. 2008;40:949-51.
60. Gloyn AL, Pearson ER, Antcliff JF, Proks P, Bruining GJ, Slingerland AS, et al. Activating mutations in the gene encoding the ATP-sensitive potassium-channel subunit Kir6.2 and permanent neonatal diabetes. N Engl J Med. 2004;350(18):1838-49. doi:10.1056/NEJMoa032922. PMID: 15115830.
61. Pearson ER, Flechtner I, Njølstad PR, Malecki MT, Flanagan SE, Larkin B, et al. Switching from insulin to oral sulfonylureas in patients with diabetes due to Kir6.2 mutations. N Engl J Med. 2006;355(5):467-77. doi:10.1056/NEJMoa061759. PMID: 16885550.
62. Babenko AP, Polak M, Cavé H, Busiah K, Czernichow P, Scharfmann R, et al. Activating mutations in the ABCC8 gene in neonatal diabetes mellitus. N Engl J Med. 2006;355:456-66.
63. Sellick GS, Barker KT, Stolte-Dijkstra I, et al. Mutations in PTF1A cause pancreatic and cerebellar agenesis. Nat Genet. 2004;36:1301-5.
64. Senée V, Chelala C, Duchatelet S, et al. Mutations in GLIS3 are responsible for a rare syndrome with neonatal diabetes mellitus and congenital hypothyroidism. Nat Genet. 2006;38:682-7.
65. Støy J, Edghill EL, Flanagan SE, Ye H, Paz VP, Pluzhnikov A, et al. Insulin gene mutations as a cause of permanent neonatal diabetes. Proc Natl Acad Sci U S A. 2007;104:15040-4.
66. Garin I, Edghill EL, Akerman I, Rubio-Cabezas O, Rica I, Locke JM, et al. Recessive mutations in the INS gene result in neonatal diabetes through reduced insulin biosynthesis. Proc Natl Acad Sci U S A. 2010;107:3105-10.
67. Allen HL, Flanagan SE, Shaw-Smith C, et al. GATA6 haploinsufficiency causes pancreatic agenesis in humans. Nat Genet. 2011;44:20-2.
68. Flanagan SE, De Franco E, Lango Allen H, Zerah M, Abdul-Rasoul MM, Edge JA, et al. Analysis of transcription factors key for mouse pancreatic development establishes NKX2-2 and MNX1 mutations as causes of neonatal diabetes in man. Cell Metab. 2014;19:146-54.
69. Weedon MN, Cebola I, Patch AM, et al. Recessive mutations in a distal PTF1A enhancer cause isolated pancreatic agenesis. Nat Genet. 2014;46:61-4.
70. Flanagan SE, Haapaniemi E, Russell MA, Caswell R, Allen HL, De Franco E, et al. Activating germline mutations in STAT3 cause early-onset multi-organ autoimmune disease. Nat Genet. 2014;46(8):812-814. doi:10.1038/ng.3040. PMID: 25038750. PMC4129488.
71. 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;386:957-63.
72. 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: Oxford University Press; 2018. p. 214–217.
73. OMIM® Online Mendelian Inheritance in Man. #610199 DIABETES MELLITUS, NEONATAL, WITH CONGENITAL HYPOTHYROIDISM; NDH. Johns Hopkins University; last edited 12/08/2016. https://omim.org/entry/610199. Accessed 28 October 2026.
74. OMIM® Online Mendelian Inheritance in Man. #606367 IMMUNODEFICIENCY 41 WITH LYMPHOPROLIFERATION AND AUTOIMMUNITY; IMD41. Johns Hopkins University; last edited 17/06/2025. https://omim.org/entry/606367. Accessed 28 October 2026.
75. OMIM® Online Mendelian Inheritance in Man. #614700 IMMUNODEFICIENCY, COMMON VARIABLE, 8, WITH AUTOIMMUNITY; CVID8. Johns Hopkins University; last edited 09/05/2017. https://omim.org/entry/614700. Accessed 28 October 2026.
76. OMIM® Online Mendelian Inheritance in Man. #616100 IMMUNE DYSREGULATION WITH AUTOIMMUNITY, IMMUNODEFICIENCY, AND LYMPHOPROLIFERATION; IDAIL. Johns Hopkins University; last edited 29/10/2021. https://omim.org/entry/616100. Accessed 28 October 2026.
77. OMIM® Online Mendelian Inheritance in Man. #618729 LIANG-WANG SYNDROME; LIWAS. Johns Hopkins University; last edited 04/10/2024. https://omim.org/entry/618729. Accessed 28 October 2026.
78. OMIM® Online Mendelian Inheritance in Man. #137920 RENAL CYSTS AND DIABETES SYNDROME; RCAD. Johns Hopkins University; last edited 21/04/2025. https://omim.org/entry/137920. Accessed 28 October 2026.
79. OMIM® Online Mendelian Inheritance in Man. *604612 NK2 HOMEOBOX 2; NKX2-2. Johns Hopkins University; last edited 07/02/2008. https://omim.org/entry/604612. Accessed 28 October 2026.
80. OMIM® Online Mendelian Inheritance in Man. *142994 MOTOR NEURON AND PANCREAS HOMEOBOX 1; MNX1. Johns Hopkins University; last edited 30/07/2022. https://omim.org/entry/142994. Accessed 28 October 2026.
81. OMIM® Online Mendelian Inheritance in Man. *604164 ONE CUT HOMEOBOX 1; ONECUT1. Johns Hopkins University; last edited 24/10/2007. https://omim.org/entry/604164. Accessed 28 October 2026.
82. OMIM® Online Mendelian Inheritance in Man. *606686 EUKARYOTIC TRANSLATION INITIATION FACTOR 2B, SUBUNIT 1; EIF2B1. Johns Hopkins University; last edited 12/06/2025. https://omim.org/entry/606686. Accessed 28 October 2026.
83. OMIM® Online Mendelian Inheritance in Man. *611099 PROTEIN DISULFIDE ISOMERASE, FAMILY A, MEMBER 6; PDIA6. Johns Hopkins University; last edited 18/01/2022. https://omim.org/entry/611099. Accessed 28 October 2026.
84. OMIM® Online Mendelian Inheritance in Man. *620875 FIC DOMAIN-CONTAINING PROTEIN ADENYLYLTRANSFERASE; FICD. Johns Hopkins University; last edited 17/06/2025. https://omim.org/entry/620875. Accessed 28 October 2026.
85. OMIM® Online Mendelian Inheritance in Man. *604917 CCR4-NOT TRANSCRIPTION COMPLEX, SUBUNIT 1; CNOT1. Johns Hopkins University; last edited 29/09/2020. https://omim.org/entry/604917. Accessed 28 October 2026.
86. OMIM® Online Mendelian Inheritance in Man. #610370 DIARRHEA 4, MALABSORPTIVE, CONGENITAL, WITH DIABETES MELLITUS AND COMBINED PITUITARY HORMONE DEFICIENCY; DIAR4. Johns Hopkins University; last edited 16/04/2025. https://omim.org/entry/610370. Accessed 28 October 2026.
87. Hattersley AT, Ashcroft FM. Activating mutations in Kir6.2 and neonatal diabetes: new clinical syndromes, new scientific insights, and new therapy. Diabetes. 2005;54:2503-13.
88. Exeter Genomics Laboratory / University of Exeter. Sulphonylurea transfer in patients with KCNJ11 and ABCC8 mutations – PNDM. DiabetesGenes. Available from: https://www.diabetesgenes.org/about-neonatal-diabetes/su-transfer-in-patients-with-kcnj11-and-abcc8-mutations-pndm/
89. Grange DK, Nichols CG, Singh GK, Guilliams K, Aggarwal M. Cantú syndrome. 2014 Oct 2 [updated 2026 Jan 29]. 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/NBK246980/
90. Gloyn AL, Diatloff-Zito C, Edghill EL, Bellanné-Chantelot C, Nivot S, Coutant R, et al. KCNJ11 activating mutations are associated with developmental delay, epilepsy and neonatal diabetes syndrome and other neurological features. Eur J Hum Genet. 2006;14:824-30.
91. Masia R, Koster JC, Tumini S, Chiarelli F, Colombo C, Nichols CG, et al. An ATP-binding mutation (G334D) in KCNJ11 is associated with a sulfonylurea-insensitive form of developmental delay, epilepsy, and neonatal diabetes. Diabetes. 2007;56:328-36.
92. Flanagan SE, Edghill EL, Gloyn AL, Ellard S, Hattersley AT. Mutations in KCNJ11, which encodes Kir6.2, are a common cause of diabetes diagnosed in the first 6 months of life, with the phenotype determined by genotype. Diabetologia. 2006;49:1190-7.
93. Flanagan SE, Patch AM, Mackay DJ, Edghill EL, Gloyn AL, Robinson D, et al. Mutations in ATP-sensitive K+ channel genes cause transient neonatal diabetes and permanent diabetes in childhood or adulthood. Diabetes. 2007;56:1930-7.
94. Patch AM, Flanagan SE, Boustred C, Hattersley AT, Ellard S. Mutations in the ABCC8 gene encoding the SUR1 subunit of the KATP channel cause transient neonatal diabetes, permanent neonatal diabetes or permanent diabetes diagnosed outside the neonatal period. Diabetes Obes Metab. 2007;9 Suppl 2:28-39.
95. Exeter Genomics Laboratory / University of Exeter. Sulphonylurea transfer in patients with KCNJ11 and ABCC8 mutations – TNDM. DiabetesGenes. Available from: https://www.diabetesgenes.org/about-neonatal-diabetes/transferring-patients-who-have-a-mutation-in-kcnj11-or-abcc8/
96. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 30–31.
97. Docherty LE, Kabwama S, Lehmann A, Hawke E, Harrison L, Flanagan SE, et al. Clinical presentation of 6q24 transient neonatal diabetes mellitus (6q24 TNDM) and genotype-phenotype correlation in an international cohort of patients. Diabetologia. 2013;56:758-62.
98. Orphanet. Paternal uniparental disomy of chromosome 6 syndrome (ORPHA:96191). https://www.orpha.net/en/disease/detail/96191
99. 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. Ferramenta: https://variantvalidator.org
100. National Center for Biotechnology Information. ClinVar. Available from: https://www.ncbi.nlm.nih.gov/clinvar/ (VCV000008666, 008669, 008679, 008680, 021193, 021196, 021200, 021201, 009102, 2574157)
101. Nicolino M, Claiborn KC, Senée V, Boland A, Stoffers DA, Julier C. A novel hypomorphic PDX1 mutation responsible for permanent neonatal diabetes with subclinical exocrine deficiency. Diabetes. 2010;59:733-40.
102. OMIM® Online Mendelian Inheritance in Man. #615935 PANCREATIC AGENESIS 2; PAGEN2. Johns Hopkins University; last edited 22/10/2024. https://omim.org/entry/615935. Accessed 28 October 2026.
103. De Franco E, Shaw-Smith C, Flanagan SE, et al. GATA6 mutations cause a broad phenotypic spectrum of diabetes from pancreatic agenesis to adult-onset diabetes without exocrine insufficiency. Diabetes. 2013;62:993-7.
104. Yorifuji T, Kawakita R, Hosokawa Y, Fujimaru R, Yamaguchi E, Tamagawa N. Dominantly inherited diabetes mellitus caused by GATA6 haploinsufficiency: variable intrafamilial presentation. J Med Genet. 2012;49:642-3.
105. 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;63:2888-94.
106. 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;23:1744-8.
107. Rubio-Cabezas O, Minton JA, Kantor I, Williams D, Ellard S, Hattersley AT. Homozygous mutations in NEUROD1 are responsible for a novel syndrome of permanent neonatal diabetes and neurological abnormalities. Diabetes. 2010;59:2326-31.
108. Philippi A, Heller S, Costa IG, Senée V, Breunig M, Li Z, et al. Mutations and variants of ONECUT1 in diabetes. Nat Med. 2021;27:1928-40.
109. Kołbuc M, Bednarek P, Motyka R, Jarmoliński T, Michalak-Kloc M, Beck BB, et al. Transient neonatal diabetes mellitus as an early diagnostic clue to HNF1B-related disease - two case reports and a literature review. Mol Cell Pediatr. 2026;13(1). doi:10.1186/s40348-026-00234-3. PMID: 41986760. PMC13083689.
110. 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, section 44.1.3.
111. 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;130:6338-53.
112. De Franco E, Caswell R, Johnson MB, Wakeling MN, Zung A, Dũng VC, et al. De novo mutations in EIF2B1 affecting eIF2 signaling cause neonatal/early-onset diabetes and transient hepatic dysfunction. Diabetes. 2020;69:477-83.
113. Wildin RS, Ramsdell F, Peake J, et al. X-linked neonatal diabetes mellitus, enteropathy and endocrinopathy syndrome is the human equivalent of mouse scurfy. Nat Genet. 2001;27:18-20.
114. National Center for Biotechnology Information. MedGen C4014795: STAT3-related early-onset multisystem autoimmune disease (OMIM 615952). Available from: https://www.ncbi.nlm.nih.gov/medgen/C4014795
115. OMIM® Online Mendelian Inheritance in Man. #615952 AUTOIMMUNE DISEASE, MULTISYSTEM, INFANTILE-ONSET, 1; ADMIO1. Johns Hopkins University; last edited 26/06/2025. https://omim.org/entry/615952. Accessed 28 October 2026.
116. OMIM® Online Mendelian Inheritance in Man. #147060 HYPER-IgE SYNDROME 1, AUTOSOMAL DOMINANT, WITH RECURRENT INFECTIONS; HIES1. Johns Hopkins University; last edited 06/10/2023. https://omim.org/entry/147060. Accessed 28 October 2026.
117. Saarimäki-Vire J, Balboa D, Russell MA, Saarikettu J, Kinnunen M, Keskitalo S, et al. An activating STAT3 mutation causes neonatal diabetes through premature induction of pancreatic differentiation. Cell Rep. 2017;19:281-94.
118. Johnson MB, De Franco E, Lango Allen H, Al Senani A, Elbarbary N, Siklar Z, et al. Recessively inherited LRBA mutations cause autoimmunity presenting as neonatal diabetes. Diabetes. 2017;66:2316-22.
119. Santer R, Klepper J. Disorders of glucose transport (8.5 Fanconi-Bickel Syndrome). In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 8, section 8.5.
120. Sansbury FH, Flanagan SE, Houghton JA, Shuixian Shen FL, Al-Senani AM, Habeb AM, et al. SLC2A2 mutations can cause neonatal diabetes, suggesting GLUT2 may have a role in human insulin secretion. Diabetologia. 2012;55:2381-5.
121. OMIM® Online Mendelian Inheritance in Man. #618434 DEAFNESS, AUTOSOMAL RECESSIVE 94; DFNB94. Johns Hopkins University; last edited 23/01/2020. https://omim.org/entry/618434. Accessed 28 October 2026.
122. OMIM® Online Mendelian Inheritance in Man. #616239 COMBINED OXIDATIVE PHOSPHORYLATION DEFICIENCY 24; COXPD24. Johns Hopkins University; last edited 04/06/2019. https://omim.org/entry/616239. Accessed 28 October 2026.
123. Frota M, Pereira SMC, Ribeiro EM. Síndrome de Pearson. In: Ribeiro EM, coord. Síndromes Genéticas: Reconhecer para Tratar. Fortaleza: EdUnichristus; 2021. p. 125–126.
124. Williams TB, Daniels M, Puthenveetil G, Chang R, Wang RY, Abdenur JE. Pearson syndrome: unique endocrine manifestations including neonatal diabetes and adrenal insufficiency. Mol Genet Metab. 2012;106:104-7.
125. Poonooru R, Park KE, Schmelzle A, Telugu BP. Functional Inactivation of PAX4 Results in Disrupted Endocrine Pancreas Development and Neonatal Diabetes in Pigs. Int J Mol Sci. 2026;27(13). doi:10.3390/ijms27135651. PMID: 42449928. PMC13361469.
126. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 6, Congenital Hyperinsulinism and Genetic Disorders of Insulin Resistance and Signalling.
127. OMIM® Online Mendelian Inheritance in Man. #240800 HYPOGLYCEMIA, LEUCINE-INDUCED; LIH. Johns Hopkins University; last edited 22/11/2024. https://omim.org/entry/240800. Accessed 28 October 2026.
128. OMIM® Online Mendelian Inheritance in Man. #616214 HYPERPROINSULINEMIA. Johns Hopkins University; last edited 06/02/2015. https://omim.org/entry/616214. Accessed 28 October 2026.
129. OMIM® Online Mendelian Inheritance in Man. #176450 CURRARINO SYNDROME. Johns Hopkins University; last edited 07/07/2023. https://omim.org/entry/176450. Accessed 28 October 2026.
130. OMIM® Online Mendelian Inheritance in Man. #603896 LEUKOENCEPHALOPATHY WITH VANISHING WHITE MATTER 1; VWM1. Johns Hopkins University; last edited 17/09/2025. https://omim.org/entry/603896. Accessed 28 October 2026.
131. OMIM® Online Mendelian Inheritance in Man. *612803 ASPARAGINYL-tRNA SYNTHETASE 2; NARS2. Johns Hopkins University; last edited 29/05/2019. https://omim.org/entry/612803. Accessed 28 October 2026.
132. Yorifuji T, Matsubara K, Sakakibara A, et al. Abnormalities in chromosome 6q24 as a cause of early-onset, non-obese, non-autoimmune diabetes mellitus without history of neonatal diabetes. Diabet Med. 2015.
133. Exeter Genomics Laboratory / University of Exeter. KCNJ11 and ABCC8 neonatal diabetes: effects on the brain. DiabetesGenes. Available from: https://www.diabetesgenes.org/about-neonatal-diabetes/effects-of-sulphonylurea-on-the-brain/
134. 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.
135. Auerbach A, Cohen A, Ofek Shlomai N, Weinberg-Shukron A, Gulsuner S, King MC, et al. NKX2-2 mutation causes congenital diabetes and infantile obesity with paradoxical glucose-induced ghrelin secretion. J Clin Endocrinol Metab. 2020;105:dgaa563.
136. 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.
137. Kumar R, Dayal D, George A, Banerjee S, Yadav J, Govier M, et al. Neonatal Diabetes: 20-Year Experience from a Tertiary Care Pediatric Diabetes Clinic in North India. Horm Res Paediatr. 2025:1-12. doi:10.1159/000548390. PMID: 40924646.
138. Shoucri BM, Thambundit A, Chia DJ. 6q24-Related Transient Neonatal Diabetes Mellitus Presenting With Severe Diabetic Ketoacidosis and Multiorgan Failure. JCEM Case Rep. 2025;3(10):luaf145. doi:10.1210/jcemcr/luaf145. PMID: 40979819. PMC12448923.
139. Carmody D, Pastore AN, Landmeier KA, Letourneau LR, Martin R, Hwang JL, et al. Patients with KCNJ11-related diabetes frequently have neuropsychological impairments compared with sibling controls. Diabet Med. 2016;33:1380-6.
140. Landmeier KA, Lanning M, Carmody D, Greeley SAW, Msall ME. ADHD, learning difficulties and sleep disturbances associated with KCNJ11-related neonatal diabetes. Pediatr Diabetes. 2017;18:518-23.
141. Desai JM, Letourneau-Freiberg LR, Wroblewski KE, Scott MN, Msall ME, Greeley SAW. Anxiety, social responsiveness, and grit among patients with KCNJ11-related neonatal diabetes compared to unaffected siblings. Acta Diabetol. 2026;63(4):621-627. doi:10.1007/s00592-025-02598-w. PMID: 41591478. PMC13092531.
142. Ben Rhaiem I, Beltrand J, Vermillac G, Kariyawasam D, Geraud D, Besançon A, et al. Safety and Effectiveness of Oral Glyburide Suspension in Neonatal Diabetes Mellitus: French Retrospective Cohort Study. J Endocr Soc. 2025;9(7):bvaf083. doi:10.1210/jendso/bvaf083. PMID: 40469835. PMC12134890.
143. Gonzalez BJ, Juanes M, Touzon MS, Alonso C, Obregon Gomez LR, Moreno N, et al. KCNJ11-associated infantile epileptic spasms syndrome and neonatal diabetes. Epileptic Disord. 2026;28(4):1290-1293. doi:10.1002/epd2.70228. PMID: 41874548. PMC13499289.
144. Bowman P, Mathews F, Barbetti F, et al. Long-term follow-up of glycemic and neurological outcomes in an international series of patients with sulfonylurea-treated ABCC8 permanent neonatal diabetes. Diabetes Care. 2021;44:35-42.
145. Bowman P, Sulen Å, Barbetti F, Beltrand J, Svalastoga P, Codner E, et al. Effectiveness and safety of long-term treatment with sulfonylureas in patients with neonatal diabetes due to KCNJ11 mutations: an international cohort study. Lancet Diabetes Endocrinol. 2018;6:637-46.
146. Warncke K, Eckert A, Kapellen T, Kummer S, Raile K, Dunstheimer D, et al. Clinical presentation and long-term outcome of patients with KCNJ11/ABCC8 variants: neonatal diabetes or MODY in the DPV registry from Germany and Austria. Pediatr Diabetes. 2022;23:999-1008.
147. McCullough ME, Letourneau-Freiberg LR, Bowden TL, Kandasamy B, Ray A, Wroblewski K, et al. Clinical characteristics and remission monitoring of 6q24-related transient neonatal diabetes. Pediatr Diabetes. 2024;2024:3624339.
148. Pietrusiński M, Grzybowska-Adamowicz J, Płoszaj T, Skoczylas S, Borowiec M, Piekarska K, et al. The Clinical and Diagnostic Characterization of 6q24-Related Transient Neonatal Diabetes Mellitus: A Polish Pediatric Cohort Study. Biomedicines. 2025;13(10). doi:10.3390/biomedicines13102492. PMID: 41153775. PMC12561608.
149. Flanagan SE, Mackay DJ, Greeley SA, McDonald TJ, Mericq V, Hassing J, et al. Hypoglycaemia following diabetes remission in patients with 6q24 methylation defects: expanding the clinical phenotype. Diabetologia. 2013;56:218-21.
150. Welters A, Meissner T, Konrad K, Freiberg C, Warncke K, Judmaier S, et al. Diabetes management in Wolcott-Rallison syndrome: analysis from the German/Austrian DPV database. Orphanet J Rare Dis. 2020;15:100.
151. Sutton H, Saban OS, Cunningham J, Bandsma R, Sayed B, Ng V. Phenotype and Long-Term Outcome in Recurrent Paediatric Acute Liver Failure: Systematic Review and Individual Participant Data Analysis. Liver Int. 2025;45(11):e70412. doi:10.1111/liv.70412. PMID: 41159810. PMC12571025.
152. OMIM® Online Mendelian Inheritance in Man. #609069 PANCREATIC AND CEREBELLAR AGENESIS; PACA. Johns Hopkins University; last edited 03/11/2023. https://omim.org/entry/609069. Accessed 28 October 2026.
153. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 266–269.
154. Poulton CJ, Schot R, Kia SK, et al. Microcephaly with simplified gyration, epilepsy, and infantile diabetes linked to inappropriate apoptosis of neural progenitors. Am J Hum Genet. 2011;89:265-76.
155. American Diabetes Association Professional Practice Committee for Diabetes. 14. Children and Adolescents: Standards of Care in Diabetes—2026. Diabetes Care. 2026;49(Suppl 1):S297–S320. doi:10.2337/dc26-S014. PMID: 41358890. PMC12690182.
156. Ghaben AL, Byer-Mendoza C, McNamara K, Huber A, Cymbaluk AL, Singh P. Implementation of a Hybrid Closed-Loop Automated Insulin Delivery System in a 29-Day Infant With Neonatal Diabetes. JCEM Case Rep. 2026;4(3):luaf338. doi:10.1210/jcemcr/luaf338. PMID: 41743174. PMC12930193.
157. Edghill EL, Gloyn AL, Goriely A, Harries LW, Flanagan SE, Rankin J, et al. Origin of de novo KCNJ11 mutations and risk of neonatal diabetes for subsequent siblings. J Clin Endocrinol Metab. 2007;92:1773-7.
158. Nalluri BT, Kudugunti N, Sahay RK, Santhoju V. Neonatal Diabetes Mellitus in a Resource-Limited Setting: Clinical Spectrum, Genetic Heterogeneity, Impact of Malnutrition, and Outcomes with Simplified Basal Insulin Therapy. Horm Res Paediatr. 2026:1-9. doi:10.1159/000553022. PMID: 42284239.
159. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 1, Clinical Approach to Inborn Errors of Metabolism in Paediatrics, section 1.2.2.
160. Jesus E Sá C, Picão de Carvalho C, Ribeiro Martins R, Martins S, Henriques AR, Robalo B, et al. Early-onset transient neonatal diabetes: safe use of continuous glucose monitoring and sensor-integrated insulin pump therapy. Acta Diabetol. 2026;63(10):1819-1823. doi:10.1007/s00592-026-02789-z. PMID: 42658247.
161. European Medicines Agency. EPAR – Product Information: Lantus, Levemir, Tresiba, NovoRapid, Humalog, Fiasp (section 4.2 paediatric population; 6.6 dilution). Available from: https://www.ema.europa.eu/en/documents/product-information/-epar-product-information_en.pdf
162. U.S. Food and Drug Administration. Prescribing information: LANTUS (insulin glargine) BLA 021081, label 2023-07-22; TRESIBA (insulin degludec) BLA 203314, label 2022-07-01; HUMALOG (insulin lispro) BLA 020563, label 2024-01-24; NOVOLOG (insulin aspart) BLA 020986, label 2023-12-15. Available from: https://dailymed.nlm.nih.gov/
163. Thurber BW, Carmody D, Tadie EC, Pastore AN, Dickens JT, Wroblewski KE, et al. Age at time of sulfonylurea initiation influences treatment outcomes in KCNJ11-related neonatal diabetes. Diabetologia. 2015;58:1430-5.
164. Babiker T, Vedovato N, Patel K, Thomas N, Finn R, Männikkö R, et al. Successful transfer to sulfonylureas in KCNJ11 neonatal diabetes is determined by the mutation and duration of diabetes. Diabetologia. 2016;59:1162-6.
165. Exeter Genomics Laboratory, University of Exeter (Hattersley AT). Transferring patients with permanent neonatal diabetes (PNDM) subtypes KCNJ11 or ABCC8 from insulin to sulfonylureas, with Appendix 1: KCNJ11 and ABCC8 neonatal diabetes — effects on the brain. Protocol of 9 June 2023. Available from: https://www.diabetesgenes.org/download/4167/
166. Naylor RN, Patel KA, Kettunen JLT, Männistö JME, Støy J, Beltrand J, et al. Precision treatment of beta-cell monogenic diabetes: a systematic review. Commun Med (Lond). 2024;4:145.
167. Ohkuma T, Aotani R, Sakamoto W, Oshiro A, Ago T. Five-year follow-up of incretin-based therapy in a patient with KCNJ11-related permanent neonatal diabetes: Achieving insulin discontinuation and sulfonylurea reduction. J Diabetes Investig. 2026;17(6):1075-1078. doi:10.1111/jdi.70300. PMID: 41914136. PMC13238849.
168. Teva Pharmaceuticals USA. Glyburide tablets USP (1.25, 2.5, 5 mg), prescribing information. NDA 017532 (reference DiaBeta, Sanofi-Aventis; original approval 1984-05-01). Label effective 2021-07-01. DailyMed set ID 50d16f1d-c814-4411-8749-62561321ecce. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=50d16f1d-c814-4411-8749-62561321ecce
169. INFARMED. Glibenclamida Generis 2,5 mg e 5 mg comprimidos — Resumo das Características do Medicamento (aprovado 25-09-2023). Infomed med_id 40785. Available from: https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=40785&tipo_doc=rcm
170. Agência Nacional de Vigilância Sanitária (ANVISA). Dados abertos — Medicamentos registados (DADOS_ABERTOS_MEDICAMENTOS.csv). Available from: https://dados.anvisa.gov.br/dados/DADOS_ABERTOS_MEDICAMENTOS.csv
171. Exeter Genomics Laboratory / University of Exeter. UK access to glibenclamide. DiabetesGenes. Available from: https://www.diabetesgenes.org/about-neonatal-diabetes/uk-access-to-glibenclamide/
172. ClinicalTrials.gov (U.S. National Library of Medicine). Registos NCT00334711, NCT02375828, NCT02624817, NCT02624830, NCT05751525, NCT02921906, NCT03705195, NCT04192292, NCT02072551, NCT07492004, NCT03988764, NCT02829268. Available from: https://clinicaltrials.gov/
173. U.S. Food and Drug Administration. Metformin hydrochloride tablets, USP: prescribing information (ANDA 209882). Label effective 2026-01-01. DailyMed set ID 0178b9dd-bafa-9e81-e063-6294a90a88b0. Available from: https://dailymed.nlm.nih.gov/
174. Julier C, Nicolino M. Wolcott-Rallison syndrome. Orphanet J Rare Dis. 2010;5:29.
175. Brown G, Plecko B. Disorders of thiamine and pyridoxine metabolism (29.1 Disorders of Thiamine Metabolism). In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 29, section 29.1.1.
176. AbbVie. CREON (pancrelipase) delayed-release capsules, prescribing information. BLA 020725 (first approval 2009-04-30). Label effective 2024-02-28. Available from: https://dailymed.nlm.nih.gov/ (openFDA brand_name CREON)
177. INFARMED. Kreon (pancreatina) 10 000 U cápsulas gastrorresistentes — Resumo das Características do Medicamento. Viatris Healthcare. Infomed med_id 4831. Available from: https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=4831&tipo_doc=rcm
178. 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. PMCID: PMC8001676.
179. Günbey C, Özön A, Gönç EN, Alikaşifoğlu A, Karahan S, Kandemir N. Basal serum thyroxine level should guide initial thyroxine replacement dose in neonates with congenital hypothyroidism. J Clin Res Pediatr Endocrinol. 2021;13(3):269-75. doi:10.4274/jcrpe.galenos.2020.2020.0194. PMID: 33374096.
180. AbbVie Inc. Synthroid (levothyroxine sodium) — prescribing information [Internet]. Revised 2024-02-20 (NDA021402). Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=1e11ad30-1041-4520-10b0-8f9d30d30fcc
181. Woon Y, Card C, Holcroft L, Power C, Somers E, McGrath N, et al. Neonatal Diabetes in Ireland over the Past 19 Years: Clinical Presentation, Management, Genetics, and Outcomes. Horm Res Paediatr. 2026:1-12. doi:10.1159/000550794. PMID: 41632727.
182. Orphanet. Thiamine-responsive megaloblastic anemia syndrome (ORPHA:49827). https://www.orpha.net/en/disease/detail/49827
183. Orphanet. Hypoplastic pancreas-intestinal atresia-hypoplastic gallbladder syndrome (ORPHA:293864). https://www.orpha.net/en/disease/detail/293864
184. Orphanet. Primary microcephaly-epilepsy-permanent neonatal diabetes syndrome (ORPHA:306558). https://www.orpha.net/en/disease/detail/306558
185. Orphanet. Fanconi-Bickel syndrome (ORPHA:2088). https://www.orpha.net/en/disease/detail/2088
186. Orphanet. Enteric anendocrinosis (ORPHA:83620). https://www.orpha.net/en/disease/detail/83620
187. Orphanet. Neonatal diabetes-congenital hypothyroidism-congenital glaucoma-hepatic fibrosis-polycystic kidneys syndrome (ORPHA:79118). https://www.orpha.net/en/disease/detail/79118
188. Orphanet. Syndromic autoimmune enteropathy due to LRBA deficiency (ORPHA:445018). Available from: https://www.orpha.net/en/disease/detail/445018
189. Slingerland AS, Shields BM, Flanagan SE, et al. Referral rates for diagnostic testing support an incidence of permanent neonatal diabetes in three European countries of at least 1 in 260,000 live births. Diabetologia. 2009.
190. Stanik J, Gasperikova D, Paskova M, Barak L, Javorkova J, Jancova E, et al. Prevalence of permanent neonatal diabetes in Slovakia and successful replacement of insulin with sulfonylurea therapy in KCNJ11 and ABCC8 mutation carriers. J Clin Endocrinol Metab. 2007;92:1276-82.
191. Kanakatti Shankar R, Pihoker C, Dolan LM, Standiford D, Badaru A, Dabelea D, et al. Permanent neonatal diabetes mellitus: prevalence and genetic diagnosis in the SEARCH for Diabetes in Youth Study. Pediatr Diabetes. 2013;14:174-80.
192. Wiedemann B, Schober E, Waldhoer T, Koehle J, Flanagan SE, Mackay DJG, et al. Incidence of neonatal diabetes in Austria – calculation based on the Austrian Diabetes Register. Pediatr Diabetes. 2010;11:18-23.