Monogenic Diabetes (MODY) - Approach and HNF1A, GCK and HNF4A MODY

A group of non-autoimmune forms of diabetes caused by heterozygous variants with dominant effects in pancreatic beta-cell genes; HNF1A-MODY as the core entity, alongside GCK-MODY, HNF4A-MODY, MODY caused by ATP-sensitive potassium channel genes (ABCC8, KCNJ11) and INS-MODY.

Literature and regulatory information updated: 10 October 2026.

0. Initial Reference Table

In the table, OGTT denotes the oral glucose tolerance test, HbA1c glycated haemoglobin, MLPA multiplex ligation-dependent probe amplification, and MANE Select the reference transcript agreed between NCBI and EMBL-EBI (Matched Annotation from NCBI and EMBL-EBI). HNF1B-related diabetes (MODY5, cystic kidney disease and diabetes), congenital hyperinsulinism, mitochondrial diabetes with deafness (MIDD), neonatal diabetes and Wolfram syndrome appear only as differential diagnoses.

1. Introduction

1.1. From “mild familial diabetes” to the MODY concept

Monogenic diabetes is an expanding category encompassing isolated diabetes and diabetes within multisystemic genetic disorders, and must be distinguished from the much more frequent type 1 diabetes (T1D) and type 2 diabetes (T2D) [17]. The main subgroups of monogenic beta-cell disorders are neonatal diabetes, maturity-onset diabetes of the young (maturity-onset diabetes of the young, MODY) and multisystemic syndromes [17]. MODY is the main cause of familial monogenic diabetes: it encompasses autosomal dominant disorders usually presenting with diabetes or hyperglycaemia before age 25, an age frequently exceeded in older generations [17].

Clinical description preceded gene identification by almost two decades. In 1974, Tattersall described mild familial diabetes with dominant inheritance [18]. The following year, Tattersall and Fajans compared 26 families of young people with “maturity-type” diabetes with 35 families with classical juvenile diabetes: in the former, 22 of 26 probands (85%) had a diabetic parent, 46% of families showed vertical transmission over three generations and 25 of 47 tested siblings (53%) had latent diabetes, with a consistent phenotype and most not requiring insulin; in families with juvenile diabetes, only 4 of 35 probands (11%) had a diabetic parent and 8 of 74 siblings (11%) were diabetic [19]. The authors considered the data consistent with autosomal dominant inheritance without excluding a multifactorial explanation [19]. Clinical resemblance to adult non-insulin-dependent diabetes, combined with young age, gave rise to the acronym [10].

1.2. Milestones in molecular discovery

The first locus was glucokinase. In 1992, Froguel and colleagues demonstrated tight linkage between the GCK locus on chromosome 7p and early-onset non-insulin-dependent diabetes in 16 French families, with significant heterogeneity and an estimated 45–95% of families linked to GCK [20]. That same year, Vionnet and colleagues described the first mutation, a nonsense GCK variant segregating with early diabetes in a French family, providing the first evidence of a mutation in a glucose metabolism gene in the pathogenesis of non-insulin-dependent diabetes [21]. In 1995, Vaxillaire and colleagues located another MODY locus on chromosome 12q, in an interval of around 7 cM between D12S86 and D12S342, in families not linked to GCK or 20q, with around half of the families studied showing linkage and a phenotype of marked hyperglycaemia with a severe secretory defect [22].

In 1996, two consecutive articles in Nature identified the transcription factor genes. Yamagata and colleagues showed that MODY3 results from mutations in the hepatocyte nuclear factor 1-alpha gene, then called TCF1 and now HNF1A, a transcription factor regulating hepatic genes and transactivating the insulin gene; the article estimated that MODY accounted for 2–5% of non-insulin-dependent diabetes, with autosomal dominant inheritance and onset by age 25 [23]. The second article attributed MODY1 to mutations in the hepatocyte nuclear factor 4-alpha gene (HNF4A, then TCF14), a member of the nuclear receptor superfamily and an upstream regulator of HNF1A expression, and described prediabetic individuals with normal insulin sensitivity and impaired glucose-stimulated secretion [24]. The list was subsequently expanded to PDX1 (MODY4), HNF1B (MODY5), NEUROD1 (MODY6), KLF11, CEL, PAX4, INS, BLK, ABCC8, KCNJ11 and APPL1, reaching a total of 14 proposed genes [1].

1.3. Numerical nomenclature and its revision

Historical numbering assigns MODY1 to HNF4A, MODY2 to GCK, MODY3 to HNF1A, MODY4 to PDX1, MODY5 to HNF1B, MODY6 to NEUROD1, MODY7 to KLF11, MODY8 to CEL, MODY9 to PAX4, MODY10 to INS, MODY11 to BLK, MODY12 to ABCC8, MODY13 to KCNJ11 and MODY14 to APPL1 [1]. Some teaching manuals, including a Brazilian genetic syndromes textbook, still present only subtypes 1 to 6 and state that most cases result from “six specific genes”, wording reflecting knowledge before large-scale sequencing [25]. Several numbered genes were associated with MODY before large population databases existed, and their validity is now contested: variants in BLK, KLF11, NEUROD1, PAX4 and PDX1 classified as pathogenic in the Human Gene Mutation Database (HGMD) appear in the Genome Aggregation Database (gnomAD) at frequencies incompatible with a rare monogenic disorder [1], and the Clinical Genome Resource (ClinGen) gene curation panel has refuted BLK, KLF11 and PAX4 as causes of monogenic diabetes [2].

For these reasons, the systematic review by the American Diabetes Association (ADA) and European Association for the Study of Diabetes (EASD) Precision Medicine in Diabetes Initiative (Precision Medicine in Diabetes Initiative, PDMI) recommends that reports and clinical communication use the gene name associated with the phenotype, for example “GCK-related hyperglycaemia” or “HNF1A diabetes”, rather than the generic term MODY or numbering [2]. Historical numbering is retained in parentheses when it facilitates reading the literature.

1.4. Definition of the entities within scope

MODY is defined as non-autoimmune, autosomal dominantly inherited diabetes caused by impaired insulin secretion with minimal or no impairment of insulin action, classically presenting before age 25, although it may arise later [1, 11]. The three main genes, GCK, HNF1A and HNF4A, account for most cases, and ADA recognises at least 14 genes with a MODY-like phenotype [11]. Patients with MODY are not insulin-dependent but are frequently treated with insulin because HNF1A and HNF4A forms present with marked symptomatic hyperglycaemia in lean young adults and are mistaken for T1D [17].

HNF1A-related diabetes (MODY3) is progressive diabetes due to a beta-cell secretory defect, with marked sulfonylurea sensitivity, glycosuria due to a low renal threshold, and a risk of microvascular and macrovascular complications dependent on glycaemic control [10, 11]. GCK-related hyperglycaemia (MODY2) is mild fasting hyperglycaemia, stable from birth and non-progressive, reflecting a shifted glucose set point rather than beta-cell failure and requiring no treatment outside pregnancy [1, 10]. HNF4A-related diabetes (MODY1) shares the progressive secretory defect and sulfonylurea sensitivity of the HNF1A form but is associated with macrosomia and neonatal hyperinsulinaemic hypoglycaemia [10, 12]. ABCC8- and KCNJ11-related MODY (MODY12 and MODY13) corresponds to mildly activating variants in the ATP-sensitive potassium channel (K_ATP), with a phenotype similar to the HNF1A and HNF4A forms and a sulfonylurea response [1, 13]. INS-related MODY (MODY10) lies at the less severe end of a spectrum of insulin gene disorders including permanent neonatal diabetes [1, 10].

2. Molecular Genetics and Pathophysiology

2.1. The beta cell as a glucose sensor and the genetic architecture of MODY

Glucose-stimulated insulin secretion depends on a chain of events in which each link corresponds to a monogenic diabetes gene. Glucokinase phosphorylates glucose to glucose-6-phosphate at the rate-limiting step of beta-cell glycolysis; the increase in ATP closes the K_ATP channel, formed by four Kir6.2 subunits (encoded by KCNJ11) and four regulatory SUR1 subunits (encoded by ABCC8); depolarisation opens voltage-dependent calcium channels and triggers insulin granule exocytosis [10, 26]. The transcription factors HNF1A and HNF4A control expression of genes in this pathway and beta-cell development, and INS encodes preproinsulin, processed into proinsulin and then insulin (A and B chains joined by two disulphide bridges) and C-peptide [10]. All known MODY subtypes result from heterozygous variants with dominant effects in beta-cell development or function genes, may arise de novo without a family history, and produce impaired insulin secretion with minimal or no impairment of insulin action [10].

The same proteins produce opposite phenotypes depending on the direction of the functional effect: inactivating GCK variants and activating K_ATP channel variants cause diabetes, whereas activating GCK variants and inactivating K_ATP channel variants cause hyperinsulinism [26]. Gene dosage adds another axis: a heterozygous inactivating GCK variant causes mild hyperglycaemia, whereas two inactivating variants cause permanent neonatal diabetes; similarly, a heterozygous PDX1 variant is associated with MODY4 and two variants with pancreatic agenesis [1, 5].

Table 2.1 brings together the 14 proposed genes, with their estimated proportion of MODY, distinguishing features and frequency of microvascular complications described in GeneReviews.

Table 2.1. Genes proposed as causes of MODY (adapted from GeneReviews) [1]

The text of the same GeneReviews chapter attributes 30–60% of MODY each to GCK and HNF1A, whereas the table gives 30–50% and 30–65%; the difference reflects dependence on the population studied and ascertainment bias, with a greater proportion of GCK in paediatric series from the US, Germany, Italy, France and Spain [1]. HNF4A and HNF1B forms together account for around 10% and the remaining ten genes around 20%, some of which may have been incorrectly associated with MODY [1]. ISPAD (International Society for Pediatric and Adolescent Diabetes) summarises the most clinically relevant subtypes in Table 2.2.

Table 2.2. Most clinically relevant MODY subtypes according to ISPAD 2022 [10]

2.2. HNF1A: structure, function and disease mechanism

HNF1A encodes hepatocyte nuclear factor 1-alpha, a homodimeric transcriptional activator directing tissue-specific expression in pancreatic islets, liver, kidney and intestine [23, 27]. The protein has a dimerisation domain (codons 1–32), two deoxyribonucleic acid (DNA)-binding subdomains (codons 107–174 and 201–280) and a transactivation domain in the C-terminal region, partly encoded by exon 9 [28]. In beta cells, heterozygous loss of function reduces transcription of genes required for secretion, partly through glucose transporter GLUT2, and produces a progressive secretory defect upstream of the sulfonylurea receptor; as peripheral insulin sensitivity is normal or increased and the K_ATP channel is intact, sulfonylureas bypass the defect and elicit a disproportionate secretory response [27, 29]. In the proximal tubule, reduced HNF1A activity decreases glucose transport and lowers the renal threshold, so postprandial glycosuria may precede significant hyperglycaemia [10]. Glucose tolerance testing shows an impaired incretin effect and inappropriate glucagon responses [10].

Diabetes penetrance in carriers identified in referred families is high: 63% develop diabetes by age 25, 79% by 35 and 96% by 55 [10]. Age at diagnosis depends on variant location, as terminal exon variants (8–10) lead to diagnosis an average of 8 years later than variants in exons 1–6 [10]. Intrauterine exposure to maternal diabetes advances onset by around 12 years: in heterozygous offspring of mothers with diabetes, age at diagnosis was 15.5 ± 5.4 years, versus 27.5 ± 13.1 years when the variant was paternally inherited (p=0.05) [10, 27].

Penetrance nevertheless depends on ascertainment context. In a study combining 1,742 referred probands, 2,194 relatives, a United States health system cohort (Geisinger, n=132,194) and UK Biobank (n=198,748), approximately 1 in 1,500 individuals carried a pathogenic HNF1A, HNF4A or GCK variant, and diabetes penetrance in HNF1A carriers was 32% in the unselected population, 49% in the health system, 86% in relatives and 98% in probands [30]. These values measure different things: probands were selected for having diabetes, and the unselected population includes carriers of variants with milder functional effects; similar relative risk across cohorts points to environmental and other genetic contributions [30]. Polygenic background also modulates phenotype, and in referred HNF1A carriers only the T2D polygenic score was associated with clinical presentation (0.33 standard deviations) [31].

The variant spectrum includes missense, nonsense, frameshift, splicing and promoter variants and exon or whole-gene deletions, the latter accounting for approximately 1.2% of cases [1, 28]. The most frequent variant in northern European populations, present in approximately 20% of HNF1A families, is a single cytosine duplication in a tract of eight cytosines in exon 4, historically designated P291fsinsC; its description is complicated by an adjacent benign synonymous polymorphism (rs56348580, c.864G>C), altering the tract's cytosine count [32]. Not all rare HNF1A variants cause MODY: in a study of 989 Europeans with non-autoimmune diabetes diagnosed before 45 years, 25 rare alleles were found in 29 individuals but only 12 alleles, in 16 probands, were considered pathogenic [33].

Table 2.3. HNF1A, HNF4A and GCK variants mentioned in the article, normalised to the MANE Select transcript (GRCh38) [3, 4]

Correspondence between historical R76W and NM_175914.5:c.187C>T p.(Arg63Trp) was obtained from ClinVar's response to a historical-name search; designation R127W, appearing in some literature for variant p.(Arg114Trp), had no unequivocal ClinVar match, returning p.Arg114Trp and p.Arg112Trp for that search, so this article uses only MANE nomenclature [4]. Descriptions NM_000457.5:c.226C>T and c.379C>T, based on the P1 isoform transcript, were rejected by VariantValidator and should not be used [3].

2.3. GCK: a glucose sensor with a shifted set point

Glucokinase catalyses glucose phosphorylation at the rate-limiting step of beta-cell glycolysis and acts as the sensor setting the threshold for glucose-stimulated insulin secretion [1, 11]. Heterozygous inactivating variants reduce phosphorylation, so beta cells regulate secretion around a higher equilibrium point (set point) and secrete insulin only at higher-than-normal glucose levels [10, 11]. The result is a regulated shift in set point, with preserved secretion and regulation rather than beta-cell failure: fasting glucose is set between 5.5 and 8 mmol/L and remains stable throughout life, with slight age-related deterioration similar to the general population [1, 35].

The GCK disease spectrum is dictated by dosage and direction of functional effect. Two inactivating variants (homozygous or compound heterozygous) cause permanent neonatal diabetes with severe intrauterine growth restriction and insulin requirement from the first days, accounting for 2–3% or less of permanent neonatal diabetes worldwide, more in consanguineous regions [10]. Heterozygous activating variants lower the secretion threshold and cause familial hyperinsulinism, from mild diazoxide-responsive to severe unresponsive forms, in fewer than 1% of non-syndromic hyperinsulinism [36]. Hyperglycaemia penetrance in inactivating-variant carriers is 89–97% in all ascertainment contexts, including unselected population cohorts, unlike HNF1A and HNF4A [30]. Approximately 1.8% of GCK-MODY causes are whole-gene or exon deletions [1]. Polygenic background explains some variability: referred carriers, but not 158 unselected UK Biobank carriers, showed enrichment of HbA1c, fasting glucose and T2D polygenic scores (0.16–0.33 standard deviations), and carriers in the top HbA1c score quintile had a 3- to 6-fold greater risk of HbA1c ≥48 mmol/mol [31].

2.4. HNF4A: isoforms, biphasic effect and specific variants

HNF4A encodes a nuclear receptor in the steroid and thyroid receptor superfamily, acting upstream of HNF1A expression [24]. The gene has two promoters, P1 and P2, producing different isoforms, and variants associated with neonatal hyperinsulinism tend to affect isoforms 7–9 dependent on that promoter [27]. The transcript used by ClinVar and ClinGen expert panel specifications is NM_175914, the P2 isoform, explaining numbering differences from older P1-based literature [4, 37].

HNF4A haploinsufficiency produces a two-stage defect: fetal and neonatal hyperinsulinism with macrosomia, followed by progressively declining secretion and diabetes from adolescence [1, 12]. In a study of 108 individuals from HNF4A families, carriers had median birth weight 790 g above non-carrier relatives, 56% were macrosomic and 8 of 54 carrier newborns had transient hyperinsulinaemic hypoglycaemia versus none of 54 non-carriers [12]. Heterozygous HNF1A variants are approximately 10 times more frequent than HNF4A variants in referral series [10]. In paediatrics, HNF4A diabetes begins at similar ages to HNF1A disease [10], and higher birth weight was associated with lower diabetes penetrance in childhood and early adulthood in a series of 186 carriers [27].

Two variants behave differently. Variant p.(Arg63Trp), known as R76W, affects a direct DNA-contact residue and causes, besides the beta-cell phenotype with neonatal hyperinsulinism and macrosomia, atypical dominant proximal tubulopathy, renotubular Fanconi syndrome type 4, with generalised aminoaciduria, low-molecular-weight proteinuria, glycosuria, hyperphosphaturia and hypouricaemia, plus hypercalciuria with relative hypocalcaemia, nephrocalcinosis, hypermagnesaemia and renal insufficiency [38]. Glycosuria in this syndrome is renal and does not indicate diabetes. Variant p.(Arg114Trp) has reduced penetrance and does not affect birth weight [39].

2.5. ABCC8 and KCNJ11: the ATP-sensitive potassium channel

The beta-cell K_ATP channel is a hetero-octamer of four pore-forming Kir6.2 subunits and four regulatory SUR1 subunits [10]. Sulfonylureas bind SUR1 and close the channel independently of glucose metabolism, while diazoxide binds SUR1 and opens it [26, 40]. Activating variants prevent channel closure in response to ATP, maintain cellular hyperpolarisation and reduce insulin secretion; depending on activation degree, they cause permanent neonatal diabetes, transient neonatal diabetes or, in milder forms, a MODY phenotype [5, 10]. Kir6.2 is also expressed in neurons and muscle, explaining DEND syndrome (developmental delay, epilepsy and neonatal diabetes) and the intermediate form without epilepsy, occurring in approximately 20% of patients with KCNJ11 variants [10].

In neonatal diabetes, approximately 90% of KCNJ11 variants cause permanent disease and approximately 10% transient disease, whereas ABCC8 variants more frequently (approximately 66%) cause the transient form [10]. Activating KCNJ11 variants are always heterozygous and approximately 90% are de novo; ABCC8 can also cause recessive disease [10]. ABCC8-MODY was described by Bowman and colleagues in 2012 in families with HNF1A-like diabetes and sulfonylurea response [13]. In dominantly inherited ABCC8 neonatal diabetes, a parent carrying the same variant may have diabetes labelled T2D, reflecting variable expressivity, and reduced penetrance has been described in ABCC8 and KCNJ11 forms [5]. KCNJ11 variant location partly predicts severity, as shown in Table 2.4.

Table 2.4. Mentioned KCNJ11 variants, normalised to MANE Select transcript NM_000525.4 [3, 5]

The catalogue of variants in both channel genes in hyperinsulinism and diabetes was updated by De Franco and colleagues in 2020 [41]. Outside the pancreas, ABCC8 was classified by ClinGen with moderate evidence as a pulmonary arterial hypertension gene, an association irrelevant to MODY but appearing in multigene panels [42].

2.6. INS: a spectrum determined by the degree of impaired biosynthesis

INS variants produce a continuous spectrum of non-autoimmune diabetes. Dominant heterozygous variants, mostly de novo with autosomal dominant family history in approximately 20% of probands, alter proinsulin folding and disulphide bonds, causing endoplasmic reticulum stress and beta-cell apoptosis; they are the second cause of permanent neonatal diabetes after the K_ATP channel [10]. Recessive biallelic variants abolish biosynthesis and are associated with lower birth weight and earlier onset [5, 10]. INS diabetes may appear after the first year, so ISPAD recommends gene testing in early-onset diabetes without autoantibodies and with a MODY-like phenotype [10]. At the mildest end, MODY10 accounts for fewer than 1% of MODY [1].

INS promoter variants described in permanent and transient neonatal diabetes and “type 1b”-like childhood diabetes (without autoimmunity, with low C-peptide and insulin dependence) use historical nomenclature, c.-218A>C and c.-331C>A/G, outside transcript NM_000207.3 sequence [3, 5]. VariantValidator proposes genomic coordinate NC_000011.10:g.2161368T>G for the first, and returned no compatible normalised description for the second, so it should be cited as in the original publication, explicitly specifying reference transcript and version [3]. The INS locus also contains the multifactorial autoimmune T1D susceptibility association known as IDDM2, so a heterozygous INS variant should not be reported as causing T1D.

2.7. Gene–disease validity and minor genes

Association validity between each gene and monogenic diabetes is assessed using the ClinGen framework (definitive, strong, moderate, limited, disputed or refuted); the ClinGen monogenic diabetes gene curation panel has curated common genes, and PanelApp and Exeter Diabetes Genes lists serve as references for panel content [2]. In an analysis of UK Biobank exomes (14,622 people with diabetes and 185,509 without), pathogenic or likely pathogenic variants according to American College of Medical Genetics and Genomics (ACMG) criteria were “consistent” with high penetrance in GCK, HNF1A, HNF4A, NEUROD1, KCNJ11 and HNF1B, “inconclusive” in ABCC8, KLF11, RFX6, PCBD1, WFS1, INS and PDX1, and “inconsistent” in CEL, BLK, APPL1 and PAX4 [43]. Only genes meeting curated criteria should be considered causal [2]. ISPAD 2022 added new genes for childhood-onset (CNOT1, ONECUT1, YIPF5, EIF2B1, KCNMA1) and later-onset monogenic diabetes (TRMT10A, DNAJC3, KCNK16, DUT) [10].

2.8. Allelic diseases

Table 2.5. Allelic diseases of genes within scope

3. Clinical Spectrum and Manifestations by System

3.1. HNF1A-related diabetes

Natural history without targeted treatment. In childhood and early adolescence, carriers' glucose tolerance is usually normal [1]. Before overt diabetes, progressive beta-cell dysfunction develops, with increased insulin sensitivity and glycosuria due to a low renal threshold [1, 10]. Fasting glucose initially remains normal, but the postprandial or 2-hour OGTT rise is large: ISPAD gives an increment above 80 mg/dL (5 mmol/L), GeneReviews an increment generally above 90 mg/dL (5 mmol/L) between 0 and 120 minutes, and ADA reproduces the value of 90 mg/dL [1, 10, 11]. The values are not entirely equivalent, because 5 mmol/L corresponds to approximately 90 mg/dL and 80 mg/dL to approximately 4.4 mmol/L; the difference lies in ISPAD's presentation and does not reflect new data. Fasting hyperglycaemia and osmotic symptoms develop over time; ketosis is rare, because residual secretion persists for many years [10]. Penetrance in referred families is 63% at 25 years, 79% at 35 and 96% at 55 years, and intrauterine exposure to maternal diabetes advances diagnosis by approximately 12 years [10]. The French PNDS places diagnosis usually around puberty, with a median age of 20 years, and describes, on the basis of a French series, two modes of presentation: acute, resembling T1D but rarely ketotic owing to residual secretion, in about 25% of cases, and resembling atypical T2D because of early onset and the absence of metabolic syndrome in about 75% [7]. In children, MODY presents in most cases after the age of 8 years [7]. Because the HNF1A defect reduces expression of the SGLT2 cotransporter in the proximal tubule and lowers the renal threshold for glucose reabsorption, isolated glycosuria without hyperglycaemia may precede diabetes and serve as an early marker in relatives, including in childhood [7].

Complications. Chronic microvascular and macrovascular complications are frequent and depend on glycaemic control, as in T1D and T2D [1, 10]. In an Irish cohort, timely initiation of a sulfonylurea was associated with a lower rate of microvascular complications than in T1D [10]. Cardiovascular risk is disproportionate to the lipid profile: in 39 British families, with 153 carriers and 241 family controls, cardiovascular death occurred in 66% of deceased carriers and 43% of deceased controls (P=0.02), with hazard ratios (HRs), adjusted for sex and smoking, of 1.9 (95% CI 1.2–2.9) for overall mortality and 2.3 (95% CI 1.3–4.2) for cardiovascular mortality [46]. These patients paradoxically have normal or elevated HDL and low hs-CRP [1, 47]. The French PNDS reports, from the same study, excess cardiovascular mortality relative to the general population expressed as an odds ratio of 2.6, attributed to prolonged chronic hyperglycaemia; this metric differs from the hazard ratio of the same study, and the two values should be neither added nor compared directly [7, 46]. The risk of microangiopathy is considered equivalent to that in T1D and T2D, and retinopathy may be the presenting manifestation of diabetes, given the long, largely asymptomatic phase [7].

Neonatal period and liver. Birth weight in HNF1A carriers does not differ from that of non-carrier siblings: in 85 heterozygotes and 49 family controls, the median difference was 10 g (p=0.86), and in Shepherd's series of 134 newborns, the median was 3,490 g [27, 39]. Diazoxide-responsive congenital hyperinsulinism is rare: it accounted for 3.4% of hyperinsulinism cases at a large paediatric centre, with median birth weight 3,815 g (interquartile range, IQR, 3,317–4,176 g), mostly paternal inheritance, presentation between day one and 3 months and diazoxide requirement up to 7.3 years [27]. In the hyperinsulinism GeneReviews, HNF1A explains 4% of diazoxide-responsive hyperinsulinism with a molecular diagnosis, with mild hypoglycaemia that generally resolves in childhood and diabetes in adolescence or adulthood [36]. Familial hepatic adenomatosis, through somatic loss of the second allele, may occur in adults and children, with or without diabetes, and present with intraperitoneal haemorrhage from rupture [27, 44]; its association with hepatic haemangiomatosis in a patient with HNF1A-MODY was described in 2025 [45]. The French PNDS, drawing on a French series, defines adenomatosis as the presence of more than 10 hepatocellular adenomas and estimates it at 6–10% of carriers of HNF1A variants, with or without diabetes; the adenomas result from biallelic inactivation of the gene in the hepatocyte, through a second acquired variant, which makes HNF1A a tumour suppressor gene, and are characterised by diffuse intratumoural steatosis [7]. The potential complications are haemorrhage and transformation into carcinoma; adenomatosis is a formal contraindication to oestrogen-containing contraception, may worsen in pregnancy, and the choice between close surveillance and preventive surgery is discussed at a multidisciplinary meeting on primary liver tumours [7]. Hepatic adenomatosis has not been described in HNF4A-MODY [7].

The problem of misdiagnosis. The clinical manifestation with the greatest impact is misclassification as T1D, with lifelong insulin, or as T2D [17]. In the United Kingdom, more than 80% of estimated MODY cases lacked molecular confirmation [8], and in the US SEARCH cohort only 3 of the 47 young people with an HNF1A, HNF4A or GCK variant had a clinical diagnosis of MODY, most of them being treated with insulin [48]. According to the French PNDS, which cites British data, 50–80% of MODY diagnoses are made during follow-up of already known diabetes classified as T1D or T2D, with a mean interval of about 10 years between the diagnosis of diabetes and molecular confirmation [7].

3.2. GCK-related hyperglycaemia

GCK-MODY is the most frequent subtype of monogenic diabetes in paediatric clinics and has a homogeneous phenotype [10]. Fasting glucose lies between 5.5 and 8.0 mmol/L (99–144 mg/dL) from birth, HbA1c between 38 and 56 mmol/mol (5.6–7.3%) up to 40 years and between 41 and 60 mmol/mol (5.9–7.6%) above 40 years, and the OGTT increment is small [1, 17]. ISPAD describes HbA1c usually below 7.5% (59 mmol/mol) and an OGTT increment usually below 60 mg/dL (3.5 mmol/L), emphasising that the OGTT is variable and the value is not an absolute criterion; ADA uses a 2-hour increment threshold below 54 mg/dL (3 mmol/L) [10, 11]. An Irish review gives an increment below 4.6 mmol/L after the glucose load [27]. The French PNDS adopts, from the same discrimination study, a fasting glucose of 5.4–8.3 mmol/L and an HbA1c of 5.6–7.6%, with almost complete penetrance from birth, whatever the type or location of the variant [7, 49]. The defect raises the glucose threshold that triggers insulin secretion without abolishing the early secretory peak [7].

Carriers are generally asymptomatic and diagnosis incidental, on routine testing, during pregnancy or through family screening [1]. Incidental hyperglycaemia of 5.5–8 mmol/L (100–145 mg/dL) in an asymptomatic child without autoimmunity is associated with minimal future T1D risk, and in a series of 172 Italian families a significant proportion of these cases had GCK-MODY [10]. The affected parent is often undiagnosed or labelled early T2D, prompting ISPAD to recommend fasting glucose measurement in parents considered unaffected [10]. The GCK variant does not protect against polygenic T2D, whose prevalence in carriers is similar to the general population [10].

Complications are rare. The largest cross-sectional study, conducted in the United Kingdom between 2008 and 2010 in adults aged 35 or older, compared 99 GCK carriers (median age 48.6 years), 91 family controls and 83 patients with young-onset T2D (diagnosis by 45 years), with median HbA1c 6.9%, 5.8% and 7.8%, respectively [50]. Clinically significant microvascular complications occurred in 1% (95% CI 0–5%) of GCK carriers, 2% of controls (P=0.52) and 36% of T2D patients (P<0.001); retinopathy, almost always background, was more frequent in GCK carriers than controls (30% versus 14%; P=0.007), but no carrier required laser versus 28% of T2D patients; microalbuminuria occurred in 1% and neuropathy in 2% of GCK carriers versus 29% in T2D [50]. Macrovascular complications were low, and after 50 years macrovascular risk appears similar to the general population [35, 50].

In pregnancy, GCK-MODY accounts for approximately 2–6% of gestational diabetes in the European series cited by ISPAD and 0.9% (95% CI 0.3–2.3%) of gestational diabetes in the Irish ATLANTIC-DIP cohort [9, 10]. The French PNDS identifies the two most frequent diagnostic settings, non-progressive hyperglycaemia in early childhood and gestational diabetes (about 1% of the latter), particularly in women of normal weight with fasting glucose above 5.5 mmol/L at screening [7]. The effects on fetal growth depend on the fetal genotype and are addressed in section 7.

3.3. HNF4A-related diabetes

The phenotype is biphasic. In the neonatal period, approximately half of carriers are macrosomic and approximately 15% have diazoxide-responsive hyperinsulinaemic hypoglycaemia, remitting in childhood, with diabetes appearing from adolescence [10]. In Pearson and colleagues' study, transient hypoglycaemia within the first 3 days occurred in 15.4% of carriers, probably underestimated, and macrosomia in 56% of carriers versus 13% of non-carriers (p<0.001) [12, 27]. Median corrected birth weight was 4 840 g when the mother was the carrier and 4 170 g when the father was, with macrosomia in 64% and 46% and extreme macrosomia (>5,000 g) in 15% and 7%, respectively, showing that the effect depends mainly on fetal genotype with an additional maternal hyperglycaemia contribution [27, 39]. Hyperinsulinism usually presents in the first week, particularly the first two days, is generally mild but may require intravenous glucose, glucagon and diazoxide or chlorothiazide for months or years, and may persist into early adulthood [27, 39]. HNF4A is the third genetic cause of diazoxide-responsive hyperinsulinism (18.6% of genetically defined cases in one series) and, in GeneReviews, explains 5% of diazoxide-responsive hyperinsulinism with molecular diagnosis [27, 36].

In adulthood, diabetes follows the course of the HNF1A form, with a progressive secretory defect and control-dependent complications [1]. Unlike HNF1A, the lipid profile resembles that of T2D, with reduced apolipoproteins A1 and A2 and HDL and increased LDL [1]. In an Irish cohort of 42 carriers, the prevalence of retinopathy was 24.4%, that of nephropathy 12.9% and that of coronary disease 22.6% [51]. The variant p.(Arg63Trp) adds the proximal tubulopathy described in section 2.4, with nephrocalcinosis and possible renal failure [38]; the French PNDS refers to it by the historical P1 isoform numbering, p.Arg76Trp, and highlights hypercalciuria and nephrocalcinosis [7].

The French PNDS adds three points. The penetrance of diabetes in HNF4A-MODY is of the order of 45% at 25 years and 96% at 60 years, compared with 63% at 25 years and 96% at 55 years in HNF1A-MODY, values derived from referred families and not from carriers identified in populations [7, 30]. Macrosomia occurs in about 50% of carriers and neonatal hypoglycaemia in about 20%, a value higher than the 15% cited by ISPAD, and a 2022 study showed that the penetrance of diabetes correlates inversely with birth weight, which does not support the hypothesis that fetal hyperinsulinism exhausts the beta cell through apoptosis [7, 10]. Some patients go through a sequence of neonatal hypoglycaemia, remission and, later, MODY-type diabetes [7]. The reduction in apolipoprotein A2 and HDL has no evident clinical consequence [7].

3.4. MODY due to ABCC8, KCNJ11 and INS

Clinical data for these forms are scarce. ABCC8- and KCNJ11-MODY resemble HNF1A and HNF4A forms, with sulfonylurea responsiveness, and should be considered when a patient responds to sulfonylureas without an HNF1A or HNF4A variant [1, 13]. ISPAD notes that probands or relatives may have a history of transient neonatal diabetes or neuropsychological difficulties [10]. In the international SWEET registry, diabetic ketoacidosis (DKA) at diagnosis occurred in 38.5% of young people with KCNJ11 variants (n=5), 22.2% with ABCC8 (n=2) and 11.8% with INS, versus 1.2% with GCK (n=3), small absolute numbers not permitting risk estimation [52]. With INS, presentation ranges from neonatal diabetes to non-autoimmune diabetes of childhood or young adulthood, and small insulin doses sometimes suffice [1, 10].

3.5. Birth weight and coexistence with other forms of diabetes

Table 3.1. Birth weight by monogenic diabetes subtype [1, 17, 39]

Obesity may coexist with any MODY form: in a treatment trial of T2D in adolescents with obesity or overweight (TODAY), at least 4.5% had MODY, mainly HNF4A, GCK or HNF1A, according to GeneReviews, and ISPAD reports 3% pathogenic variants among obese young people with presumed T2D [1, 10]. MODY may coexist with insulin resistance, and MODY families may also include members with T1D or T2D, requiring each relative to be interpreted separately [1]. T1D autoantibodies exclude the need for further testing in most cases, but autoantibodies have been reported in monogenic diabetes [11].

4. Clinical Recognition and Warning Signs

4.1. General principle

There is no pathognomonic MODY sign except age below 6 months for neonatal diabetes, and suspicion arises from combined traits making T1D or T2D unlikely [10]. ISPAD recommends considering MODY when diabetes occurs in a parent and that parent's first-degree relatives in someone without T1D or T2D characteristics (grade B); testing GCK with mild, stable, non-progressive fasting hyperglycaemia (grade B); and considering HNF1A first in symptomatic familial diabetes with an autosomal dominant pattern (grade B) [10]. Obesity alone should not prevent testing, particularly when family history strongly suggests dominant inheritance, some affected relatives are not obese or other metabolic syndrome traits are absent (grade C) [10].

4.2. When to suspect the disease, by age and context

Table 4.1. Monogenic diabetes warning signs by age and context

ADA 2026 summarises these in three situations: diabetes in the first 6 months; diabetes without typical T1D or T2D features (negative autoantibodies, no obesity or other metabolic traits), particularly with strong family history; and mild stable fasting hyperglycaemia of 100–150 mg/dL (5.6–8.5 mmol/L), with stable HbA1c of 5.6–7.6% (38–60 mmol/mol), particularly without obesity [11]. Although T1D may be familial, only 2–6% of people with T1D have an affected parent [1]. Familial T2D typically begins after 45 years and is associated with obesity, but outside high-prevalence populations, such as south and south-east Asian populations, it is rarely diagnosed before 40 years in people without obesity [1, 17].

The 2022 French PNDS defines two main criteria for suspicion: adult diabetes with onset before 40 years, without anti-GAD or anti-IA2 antibodies and, in most cases, without overweight or features of metabolic syndrome (hypertension, hypertriglyceridaemia, low HDL, increased waist circumference); and a family history of diabetes in several generations, without overweight or obesity at diagnosis in the patient or relatives [7]. It adds seven settings in which one of the forms should be considered: non-insulin-requiring diabetes diagnosed in childhood; gestational diabetes without risk factors; incidentally discovered glycosuria without evident hyperglycaemia; glycaemic stability long after diagnosis, or moderate fasting hyperglycaemia that does not progress; early-onset diabetes already with severe complications, particularly retinal, at diagnosis; diabetes associated with hepatic adenomatosis; and a personal history of macrosomia or neonatal hypoglycaemia without maternal hyperglycaemia in pregnancy [7]. The initial hyperglycaemia is almost always asymptomatic and discovered in routine tests, on urinalysis, in pregnancy or through family screening; decompensation with polyuria and polydipsia is rare, usually without ketoacidosis, and metabolic control is achieved with limited doses of insulin or antidiabetic drugs for years [7]. The patient with MODY usually has no hepatic steatosis, obstructive sleep apnoea syndrome or acanthosis nigricans, unlike the young person with T2D [7]. The PNDS decision tree (appendix 3) summarises referral for genetic testing (Table 4.2).

Table 4.2. Decision tree of the French PNDS for the genetic diagnosis of MODY [7]

4.3. Targeted history and physical examination

History should record age at diagnosis, presentation, initial and subsequent treatments, glycaemic control, birth weight and gestational age, neonatal hypoglycaemia history, associated autoimmune diseases (thyroid disease, vitiligo, Addison disease, lupus, rheumatoid arthritis) and signs suggestive of mitochondrial disease such as deafness, muscle weakness, stroke-like episodes, ophthalmoplegia or pigmentary retinopathy [17]. Family history should cover three generations and include, for each diabetic relative, onset age, body habitus, insulin independence, previous genetic testing and renal disease [1]. A family with several members with kidney disease and others with diabetes suggests HNF1B [1].

Physical examination includes height, weight and BMI, assessment for acanthosis nigricans (insulin resistance), vitiligo, abnormal fat distribution (generalised hereditary lipodystrophy or involving limbs and buttocks; acquired patchy lipodystrophy of face, shoulders and arms), pseudoacromegaly, hyperandrogenism (hirsutism, acne, polycystic ovary syndrome), ptosis, diabetes complications and syndromic traits [17]. Except for gout findings suggesting HNF1B, no physical finding distinguishes different MODY causes [1].

4.4. Emergency signs and newborn screening

Ketoacidosis at diagnosis is unusual in MODY and, when present in a young person, suggests T1D or K_ATP- or INS-related neonatal diabetes; in SWEET, overall ketoacidosis frequency at monogenic diabetes diagnosis was 3.6% [52]. In infants, persistent hyperglycaemia with severe dehydration, glycosuria, osmotic polyuria and poor weight gain requires immediate insulin and rapid genetic testing because results may enable transition to sulfonylureas [5, 17]. In a macrosomic newborn with hyperinsulinaemic hypoglycaemia without perinatal stress, assessment of young-onset diabetes family history and HNF4A or HNF1A testing guide follow-up [27]. The sources consulted describe no MODY newborn screening programmes, and HbA1c is unsuitable for diabetes diagnosis before 6 months because of fetal haemoglobin [5].

5. Diagnosis: Criteria, Tests and Algorithms

5.1. Criteria and recommendations on whom to test

Diabetes diagnosis follows general criteria: in an at-risk relative, HbA1c ≥48 mmol/mol (6.5%) defines diabetes and 43–47 mmol/mol indicates intermediate hyperglycaemia, to be followed in primary care [17]. Subtype diagnosis is exclusively molecular: no biochemical or functional test confirms MODY, and autoantibodies, C-peptide, OGTT, glycosuria, hs-CRP, probability calculators and genetic scores only select whom to test [2, 10]. ISPAD considers molecular testing sensitive and specific, recommends next-generation sequencing as the methodology of choice except in very specific scenarios such as suspected GCK (grade B), and suggests referral to a monogenic diabetes specialist or clinical genetics unit (grade E) [10].

ADA makes three recommendations in its 2026 Standards of Care: everyone diagnosed with diabetes in the first 6 months, irrespective of current age, should undergo neonatal diabetes genetic testing (2.29a, grade A); children and young adults lacking typical T1D or T2D characteristics and with diabetes in successive generations should undergo MODY genetic testing (2.29b, grade A); and in both cases, consultation with a specialist diabetes genetics centre is recommended (2.29c, grade E) [11]. The PDMI systematic review specified criteria based on demonstrated diagnostic yield (Table 5.1).

Table 5.1. PDMI recommendations on whom to test [2]

Excluding people with obesity increases specificity but reduces sensitivity for GCK [2]. The 2020 ADA/EASD consensus also proposed testing paediatric diabetes with three negative autoantibodies based on Swedish data and emphasised that universal testing is not cost-effective, so selection should combine clinical characteristics and non-genetic tests [15].

5.2. Non-genetic selection tests

Autoantibodies. In T1D, autoantibodies against glutamic acid decarboxylase (anti-GAD) and/or insulinoma-associated antigen 2 (anti-IA2) are present in approximately 80% of northern Europeans at diagnosis and become less frequent after 5–10 years of disease; islet cell antibodies measured on rat pancreas have low sensitivity and little value [17]. ADA recommends starting with anti-GAD and, if negative, measuring anti-IA-2 and/or antibodies against zinc transporter 8 (anti-ZnT8), with anti-insulin useful only before insulin treatment; 5–10% of people with T1D have no antibodies, so that, below 35 years and without features of T2D or monogenic diabetes, a negative result does not change the diagnosis of T1D [11]. The French PNDS requests anti-GAD and anti-IA2 at diagnosis and, in children, also anti-insulin antibodies, and emphasises that, in paediatric studies, the absence of autoantibodies is the factor that most points to MODY, far more discriminating than any clinical criterion; the absence of markers of insulin resistance (normal HDL and triglycerides, normal or low insulinaemia) completes the profile [7].

C-peptide. C-peptide is indicated only in insulin-treated people and may be measured in a random sample, with concurrent glucose, up to 5 hours after a meal [11]. Values ≥600 pmol/L (≥1.8 ng/mL) indicate preserved secretion; values <600 pmol/L with glucose <4 mmol/L (<70 mg/dL) or in the fasting state warrant repetition; values <80 pmol/L (<0.24 ng/mL) indicate severe deficiency with no need for repetition; and values of 200–600 pmol/L (0.6–1.8 ng/mL) are compatible with T1D or MODY, but also with insulin-treated T2D [11]. C-peptide should be measured before stopping insulin, to exclude severe deficiency, and not within the 2 weeks following a hyperglycaemic emergency [11]. In T1D, C-peptide <200 pmol/L indicates absolute deficiency, which may not be observed in the first 5 years [17]. GeneReviews uses a threshold of ≥0.2 nmol/L (≥0.60 ng/mL) with hyperglycaemia, beyond the honeymoon period, as a warning sign [1]. The French PNDS uses the same value, but for stimulated C-peptide: above 200 pmol/L, even after several years of pharmacological treatment, it indicates partially preserved beta-cell function [7].

Abdominal ultrasound. The French PNDS includes abdominal ultrasound in the initial assessment, because it may show hepatic adenomatosis, favouring HNF1A, or renal cysts and urogenital tract malformations, favouring HNF1B [7].

Urinary C-peptide/creatinine ratio (UCPCR). In a urine sample collected 120 minutes after the day's largest meal, preserved with boric acid and posted, UCPCR ≥0.2 nmol/mmol distinguished HNF1A and HNF4A forms from T1D of at least 5 years' duration with 97% sensitivity and 96% specificity (area under the curve, AUC, 0.98), but poorly distinguished them from T2D (AUC 0.64) [54].

High-sensitivity C-reactive protein. hs-CRP is lower in HNF1A-MODY than other diabetes forms [1, 47]. In 750 patients, median was 0.3 mg/L (IQR 0.1–0.6) in HNF1A, versus 1.40 mg/L in T2D, 1.10 mg/L in T1D, 1.45 mg/L in HNF4A and 0.60 mg/L in GCK; a threshold <0.75 mg/L distinguished HNF1A from T2D with 79% sensitivity, 70% specificity and AUC 0.84 [47]. A seven-centre European study confirmed the difference irrespective of assay, with C-statistics between 0.79 and 0.97 depending on centre after excluding values >10 mg/L; hs-CRP rises with infection and inflammation, limiting interpretation [55]. In a third study, hs-CRP had AUC 0.83 (cut-off 0.81 mg/L) and plasma fucosylated glycan GP30 AUC 0.90 (cut-off 0.70%), with an estimated 70–75% reduction in sequencing [33].

Glucose, HbA1c and OGTT. In GCK-MODY, narrow fasting glucose and HbA1c intervals have been formally evaluated as discriminators [49]. OGTT shows a small increment in GCK and large increment in HNF1A, with divergent thresholds described in section 3 [10, 11].

T1D genetic risk score. In 1,129 insulin-treated people referred for MODY testing in the United Kingdom, a 10-polymorphism score included in the panel estimated approximately 20% (95% CI 14.9–25.2%) had T1D; as pretest triage, the score had overall AUC 0.60 but 0.83 below 10 years, an age at which more than half the tests could be avoided without missing MODY [56]. Genetic scores were developed in homogeneous White European populations [1].

5.3. Probability calculators and selection pathways

The Exeter MODY probability calculator was developed by logistic regression in 1,191 patients diagnosed by 35 years (594 MODY, 278 T1D and 319 T2D) and externally validated in 350 [57]. MODY predictors versus T1D were lower HbA1c, a diabetic parent, female sex and later diagnosis; versus T2D, lower BMI, younger diagnosis, female sex, lower HbA1c, a diabetic parent and no oral antidiabetic or insulin treatment [57]. C-statistics were 0.95 versus T1D and 0.98 versus T2D, with 0.95 and 0.94 in external validation, and optimal cut-offs gave 91% sensitivity and 94% specificity versus 72% and 91% with classical criteria (diagnosis before 25 years and affected parent) [57]. The calculator was not validated outside European populations, above 35 years at diagnosis, for other monogenic forms or in insulin-treated people with high T1D probability [2]. In Italy, discrimination was substantially lower than in the United Kingdom [58].

In 2026, the Exeter team updated the model with clinical characteristics and biomarkers (C-peptide and autoantibodies) through Bayesian recalibration: development used 594 monogenic diabetes cases and 597 controls, recalibration used the UNITED paediatric population (n=1,299) and external validation 1,025 people [59]. In early-insulin-treated people, AUC was 0.98 with biomarkers versus 0.80 with clinical data alone and 0.96 with biomarkers alone; in those not treated early with insulin, it was 0.86, and 0.98 and 0.92 in respective external validations [59]. At a probability threshold ≥5%, positive test rate was 16–19%; the model applies to White Europeans diagnosed by 35 years [59]. ADA incorporates this threshold into its adult flowchart, considering monogenic diabetes suggested by probability above 5% in the calculator (https://www.diabetesgenes.org/exeter-diabetes-app/ModyCalculator), HbA1c <58 mmol/mol (<7.5%) at diagnosis, a diabetic parent or traits of a specific monogenic cause [11].

Table 5.2. Selection strategies compared by PDMI [2, 57]

5.4. Diagnostic yield by population

Genetic testing yield depends on selection. Table 5.3 summarises PDMI and subsequent study data.

Table 5.3. Diagnostic yield of genetic testing by population [2, 48, 53, 58, 60, 61]

In the late-onset diabetes study, therapeutically relevant subtypes (GCK, HNF1A, HNF4A, ABCC8, KCNJ11) accounted for 1/234 United Kingdom cases and 1/935 United States cases; strict criteria (BMI <25 kg/m², no insulin, diabetic parent) raised yield to 2.64% and 0.87% but would miss more than 86% of cases [53]. Data from different populations are not directly comparable because selection criteria, testing method and variant classification differ.

5.5. Molecular testing methods and their limitations

Next-generation sequencing has replaced gene-by-gene Sanger sequencing, but cost still warrants careful selection; in limited-resource settings, Sanger sequencing of genes with the greatest therapeutic impact remains acceptable and is appropriate for cascade testing of a familial variant [10]. Rapid targeted testing, such as GCK in a pregnant woman with mild fasting hyperglycaemia, may be appropriate [10]. GeneReviews describes three approaches: sequencing the most likely genes when phenotype is distinctive, followed by deletion/duplication analysis; a multigene panel with deletion/duplication analysis when phenotype does not identify a gene; and exome sequencing, avoiding gene selection and enabling reanalysis [1]. Chromosomal microarray is indicated when phenotype suggests a contiguous deletion such as the ≥1.2 Mb 17q12 deletion including HNF1B [1].

Table 5.4. Yield and limitations by method (PDMI systematic review) [2]

NGS increased yield by approximately 30% versus GCK, HNF1A and HNF4A Sanger sequencing, particularly by identifying rare syndromic forms, and diagnosed cases missed by Sanger because of allelic drop-out caused by primer-binding-site variants [2]. Exome and genome sequencing rarely added yield and are expensive [2]. With grade A, PDMI recommends targeted NGS as the preferred method; panels including all known causes, including mitochondrial, known non-coding variants (promoters, deep introns and distal enhancers) and CNVs, with all recessive genes in consanguineous populations; acceptance of separate MLPA for CNVs or pyrosequencing for m.3243A>G, although more expensive; provision of 6q24 methylation assays by neonatal diabetes services; and rapid initial Sanger of GCK when suspected and KCNJ11, ABCC8 and INS in neonatal diabetes diagnosed by 6–12 months [2]. Integration of mitochondrial DNA structural variant analysis into routine genomic sequencing was described in 2026 as increasing diagnosis without additional cost [62].

Test limitations must be explicit. Exon or whole-gene deletions account for approximately 1.8% of GCK-MODY causes, 1.2% of HNF1A-MODY, 1.9% of HNF4A-MODY and 33% of HNF1B-MODY; these are proportions of variant types rather than clinical sensitivities but show that panels without CNV analysis miss cases [1]. Deep intronic variants, regulatory-region variants and variants in genes absent from the panel escape most panels; m.3243A>G with low blood heteroplasmy may be undetected; and included genes and sensitivity vary between laboratories and over time [1, 2]. Some clinical MODY may reflect unknown genes or undetected complex alterations, so a negative result does not exclude monogenic diabetes [1]. Approximately 80% of insulin-treated people referred for testing in the United Kingdom remain without a genetic diagnosis [56].

In France, the PNDS stipulates that the test be prescribed by a specialist physician, after consent, and performed in a specialised laboratory, which may request a second sample for confirmation, in line with good practice for any molecular test; the list of laboratories is available on the Orphanet portal [7]. The PNDS describes a two-stage analysis, with sequencing of the coding regions, the flanking intronic junctions and, for some genes, the promoters, and a search for deletions of one or more exons or of the whole gene, on an NGS panel that includes the three main genes, genes for rare causes and genes for syndromic diabetes, with possible confirmation by a second technique and segregation analysis in the parents [7]. It attributes about 75% of identified MODY cases to the GCK, HNF1A and HNF4A genes, reports variants in some twenty other genes and regards the variants as private (specific to each family), with the exception of one HNF1A variant present in about 15% of HNF1A-MODY cases [7]. It estimates large deletions at less than 0.1% of cases, a value far lower than the per-gene proportions of 1.2–1.9% in GeneReviews; the difference may reflect different denominators (all tested cases versus cases with a confirmed diagnosis), and CNV analysis remains recommended by both sources [1, 7].

5.6. Variant classification and expert panel specifications

Laboratories should classify variants according to 2015 ACMG/AMP (American College of Medical Genetics and Genomics/Association for Molecular Pathology) standards and ClinGen Sequence Variant Interpretation group updates [2, 10]. The ClinGen Monogenic Diabetes Variant Curation Expert Panel (VCEP 50016) adapted rules for HNF1A, HNF4A and GCK, incorporating unpublished case, phenotype and prevalence data; its classifications, covering hundreds of variants, are freely accessible and recognised by the FDA (Food and Drug Administration) [2, 10]. Current specifications appear in the Criteria Specification registry (CSpec) and were not identified as a published article in an October 2026 search [28, 37, 63].

Table 5.5. Main ClinGen gene-specific ACMG/AMP rule specifications [28, 37, 63]

In approximately half of cases, the rare HNF1A variants found in young-onset diabetes are not causal, which makes the application of these rules decisive [33]. In a Brazilian study of 109 probands, application of the ACMG/AMP standards classified 16% of the 32 variants as variants of uncertain significance (VUS) [61], and reanalysis of GCK VUS has been described as a way of improving the detection of monogenic diabetes [64]. VUS may be reclassified through segregation, biochemical data or functional studies; the interpretation of novel rare variants is difficult, functional studies are slow and poorly standardised, and variant effect maps from multiplexed assays (MAVE) are a future approach [2]. Misclassifications are perpetuated in the literature, and databases contain less information for non-European populations [2]. The French PNDS summarises the three possible outcomes of the analysis: a class 4 or 5 variant (likely pathogenic or pathogenic), which confirms the diagnosis; absence of a relevant variant within the limits of the regions and techniques used, which does not confirm the diagnosis; and a class 3 variant (VUS), which does not allow a conclusion, may justify segregation or functional studies and cannot be used for presymptomatic testing in the family [7]. The interpretation may change with family studies and with advances in knowledge [7].

5.7. Laboratory report content

The report should occupy one page with appendices and restate indication and phenotype; provide a clear headline result without “positive” or “negative”; include a table with HUGO gene name, zygosity, HGVS c. and p. descriptions with transcript and version, genomic coordinates and ACMG/AMP class; note subtype-specific treatment response; propose testing at-risk relatives (diagnostic, predictive or carrier) and specify offspring risk; not report benign or likely benign variants; report VUS according to professional judgement with possible reclassification; and state limitations when not all genes or variant types were covered [2, 65]. Recommendations are to avoid “MODY” in the report and use gene name with phenotype (for example, “GCK-related hyperglycaemia” or “HNF1A diabetes”), exercise caution when reporting syndromic-gene variants in isolated diabetes and participate in the European EMQN (European Molecular Genetics Quality Network) external quality assessment scheme for monogenic diabetes [2]. With a VUS or a predictive testing request in an asymptomatic person, ISPAD advises discussion with a specialist centre [10].

5.8. Prenatal and preimplantation diagnosis

Invasive prenatal diagnosis for MODY is rarely requested, and genotyping the fetus solely to guide obstetric management is not recommended, because of the risk of pregnancy loss from the procedure [1, 17]. The French PNDS states the same rule for maternal GCK-MODY: amniocentesis should not be offered for this indication, but, if performed for another reason, it may be used to test for the maternal variant in fetal cells from the amniotic fluid [7]. Non-invasive fetal genotyping from cell-free fetal DNA in maternal plasma is accredited in the United Kingdom National Health Service for pregnancies with maternal GCK-MODY and is described, with its results, in section 7.6 [66]. In neonatal diabetes, prenatal diagnosis is reserved mainly for syndromic forms that are potentially fatal in childhood, such as Wolcott–Rallison and IPEX syndromes [17]. The sources consulted do not describe the application of preimplantation genetic diagnosis to MODY.

5.9. Practical algorithm

Table 5.6. Investigation algorithm for suspected monogenic diabetes

6. Differential Diagnosis

6.1. Comparative table

Table 6.1. Differential diagnosis of MODY-type monogenic diabetes

6.2. Type 1 and type 2 diabetes

T1D is diagnosed clinically, by insulin dependence with very low C-peptide or anti-GAD or anti-IA2 antibodies [17]. Prevalence is approximately 0.3% in Caucasian people, highest in northern Europe (1–1.5% in Finland), and it is approximately 10 times less frequent in Japan than the USA [17]. Criteria suggesting that “T1D” is not T1D were described in section 4; ISPAD emphasises that, except diagnosis before 6 months, none is pathognomonic [10]. The T1D genetic risk score integrated into the panel estimated approximately 20% of insulin-treated people referred for MODY testing had T1D, reaching 76.2% in children without a genetic diagnosis [56].

T2D has no specific test, and an estimated 1 in 10 Europeans and North Americans will develop it [17]. Variants in more than 60 genes increase risk but explain only approximately 10% of heritability, so the geneticist's role is mainly recognising monogenic forms [17]. Low hs-CRP and BMI, HDL and fasting insulin criteria help separate HNF1A from young-onset T2D (section 5.3) [2, 47].

The French PNDS provides additional figures and clues. GCK-MODY may be mistaken for early T1D in a young person without overweight, and negative antibodies and the stability of the hyperglycaemia resolve the misunderstanding; about 25% of HNF1A and HNF4A forms present as insulinopenic T1D (cardinal syndrome, young age, no overweight), and it is the negative antibodies, persistent residual secretion and dominant transmission that point the way, in a context in which familial cases of T1D are around 10% [7]. GCK-MODY is rarely confused with T2D, but does not protect against later T2D, which should be considered when control deteriorates, particularly if metabolic syndrome develops [7]. The HNF1A and HNF4A forms present as T2D in most cases; a strong family history without obesity, the absence of metabolic syndrome, early onset, severe complications at diagnosis, prolonged efficacy of low-dose sulfonylureas and, in HNF4A, a history of macrosomia or neonatal hypoglycaemia are the atypical features that should prompt testing [7].

Ketosis-prone T2D, or “African diabetes”, described in people from sub-Saharan Africa and also in Asian and Hispanic populations, presents acutely with ketoacidosis or a hyperosmolar state and goes into remission after a few weeks of insulin, remaining controlled thereafter for years with diet or metformin [7]. It predominates in men aged 30–40 years with prior overweight and a family history of diabetes, and is distinguished from monogenic diabetes and T1D by the combination of negative antibodies, metabolic syndrome and preserved residual insulin secretion despite the initial ketosis, in a person of at-risk ancestry [7].

6.3. Neonatal and infancy-onset diabetes

More than 80% of diabetes diagnosed before 6 months has a monogenic cause [11, 15]. Permanent neonatal diabetes presents with hyperglycaemia in the first 6 months (mean 7 weeks), intrauterine growth restriction, glycosuria, osmotic polyuria, severe dehydration and poor weight gain; “diabetes of infancy” has been proposed because diagnosis may occur between 3 and 6 months [5]. According to GeneReviews, the most frequent causes are KCNJ11 (approximately 25%), INS (20–25%) and ABCC8 (10–15%), followed by EIF2AK3 (Wolcott–Rallison syndrome, <10%), GATA6, biallelic GCK, GLIS3 and RFX6 (approximately 5% each), PDX1 (approximately 4%), SLC19A2 (2–3%), NEUROD1 (approximately 2%) and rarer causes, with fewer than 20% unexplained [5]. Syndromic forms include those associated with FOXP3 (IPEX), IER3IP1 (microcephaly and lissencephaly), NEUROG3 (malabsorptive diarrhoea), CNOT1, EIF2B1, YIPF5, activating STAT3 variants and WFS1 [5].

The most frequent transient neonatal diabetes causes are 6q24 abnormalities with overexpression of paternally imprinted PLAGL1 and HYMAI genes and ABCC8 and KCNJ11 variants [5]. In 6q24 disease, intrauterine growth restriction develops in the third trimester, mean birth weight was 1,930 g at 39 weeks in a 30-patient series, diabetes begins in week one and lasts a mean 3 months, and recurs in late childhood or young adulthood in possibly half or more (mean recurrence age 14 years in one series), potentially presenting as gestational diabetes [67]. Some carriers of 6q24 abnormalities without known neonatal diabetes develop diabetes later, and intrauterine growth restriction or low birth weight in a young adult with non-autoimmune diabetes is a clue to this differential diagnosis [67].

6.4. Mitochondrial diabetes, syndromes and other monogenic forms

In mitochondrial diabetes, the most common symptom besides diabetes is sensorineural deafness, which usually precedes it; other mitochondrial signs should be sought in all affected relatives, and the most frequent variant is the point variant m.3243A>G, which also causes MELAS syndrome (mitochondrial encephalomyopathy with lactic acidosis and stroke-like episodes) [17]. This variant is the fourth most frequent diagnosis in people referred for monogenic diabetes testing, and PDMI recommends testing for it in all suspected cases, even without familial deafness [2]. Transmission is maternal: the risk to the children of a carrier father is practically zero, and all children of a carrier mother inherit the variant, with variable heteroplasmy [2]. The French PNDS describes a presentation ranging from pseudo-T1D with negative antibodies to pseudo-T2D without metabolic syndrome, and lists as clues maternal transmission, familial sensorineural deafness, macular dystrophy (to be specifically requested from an experienced ophthalmologist), neurological or muscle disease, cardiomyopathy and nephropathy unrelated to diabetes [7].

The French PNDS adds clues for the remaining monogenic forms [7]. In KCNJ11 or ABCC8 MODY, the phenotype overlaps with that of the main forms, from moderate hyperglycaemia to more severe diabetes, and a personal or family history of neonatal diabetes or, conversely, of neonatal or early-childhood hypoglycaemia should prompt its consideration [7]. In INS, diabetes is mainly of neonatal or early-childhood onset, with a phenotype ranging from pseudo-T1D to moderate hyperglycaemia similar to GCK-MODY, and a family history is often absent because the variants frequently arise de novo [7]. In HNF1B-related disease, there is renal involvement in 90% of cases (cysts, malformations, renal failure, hypokalaemia, hypomagnesaemia), genitourinary malformations, pancreatic atrophy with exocrine insufficiency, fluctuating liver function test abnormalities, early-onset gout and, in 10%, intellectual disability; kidney disease without diabetic retinopathy is the main clue, and the absence of a family history does not exclude the diagnosis, because the variant is de novo in about 50% of cases [7].

Insulin receptor diseases include biallelic forms with extreme resistance (Donohue syndrome, with prenatal and postnatal growth retardation, hirsutism, aged facies, thick lips and prominent ears, and Rabson–Mendenhall syndrome, with acanthosis nigricans, polycystic ovaries and virilisation, of later presentation) and monoallelic dominant-negative forms with type A insulin resistance [17]. The lipodystrophies include Dunnigan-type familial partial lipodystrophy (LMNA), with loss of fat from the extremities and trunk and accumulation in the head and neck, the PPARG-associated form and the congenital generalised forms (BSCL2, AGPAT2) [17]. PDMI found LMNA variants in 9.7% of a French series of lipodystrophy or severe insulin resistance and PPARG variants in 0.6% of a Chinese series of young-onset diabetes [2]. The French PNDS points out that the clinical diagnosis of lipodystrophy can be difficult in partial forms and in men, and that in some families monogenic diabetes is the first hypothesis; with INSR variants, insulin resistance manifests as high insulinaemia, acanthosis nigricans and hyperandrogenism in women, but usually without dyslipidaemia or steatosis, which may lead to MODY being wrongly suspected [7].

Wolfram syndrome (diabetes insipidus, diabetes mellitus, optic atrophy and deafness) is autosomal recessive, caused by biallelic WFS1 variants at 4p16.1, and is frequently associated with urinary tract abnormalities such as hydronephrosis [17]. Orphanet also records a dominantly inherited Wolfram-like syndrome (ORPHA:411590), whose manifestations were not described in the sources consulted [69]. According to the French PNDS, the classic form combines early-onset insulin-requiring diabetes, optic atrophy, sensorineural deafness, diabetes insipidus, neurological manifestations and urinary tract abnormalities, but much milder forms exist; the recessive inheritance of most cases distinguishes it from MODY [7].

Table 6.2. Diabetes syndromes grouped by associated trait [17]

The original Oxford manual table also places Prader–Willi syndrome in the ocular signs column with a pigmentary retinopathy annotation, not a trait of this syndrome; that entry was omitted.

6.5. Congenital hyperinsulinism

Congenital hyperinsulinism has an incidence of 1:27,000 to 1:50,000 live births in non-consanguineous Western populations and up to 1:2,500 in populations with consanguinity, a mean birth weight of 3.7 kg and presentation with hypoglycaemic seizure or coma in half of cases [40]. The genes for dominantly inherited hyperinsulinism include GLUD1, GCK, SLC16A1, HNF4A, HNF1A, UCP2 and dominant forms of ABCC8 and KCNJ11; the most severe recessive forms result from inactivating variants of ABCC8 and KCNJ11 and of HADH [26]. In diazoxide-responsive hyperinsulinism, a genetic cause is identified in up to 47% of cases, predominantly ABCC8 and GLUD1, and lack of response to diazoxide suggests ABCC8, KCNJ11 or GCK [40]. The recessive forms of ABCC8 have founder variants in Ashkenazi Jewish and Finnish populations [36]. The HNF1A, HNF4A, GCK, ABCC8 and KCNJ11 genes cause both hyperinsulinism and monogenic diabetes, so that children with diffuse hyperinsulinism in remission warrant lifelong glycaemic follow-up [40].

7. Treatments and Risk Management Strategies

7.1. General principles and evidence quality

Molecular diagnosis changes treatment specifically: no therapy for GCK-MODY and first-line sulfonylureas for HNF1A and HNF4A forms, with lower costs and identification of at-risk relatives [10, 11]. Among reasons for molecular diagnosis, ISPAD adds more accurate complication-risk prognosis and prevention of stigma and employment restrictions associated with a diabetes label, particularly in GCK-MODY (grade C) [10]. Evidence is nevertheless low quality. The PDMI treatment systematic review identified 2,389 MODY studies, of which 33 were extracted (12 on GCK, 23 on HNF1A and 3 on HNF4A); only 6 were experimental and 4 randomised clinical trials (RCTs), none with power calculation, and most had high risk of bias [14].

No drug has FDA or European Medicines Agency (EMA) authorisation specifically for MODY; sulfonylureas (SUs), glinides, glucagon-like peptide 1 (GLP-1) receptor agonists, dipeptidyl peptidase 4 (DPP-4) inhibitors and sodium–glucose cotransporter 2 (SGLT2) inhibitors are used off-label, with labels referring to adult T2D and not establishing paediatric safety and efficacy [70–72]. The sole formulation authorised for a monogenic diabetes form is Amglidia® oral glibenclamide suspension in the EU for neonatal diabetes [73].

7.2. GCK-related hyperglycaemia

Outside pregnancy, do not treat. GCK-MODY hyperglycaemia does not progress, causes no relevant complications and does not respond to low-dose insulin or oral antidiabetics, so these patients should not receive treatment (ISPAD, grades B and C) [10]. PDMI makes the same recommendation (“no medical treatment should be given for GCK-related hyperglycaemia”, grade C) and found no basis for dietary recommendations [14]. In a study of 799 patients, 168 (21%) were medicated at referral (13.5% oral agents and 7.5% insulin), without an HbA1c difference versus untreated patients (48 versus 46 mmol/mol, or 6.5% versus 6.4%; p=0.11), and in 16 stopping treatment (10 insulin and 6 oral agents), mean change was −0.68 mmol/mol (95% CI −2.97 to 1.61), approximately −0.06% [74]. PDMI combined 175 individuals followed untreated for 3 to 126 months and 35 after withdrawal, without significant HbA1c change [14]. In the United States national monogenic diabetes registry, GCK-MODY was frequently misdiagnosed and unnecessarily treated [75]. A small continuous glucose monitoring (CGM) study compared a high-carbohydrate (60%) and low-carbohydrate (25%) diet over 2 days in 10 adults and found, in a non-prespecified analysis of 7 with HbA1c >6.5%, mean glucose 0.76 mmol/L higher with the high-carbohydrate diet, insufficient data to recommend a diet [14].

The commonest clinical error is treatment: after genetic diagnosis, glucose-lowering drugs should be stopped and the patient reassured, except during pregnancy (section 7.6) [10, 35]. When GCK-MODY coexists with T1D or T2D, treatment follows the coexisting type, but the HbA1c target should lie within the GCK-MODY expected range because set point is higher and counterregulatory threshold lower [1]. PDMI identified absent recommendations for this coexistence and long-term cardiovascular effects in older people [14].

The French PNDS explains the ineffectiveness of glucose-lowering drugs by counter-regulation: in GCK-MODY, glucagon and catecholamines rise earlier and more strongly than in healthy people, at higher glucose levels, and oppose the fall in glucose [7]. It recommends stopping antidiabetic drugs after genetic confirmation, provided there are no arguments for associated T2D, notably overweight, and checking that HbA1c remains ≤7% [7]. This operational threshold does not appear in ISPAD or PDMI, which simply recommend no treatment, and coincides with the upper limit of the HbA1c range characteristic of GCK-MODY [10, 14]. Lifestyle advice is that given to the general population, with a balanced diet without carbohydrate restriction and regular physical activity, to prevent later T2D and cardiovascular disease [7].

7.3. HNF1A-related diabetes

Non-pharmacological measures. The French PNDS, which addresses the HNF1A and HNF4A forms together, applies to them the advice given in T1D: combating sedentary behaviour with regular physical activity, ideally 30 minutes a day on 5 days a week, and dietary education focused on the carbohydrate content of foods [7]. As loss of the early secretory peak causes marked postprandial hyperglycaemia, it recommends a carbohydrate-calibrated diet, with a preference for low-glycaemic-index meals, an initial consultation with a dietitian and regular reassessment of knowledge [7]. If there is overweight, a calorie reduction relative to usual intake is added, with restriction of fats and a preference for polyunsaturated fats, to combat metabolic syndrome [7].

First-line low-dose sulfonylurea. Initially, diet may suffice because hyperglycaemia is mainly postprandial after carbohydrate intake, but most eventually require medication [10]. Patients are extremely sulfonylurea-sensitive and generally achieve better control than with insulin, particularly children and young adults (ISPAD, grade B) [10]. PDMI recommends sulfonylurea first-line (grade C), glinides with frequent sulfonylurea hypoglycaemia (grade D), GLP-1 receptor agonists as an option (grade C) and DPP-4 inhibitors as add-on therapy (grade C) [14]. Sulfonylureas act downstream of the genetic defect by closing K_ATP through SUR1 and increasing secretion through a glucose-independent mechanism [1, 27].

The trial underpinning this practice is Pearson and colleagues (2003): a crossover superiority RCT, with 6-week periods separated by 1-week washout, compared gliclazide and metformin in 18 HNF1A adults and 18 T2D patients matched by BMI and fasting glucose; in HNF1A patients, fasting glucose fell 4.7 mmol/L with gliclazide and 0.9 mmol/L with metformin (5.2-fold greater response; p=0.0007), and gliclazide response was 3.9-fold greater than in T2D (p=0.002) [14, 29]. HNF1A patients had a strong secretory response to intravenous tolbutamide despite little response to intravenous glucose and were more insulin-sensitive than T2D patients, locating the defect upstream of the sulfonylurea receptor [29].

Dose. ISPAD recommends starting at approximately one-quarter of the usual adult starting dose to avoid hypoglycaemia and mentions patients maintained for decades at low doses, for example gliclazide 20–40 mg/day [10]. This corresponds to one-quarter to one-half of a Diamicron® 80 mg tablet, whose adult T2D starting dose is 40–80 mg [76]. In the USA, where gliclazide is not marketed, glibenclamide (glyburide) is the most used sulfonylurea, started at low dose with insulin reduction or withdrawal to avoid hypoglycaemia [1]. If hypoglycaemia occurs despite titration in one or two daily doses, ISPAD suggests a prolonged-release preparation or short-acting secretagogue (meglitinide) with meals [10].

Table 7.1. Drugs used in HNF1A- and HNF4A-related diabetes: labelled dosing (adult T2D) and MODY adaptation

The main sulfonylurea adverse effects are hypoglycaemia, enhanced by these patients' increased insulin sensitivity, and weight gain, associated with loss of efficacy [1, 14]. The United States glibenclamide label contraindicates it in T1D or ketoacidosis and with bosentan and retains a cardiovascular mortality warning derived from the tolbutamide University Group Diabetes Program (UGDP) study [70]. The British gliclazide summary of product characteristics (SmPC) contraindicates it in sulfonylurea or sulfonamide hypersensitivity, T1D, diabetic precoma or coma, ketoacidosis, severe renal or hepatic insufficiency, with miconazole and during breastfeeding, and recommends considering an alternative in glucose-6-phosphate dehydrogenase deficiency because of haemolytic anaemia risk and caution in porphyria [72]. Glimepiride is contraindicated in sulfonamide hypersensitivity; in a 24-week trial of 272 T2D patients aged 8–17 years, it failed the objective versus metformin (HbA1c difference 0.2% favouring metformin; 95% CI −0.3 to +0.6) [71]. Hypoglycaemia management involves dose reduction, divided dosing or switching to a glinide or modified-release form [10, 14].

Transition from insulin to sulfonylurea. Patients misdiagnosed as T1D may stop insulin and start sulfonylurea without ketoacidosis risk [1]. In an uncontrolled study, 8 HNF1A patients treated with insulin from diagnosis (median 0.5 U/kg/day for 20 years) switched to gliclazide (median 80 mg/day): all stopped insulin and median HbA1c fell 0.8%, improving in 6 of 8 and worsening by 3.2% in one [14, 79]. In a subsequent prospective study of 27 HNF1A and 7 HNF4A patients, 25 of 31 receiving insulin stopped it, but only 12 (48%) maintained HbA1c ≤7.5%; success was associated with shorter diabetes duration and lower transition HbA1c [14]. Before insulin withdrawal, C-peptide should be measured to exclude severe deficiency [11].

Durability and treatment intensification. Sulfonylurea control may deteriorate over time, particularly with diabetes duration and weight gain: in a cohort, patients with median duration 4.6 years moved from HbA1c 7.5% to 6.4% over 24 months, whereas those with median duration 18.1 years moved from 8.8% to 9.2% [14]. In the German and Austrian DPV registry, with 114 HNF1A young people, median HbA1c was 7.5% with insulin, 6.7% sulfonylurea, 7.9% insulin plus sulfonylurea, 6.8% glinides and 6.1% untreated [14]. The best intensification is uncertain, with GLP-1 receptor agonists or insulin mentioned [1]. PDMI considered SGLT2 inhibitor safety uncertain in HNF1A because of insulin deficiency, already reduced SGLT2 expression and glycosuria, and advocated trials of drugs with proven cardiovascular benefit [14]; a 2026 RCT partly addressed this (section 7.5).

Position of the French PNDS. In the HNF1A and HNF4A forms, the PNDS recommends a low-dose sulfonylurea as first line, about a quarter of the usual dose, and considers that glinides have similar efficacy with less hypoglycaemia [7]. It identifies long duration of diabetes, signs of insulin resistance and high HbA1c as factors predicting poor response to sulfonylurea [7]. GLP-1 receptor agonists and DPP-4 inhibitors may be used outside the marketing authorisation (off-label), in coordination with the diabetologist, but remain second line for lack of evidence; linagliptin, tested in the trial with glimepiride, is not marketed in France [7]. SGLT2 inhibitors are not advocated: HNF1A already reduces SGLT2 expression in the tubule, a single 10 mg dose of dapagliflozin caused more glycosuria in HNF1A patients than in T2D, and evidence of efficacy was lacking, with a potential risk of ketosis [7]. This position predates the 2026 MOD3ST RCT, which showed a short-term benefit of adjunctive empagliflozin without an increase in β-hydroxybutyrate, and the 2026 French series without ketoacidosis; neither of these publications is large or long enough to replace the PNDS recommendation with a routine indication [78, 80].

Insulin and advanced therapies. In the long term, insulin often becomes necessary. The PNDS recommends an optimised regimen, with multiple daily injections (basal-bolus) or a subcutaneous pump, and education in functional insulin therapy, as in T1D [7]. A closed-loop system, which couples the pump to an algorithm fed by continuous glucose monitoring, may be considered only after discussion with specialist teams, as it is an off-label indication, and recourse to combined kidney–pancreas or kidney–islet transplantation is decided at a multidisciplinary consultation meeting [7].

Cardiovascular prevention. Increased cardiovascular risk despite high HDL and low hs-CRP justifies statin initiation no later than 40 years [1, 46].

7.4. HNF4A-related diabetes, K_ATP MODY, INS and rare genes

HNF4A-MODY. Sulfonylureas are the established first line, and it is reasonable to assume a response to meglitinides and GLP-1 receptor agonists, without formal data [1]. PDMI found only 3 studies with 16 individuals and could formulate no recommendation [14]. The 2025 Irish cohort, with 42 carriers phenotyped by OGTT, showed that 51.6% of patients on sulfonylurea monotherapy had a significant reduction in HbA1c (p=0.045), maintained after a median of 6 years (IQR 3.5–11), whereas 48.4% required additional drugs, including insulin; the latter had higher BMI (p=0.037), longer diabetes duration (p=0.03) and lower secretory capacity (C-peptide area under the curve; p=0.01) [51]. A normal BMI predicted response, and avoiding weight gain was associated with a sustained response; the authors propose GLP-1 receptor agonists and SGLT2 inhibitors as possible adjuncts, a hypothesis not tested in a trial [51]. The French PNDS, which manages HNF4A-MODY like HNF1A-MODY outside the neonatal period, reports favourable effects of GLP-1 receptor agonists in some patients, without large-scale studies, and keeps them as off-label second line; it does not advocate SGLT2 inhibitors in this form [7]. The associated neonatal hyperinsulinism is treated with intravenous glucose, diazoxide and other measures; in carriers of the p.(Arg63Trp) variant, renal follow-up of the tubulopathy, nephrocalcinosis and renal function is warranted [38].

MODY due to ABCC8 and KCNJ11. The ADA/EASD consensus considers these forms, like HNF1A and HNF4A, acutely sulfonylurea-sensitive [15]. GeneReviews indicates sulfonylureas for ABCC8 and diet, sulfonylureas or insulin for KCNJ11, without specific doses [1]. ISPAD classifies these patients, when neonatal diabetes occurs in the family, as high-dose sulfonylurea candidates [10].

K_ATP neonatal diabetes (differential and context). Approximately 90% of K_ATP neonatal diabetes patients switch from insulin to sulfonylurea according to ISPAD, with higher weight-based doses than adult T2D, typically approximately 0.5 mg/kg/day glibenclamide and occasionally up to 2.3 mg/kg/day; dose depends on onset age and variant, and many progressively reduce it [10]. ADA 2026 reports that 30–50% of people with K_ATP neonatal diabetes improve with high-dose oral sulfonylureas instead of insulin, differing from ISPAD with no origin clarified in the sources consulted [10, 11]. Reported glibenclamide adverse effects in infants are limited to transient diarrhoea and tooth discolouration; coeliac disease may cause secondary failure; sulfonylurea-stimulated secretion depends mainly on meals, so most should contain carbohydrate; neurological benefit is partial and greater with early initiation [10]. With 5–10 years' follow-up, mean HbA1c was 5.9–6.0% with minimal hypoglycaemia, and later sulfonylurea initiation was associated with greater add-on insulin requirement [5, 81].

Table 7.2. Oral glibenclamide suspension (Amglidia®) dosing for neonatal diabetes according to European SmPC [73]

INS-MODY and rare genes. In MODY10, diet, oral antidiabetics or insulin, sometimes small doses, are used [1]; in INS neonatal diabetes, insulin is preferred [11]. Rare causes have no therapeutic outcome data, and treatment follows clinical judgement [1].

Table 7.3. Treatment by gene according to GeneReviews [1]

7.5. Clinical trials, developments and regulatory status

Table 7.4. Main treatment trials and studies

These studies have small samples, short durations and, in crossover trials, intermediate outcomes (fasting glucose, mean glucose, MAGE), so they do not inform about complications, cardiovascular events or mortality. Results do not permit indirect comparisons between drugs tested in different trials, and absence of a significant difference in Østoft's 15-patient trial does not demonstrate equivalence. ISPAD describes the liraglutide trial as showing lower fasting glucose with the GLP-1 agonist, but trial data reproduced in PDMI's systematic review show a numerically greater reduction with glimepiride without significant difference, and these data take precedence [10, 14]. The negative PRISM trial also included monogenic diabetes (23 of 884) but aimed to evaluate a precision medicine strategy for all young-onset diabetes rather than a MODY drug [82].

Developments over the past 12 months (October 2025 to October 2026), by evidence level.

Published randomised clinical trial. MOD3ST-TRIAL, conducted in Denmark (Steno Copenhagen and Aarhus), tested empagliflozin 25 mg once daily versus placebo for 4 weeks each, with 2-week washout, in HNF1A-MODY adults with HbA1c ≥6.5% on at least one stable antidiabetic (89% sulfonylurea) [78]. Besides the primary outcome, time in range (3.9–10.0 mmol/L) rose from 52% to 78% (+26 percentage points; 95% CI 16–37), 24-hour glycosuria from 4 to 102 g and renal glucose threshold from 12.0 to 4.0 mmol/L, with 0.7 kg weight reduction [78]. Hypoglycaemia and β-hydroxybutyrate did not differ (no participant ≥1.5 mmol/L); adverse effects were mild (balanitis in 2 of 8 men, vulvovaginitis in 1 of 10 women, urinary infection in 1 of 18), with one renal stone during washout after placebo [78]. Limitations are 4-week duration, small sample and add-on use only [78].

Observational study. A retrospective French national series of 40 adults with monogenic diabetes receiving SGLT2 inhibitors (16 MIDD, 9 HNF1B, 11 HNF1A or HNF4A, 4 other genes), followed for 23.9 ± 5.9 months, showed overall stable HbA1c, declining from 9.6 ± 1.3% to 7.6 ± 0.7% in those starting at ≥8% (p=0.001), albuminuria reductions of 35.6% at 1 year and 43.5% at 2 years (p=0.004), non-serious adverse effects in 15%, discontinuation in 7.5% and no ketoacidosis [80]. Authors note euglycaemic ketoacidosis risk in insulinopenic patients and that routine use is not recommended [80].

Ongoing trial. Dorzagliatin, an allosteric glucokinase activator, is being tested against placebo in adults with GCK-MODY in a phase 2 randomised crossover trial at the Chinese University of Hong Kong (NCT06976658; estimated n=44; primary outcome fasting glucose at 8 weeks; started April 2025, primary completion expected December 2026; recruiting without results) [83]. An earlier single-dose 75 mg trial studying first-phase insulin response in recent T2D and GCK-MODY completed in 2022 without registry results [84].

Preclinical study. In a murine GCK-MODY model, dorzagliatin and GLP-1 receptor agonism had additive effects [85]. A Norwegian phase 4 trial on switching insulin to sulfonylureas in HNF1A, HNF4A and HNF1B has had unknown status since 2020 [86]. The ClinicalTrials.gov search identified no other active MODY drug trials, which does not prove their absence.

Table 7.5. Comparative regulatory status (USA and EU)

The sources consulted include no national access data for Portugal (INFARMED) or Brazil (ANVISA).

7.6. Pregnancy and breastfeeding

General principles. In pregnancy with diabetes, ADA recommends fasting glucose <95 mg/dL (<5.3 mmol/L), postprandial <140 mg/dL (<7.8 mmol/L) at 1 hour or <120 mg/dL (<6.7 mmol/L) at 2 hours, and HbA1c <6% (<42 mmol/mol) without significant hypoglycaemia, allowing up to <7% (<53 mmol/mol) (grade B) [87]. Insulin is preferred; metformin and glibenclamide should not be first-line because they cross the placenta (grade A) and may not achieve targets (grade B) [87]. Cord blood glibenclamide concentration is approximately 50–70% of maternal levels, and meta-analyses associate it, versus insulin or metformin, with more neonatal hypoglycaemia, macrosomia and increased neonatal abdominal circumference, without long-term offspring safety data [87]. ADA does not specifically address MODY pregnancy, and the following recommendations come from specialist centre reviews and observational studies [27, 39, 87].

Maternal GCK-MODY: treatment depends on fetal genotype. If the fetus inherits the maternal variant, it shares the higher set point, secretes insulin normally and grows normally; if not, it responds to maternal hyperglycaemia with hyperinsulinaemia and excessive growth, with birth weight approximately 600–700 g higher [1, 27]. If paternally inherited or de novo, the carrier fetus of a non-carrier mother secretes less insulin and is born approximately 400–500 g lighter, without indicated treatment [1, 27]. Maternal insulin is appropriate only when the fetus is not a carrier: treating a mother with a carrier fetus is associated with growth restriction (small for gestational age) [35, 66].

Table 7.6. GCK-MODY pregnancy management according to variant origin and fetal genotype [1, 27, 66]

Fetal genotype is rarely known. The classical strategy infers it by ultrasound: fetal abdominal circumference disproportionately above the 75th centile at 26 weeks suggests a non-carrier fetus and warrants insulin and delivery at 38 weeks, whereas circumference below the 75th centile suggests a carrier fetus [1, 27]. This inference is imprecise: in an international study, abdominal circumference >P75 at 28 weeks had 53% sensitivity and 61% specificity for fetal genotype [66]. Third-trimester insulin has not demonstrated prevention of macrosomia: in an Irish series, macrosomia in non-carrier fetuses was 33.3% with insulin and 62.5% without, and another study found no reduction in large-for-gestational-age newborns [27]. Some authors advocated insulin for all pregnant women with GCK-MODY from early pregnancy, a position GeneReviews considers to require further study [1].

Non-invasive fetal genotyping using cell-free fetal DNA changed this approach. In Hughes and colleagues' study, relative variant dosage by ddPCR, a non-invasive prenatal test (NIPT-GCK), was concordant in 37 of 38 pregnancies, with 100% sensitivity (95% CI 78–100%), 96% specificity (95% CI 78–100%), PPV 94% and NPV 100%, using offspring genotyping as reference [66]. In 43 consecutive pregnancies (July 2019 to September 2021), valid genotypes were obtained in all 38 with analysable variants, with median time to result 5 weeks (IQR 3–8), 2 weeks when a variant assay already existed; in 23 of 25 samples received before 20 weeks, results arrived before 28 weeks [66]. Estimated cost was approximately 2,000 pounds, and the United Kingdom National Health Service formally accredited testing for GCK-MODY pregnancies [66]. An earlier 33-pregnancy series showed 100% cord blood concordance [2, 88]. In Geneva, gene-specific rather than variant-specific relative haplotype dosage gave results within 4–11 days in 10 samples from 7 pregnant women, with testing recommended at 12–14 weeks, one insufficient-DNA failure successfully repeated, cost 920 Swiss francs and a vanishing twin identified as a limitation after publication [89]. In that series, carrier fetal genotype allowed moderate maternal hyperglycaemia (fasting 5.8–7.1 mmol/L and 1-hour postprandial <8.5 mmol/L) without insulin, while non-carrier fetal genotype led to gestational diabetes targets with mealtime insulin, without growth restriction or macrosomia [89]. In Australia, where testing is unavailable locally, a 2-case series described sending samples to Exeter and estimated the strategy cost-effective [90].

The prospective French study of 46 pregnant women (NCT02556840) [91] compared maternal-glucose-guided insulin (n=25) with fetal abdominal circumference ≥P75 or safety threshold guidance (glucose ≥6.7 mmol/L before or ≥11.1 mmol/L 2 hours after meals; n=21): insulin was used in 100% versus 75% (p=0.01), earlier and with more maternal hypoglycaemia in the first group, without outcome differences [92]. Overall, 22.2% of newborns were large for gestational age, 0% among carriers versus 36% among non-carriers (p=0.005), a fetal-genotype-dependent difference rather than strategy- or control-dependent [92]. A Chinese meta-analysis of 8 studies (257 mothers and 499 children) found congenital malformations in 2.4%, without difference between carrier and non-carrier offspring (OR 0.56; 95% CI 0.07–4.51), and less frequent macrosomia, neonatal hypoglycaemia and composite adverse outcome in carriers [93].

Maternal insulin therapy is difficult in GCK-MODY: exogenous insulin suppresses endogenous secretion, the counterregulatory threshold is higher and autonomic symptoms appear at higher glucose, making symptomatic hypoglycaemia frequent and potentially requiring high doses [1, 27]. CGM has been proposed to guide insulin when fetal genotype is unknown, but validated thresholds are lacking [94]. In a Polish series of 18 GCK pregnancies, all switched to insulin, differing from the fetal-genotype-guided approach [95]. Invasive fetal testing is not recommended solely for this purpose [1, 27].

The French PNDS, published before non-invasive fetal genotyping became widespread, acknowledges that there is no formal recommendation for GCK-MODY in pregnancy and describes two practices: surveillance of fetal abdominal circumference by repeated ultrasound from 26 weeks of gestation, with insulin if it exceeds the 75th centile and induction of labour if the weight is high for gestational age; and insulin from the outset to reach pregnancy glycaemic targets, irrespective of fetal growth [7]. It considers reasonable a decision that integrates the degree of maternal hyperglycaemia before conception and in early pregnancy (with a risk of malformation if HbA1c exceeds 6.5%), the insulin resistance presumed from maternal BMI and the assessment of fetal growth, and anticipated that fetal DNA in maternal blood would come to refine fetal risk [7]. It further recommends that, when MODY is suspected in pregnancy, the usual management of gestational diabetes be applied until the test result is available [7]. The PNDS estimates MODY at 2–6% of gestational diabetes, depending on the selection criteria, and GCK-MODY at about one third of women with fasting glucose >5.5 mmol/L and BMI <25 kg/m² [7].

Maternal HNF1A-MODY and HNF4A-MODY. Sulfonylureas have not been studied in MODY pregnancies, and insulin is considered first-line [27]. Shepherd and colleagues proposed two options: discontinue sulfonylurea before pregnancy and switch to insulin; or, if preconception HbA1c meets local targets, continue or switch to glibenclamide before or early in pregnancy, introduce basal insulin in the 2nd trimester and subsequently replace glibenclamide with boluses, achieving excellent control well before 26 weeks [39]. If already pregnant on sulfonylurea with HbA1c <48 mmol/mol (6.5%), there are arguments for continuation until the end of the 1st trimester because switching to insulin then may transiently worsen control during organogenesis, and no increase in congenital abnormalities or pregnancy loss was observed with 1st-trimester sulfonylureas [27, 39]. Other sulfonylureas should be replaced by equivalent-dose glibenclamide and, if continued into the 3rd trimester, the dose should be as low as possible (≤5 mg/day) [39]. In an Irish series, insulin was initiated in 26% of pregnancies, with median requirements 0.4 U/kg/day in the 1st trimester and 0.8 U/kg/day in the 3rd [27].

Baseline congenital defect risk is 3–4%, increasing to 6–8% in pregestational insulin-dependent diabetes and intermediate in non-insulin-dependent pregestational diabetes; one meta-analysis found no congenital defect rate difference between oral hypoglycaemics, including glibenclamide, and insulin, whereas another associated glibenclamide with more macrosomia and neonatal hypoglycaemia [1]. Rigorous maternal hyperglycaemia treatment, high-resolution ultrasound with fetal echocardiography and maternal–fetal medicine follow-up are recommended [1].

Position of the French PNDS for the HNF1A and HNF4A forms. As in any pregestational diabetes, management aims at pregnancy glycaemic targets, screening for and treatment of nephropathy and retinopathy and the prevention of obstetric complications, including gestational hypertension [7]. The preconception consultation defines the goals, replaces the sulfonylurea with insulin, because sulfonylureas remain contraindicated in pregnancy in France, prescribes folic acid, assesses complications and addresses the risk of transmission; preconception HbA1c should be below 6.5%, because of the risk of malformation associated with hyperglycaemia at conception [7]. In women with HNF1A, screening liver ultrasound is required, or specialist surveillance if adenomatosis is already present, because of the risk of worsening with the rise in oestrogens [7]. Joint follow-up by an endocrinologist-diabetologist and an obstetrician in a maternity unit with a neonatal unit, of at least level 2, is considered reasonable [7]. The mandatory switch to insulin before conception diverges from the option, described by Shepherd and colleagues, of maintaining or introducing glibenclamide in the 1st trimester when HbA1c is at target (Table 7.9) [39].

Table 7.7. Pregnancy management when a parent has HNF1A-MODY or HNF4A-MODY [27, 39]

In HNF4A, neonatal hypoglycaemia occurs in at least 10% of carriers irrespective of maternal glucose, and macrosomia exposes to shoulder dystocia, brachial plexus injury, instrumental delivery and emergency caesarean [27, 39]. Cell-free fetal DNA is already used in GCK and ABCC8 and may be used when the father carries HNF4A [27]. Hyperinsulinism without perinatal stress in a child of an HNF1A carrier should prompt newborn HNF1A testing [27].

The newborn according to the French PNDS. As the genotype of the newborn of a mother with MODY is usually unknown, the PNDS recommends systematic glucose monitoring in the first days of life in all of them, because of two risks of hypoglycaemia: hyperinsulinism reactive to maternal hyperglycaemia and congenital hyperinsulinism in a newborn carrying an HNF1A or HNF4A variant [7]. Hyperinsulinism also exposes the newborn to neonatal hypocalcaemia and jaundice, and the neonatal paediatrician should be aware of these risks and call on paediatric diabetology if necessary [7]. The newborn carrying HNF4A, of maternal or paternal inheritance, is at risk of persistent hyperinsulinism in the first weeks, which may require dietary and pharmacological treatment with specialist follow-up and disappears in the first years of life [7]. The PNDS attributes the risk of fetal hyperinsulinism with macrosomia, even with paternal transmission, to both HNF4A and HNF1A, which for HNF1A is not confirmed by the birth weight data cited in section 3.1, which show no difference from siblings [7, 27]. The newborn carrying a paternally inherited GCK variant has intrauterine growth restriction due to relative hypoinsulinism in the face of normal maternal glucose levels, and is managed like any small-for-gestational-age newborn: an incubator to prevent hypothermia and prevention of nutritional hypoglycaemia with early feeding, glucose monitoring and enteral nutrition with intravenous supplementation if necessary [7].

Maternal K_ATP diabetes. The carrier fetus, maternally or paternally inherited, has reduced weight (median 2,580 g at 39 weeks) [39]. In 4 pregnancies of 3 women with a KCNJ11 variant receiving glibenclamide (2.8–90 mg/day), the unaffected fetus developed hyperinsulinism, excessive growth and neonatal hypoglycaemia, whereas the affected fetus normalised growth (median 3,010 g at 38 weeks) [39]. Recommendations are to reduce glibenclamide to the minimum maintaining HbA1c ≤48 mmol/mol; continue if fetus affected (genotyped by amniocentesis or chorionic villus sampling performed for another reason, or with reduced growth on ultrasound from 28 weeks); and, if fetus unaffected or without evidence, switch to insulin in the 3rd trimester with rapid transition (<3 days) to basal–bolus treatment [39]. Delivery should occur between 37 and 38+6 weeks, with glucose at birth and at least every 8 hours for ≥48 hours and rapid cord blood genetic testing, freely available in Exeter with results within approximately 1 week [39]. A 2025 report described better neurodevelopment in a KCNJ11 neonatal diabetes patient exposed to sulfonylurea in utero [96].

Breastfeeding. Glibenclamide 5–10 mg/day was neither detected in milk nor associated with neonatal hypoglycaemia; gliclazide has not been studied and is contraindicated during breastfeeding according to SmPC, so persistent dysglycaemia during breastfeeding should be treated with modest-dose glibenclamide or continued insulin [27, 72]. In K_ATP diabetes, hypoglycaemia was described in an unaffected newborn of a mother receiving 90 mg/day, and doses above 0.2 mg/kg/day require attention during breastfeeding [39]. Another sulfonylurea may be resumed after weaning [39].

7.7. Agents to avoid, perioperative care and transition

The US labels recommend stopping glibenclamide and glimepiride at least 2 weeks before the expected date of delivery, because of the risk of severe and prolonged neonatal hypoglycaemia (4–10 days) [70, 71]. Gliclazide should be avoided with miconazole and in severe renal or hepatic impairment, and sulfonylureas require caution in glucose-6-phosphate dehydrogenase deficiency [72]. SGLT2 inhibitors require monitoring for euglycaemic ketoacidosis in insulin-deficient patients [80]. Insulin should be neither initiated nor maintained in isolated GCK-MODY outside pregnancy [10]. In women with HNF1A and hepatic adenomatosis, oestrogen-containing contraception is formally contraindicated [7]. Self-monitoring of blood glucose becomes indispensable when corticosteroids or immunosuppressants, which alter glucose homeostasis, are combined [7]. The sources consulted do not describe specific perioperative recommendations for MODY. The French PNDS includes the organisation of the transition from paediatric to adult care among the goals of follow-up and asks that situations requiring specific management, such as pregnancy, be anticipated [7]. During the transition from paediatric to adult care, the annual follow-up suggested by PDMI makes it possible to update the diagnosis, treatment and variant classification and to maintain counselling on the risk to offspring [2].

7.8. Guideline recommendations and divergences

Table 7.8. Guideline recommendations and divergences between sources

The 2022 French PNDS converges with the other guidelines on the essential points (no treatment of GCK-MODY outside pregnancy; sulfonylurea at about a quarter of the dose as first line in HNF1A and HNF4A; NGS panel with CNV), but diverges on several details (Table 7.9).

Table 7.9. Divergences between the French PNDS and the other sources [7]

7.9. Therapeutic education, goals and daily life

The French PNDS sets as goals of treatment, besides metabolic control, multidisciplinary coordination, support for adherence and detection of adverse effects, early integration of therapeutic education of the patient, parents and close relatives, shared decision-making, discussion of participation in trials, assessment of quality of life, self-esteem and social and school or working life, psychological support when needed, referral to patient associations, liaison with the school doctor and the occupational physician and, where relevant, recognition of disabled worker status [7]. The reference specialist is usually the endocrinologist-diabetologist, and the nurses and dietitians involved in therapeutic education have a central role in initiation and follow-up [7].

Therapeutic education is organised by a team with a physician, an education nurse, a dietitian, a psychologist, a social worker and, if possible, a sports educator, and should enable the patient to understand the disease, the treatment (lifestyle adaptation and drugs) and surveillance, which varies with the gene; to recognise the signs of decompensation and the associated manifestations in the HNF1A and HNF4A forms; to understand the mode of transmission; to know the periods that require special follow-up, such as pregnancy in GCK, HNF1A and HNF4A (risk of macrosomia and neonatal hypoglycaemia); to know about follow-up in adulthood; and to be able to talk about the disease [7]. When there is prior overweight or obesity, education aims at weight loss and stabilisation with adapted physical activity [7]. Insulin-treated patients are trained in functional insulin therapy, as in T1D [7].

In children, an individualised reception protocol (projet d'accueil individualisé, PAI) should be drawn up and given to the family when drugs must be taken during school hours; meals in the school canteen require no adaptation, because the recommended diet is a balanced one, the same as that of other children [7].

8. Reproductive Options and Family Counselling

8.1. Inheritance pattern and recurrence risks

The HNF1A, HNF4A, GCK, ABCC8, KCNJ11 and INS forms of MODY are transmitted in an autosomal dominant manner: each child of a carrier has a 50% probability of inheriting the variant and, when a parent is a carrier, the proband's siblings have the same risk [1, 2]. The estimated overall penetrance of MODY in referred families is about 95%, with environmental factors influencing the age at onset [17]. De novo variants occur, but their proportion is unknown for most genes: they are frequent in the 17q12 deletion (about 70%) and rare in GCK, HNF1A and HNF4A, with a rate that, in a small study, approached 7% [1]. When neither parent has the variant in blood and parentage is confirmed, the recurrence risk for siblings is about 1%, owing to the possibility of germline mosaicism [1, 2]. In the absence of the variant in both parents, alternative paternity or maternity, including assisted reproduction, or undisclosed adoption should also be considered [1]. The French PNDS recommends that, when the variant is identified in a family with endogamy, testing be offered to both parents, because homozygosity in the offspring requires specific management (in GCK, the biallelic form causes permanent neonatal diabetes) [7, 10]. The same document emphasises the variable expressivity of a given variant between and within families, and the frequent coexistence of MODY and T2D in the same family [7].

Other patterns apply to differential diagnoses: recessive forms such as Wolcott–Rallison syndrome (EIF2AK3), biallelic GCK, INS and ABCC8 forms and thiamine-responsive megaloblastic anaemia imply 25% risk for each sibling [2, 10]; in KCNJ11 neonatal diabetes, approximately 90% of variants are de novo, but reported germline mosaicism makes sibling risk low rather than zero [10]; and in m.3243A>G mitochondrial diabetes, risk is practically zero with a carrier father, whereas all children of a carrier mother inherit the variant with variable heteroplasmy [2].

Table 8.1. Assessment of relatives of a MODY proband [1]

8.2. Cascade testing and carrier surveillance

Familial variant testing should be offered to affected relatives to confirm aetiology and adjust treatment, and may be offered predictively to asymptomatic relatives at high risk of HNF1A, HNF4A or HNF1B [2, 17]. For GCK, PDMI proposes discussing cascade testing with relatives who had gestational diabetes or have prediabetes and starting with fasting glucose in apparently unaffected adult first-degree relatives: if normal, GCK-MODY is highly unlikely and genetic testing unnecessary; if elevated, test the variant [2]. A diabetic relative negative for the variant has another aetiology such as T1D or T2D [2].

The French PNDS organises family screening in three situations. Relatives with hyperglycaemia or diabetes are offered targeted testing for the index case's variant, particularly in HNF1A and HNF4A, because of variable expression and the difficulty of distinguishing these forms from T1D and T2D [7]. In relatives of patients with GCK-MODY, the first step is fasting glucose, and genetic testing is performed if it is above 5.5 mmol/L; it should be specifically offered to women with fasting hyperglycaemia, a family history of GCK-MODY and plans for pregnancy, because the result changes the management of pregnancy [7]. In relatives of patients with HNF1A or HNF4A, genetic testing is offered in a specialist setting to all who wish it, from the age of 10 years. When testing has not been performed or is positive, annual fasting glucose and HbA1c are recommended; absence of the variant returns the person to general population risk [7].

Early genetic status knowledge enables glucose monitoring and subtype-appropriate treatment, reducing long-term complication risk according to GeneReviews, and avoids confusion in families with coexisting MODY and T1D or T2D [1]. At-risk relatives generally favour early predictive testing [1]. An asymptomatic carrier has a high diabetes or hyperglycaemia risk and requires a glucose surveillance plan, with annual HbA1c according to the Oxford manual; in HNF1A carriers, glycosuria is detectable in urine [2, 17]. A test-negative relative retains only common-diabetes risk except in mitochondrial diabetes, where variable tissue heteroplasmy means a negative blood test does not exclude risk [2].

Risk interpretation depends on variant identification context. In HNF1A carriers found in population cohorts, diabetes penetrance was only 32% versus 98% in referred probands, and HNF4A penetrance was also reduced outside clinical settings; GCK penetrance was 89–97% across contexts [30]. Authors propose counselling for HNF1A and HNF4A according to ascertainment and considering GCK for the ACMG secondary findings list [30]. ClinGen scored HNF1A diabetes actionability at 10 in adults and children, and HNF4A at 8 in adults and 10 in children, with specialist evaluation to guide sulfonylurea treatment as the intervention [97].

8.3. Predictive testing in minors

Predictive testing for HNF1A or HNF4A has minimal clinical relevance before adolescence and is generally discouraged in young children; adolescents from HNF1A families preferred to be tested in adolescence, with shared decision-making [2]. Testing asymptomatic children for GCK is not recommended, because the condition is benign and labelling carries psychosocial risks and implications for life, long-term care or disability insurance [2]. GCK-MODY often meets the glycaemic criteria for prediabetes, but lacks the cardiovascular risk and the progression to T2D of prediabetes; if a child from a GCK family has incidental hyperglycaemia, the paediatrician should be informed of the familial variant and the child tested, to avoid unnecessary treatment [2]. The Oxford manual emphasises issues of consent and autonomy in predictive testing of children [17]. The French PNDS sets at 10 years the age from which presymptomatic genetic testing for HNF1A and HNF4A is offered, points out that in France this test is subject to different legal rules for adults and minors, and excludes the use of a VUS for this purpose; in GCK, fasting glucose precedes testing [7]. The age of 10 years is compatible with the PDMI recommendation not to test young children and to decide jointly in adolescence, but is more explicit [2].

8.4. Communication of results

PDMI recommends including a genetic counsellor in the team and communicating a pathogenic or likely pathogenic result in a 30–60-minute consultation, in person or by telehealth, and not through the electronic health record portal or by non-clinical staff [2]. The consultation should explain the variant in plain language, the mechanism, the evidence for the classification, the inheritance pattern, variable expressivity and reduced penetrance; a copy of the report and a document for relatives should be provided, the frustration caused by a previous misdiagnosis acknowledged and peer support and the MDRAC patient consortium indicated [2]. The risks of testing include anxiety, unexpected information and insurance discrimination, with rules that vary by country; the benefits include more appropriate treatment, less anxiety and stigma and knowledge of the recurrence risk [2]. ISPAD asks that results be communicated to families clearly and unequivocally (grade E) [10]. The French PNDS asks that the disclosure be made by an experienced physician, in an individual consultation, with a close relative if the adult so wishes and, for a child, with the child and both parents whenever possible; the consultation covers the mode of transmission, the course, pregnancy, associated manifestations, the gene-specific surveillance and treatment strategy and the value of family investigation, and organises the care network with the attending physician [7]. In the HNF1A and HNF4A forms, it should be stressed that their apparently more benign character does not protect against complications if long-term control is poor [7]. Annual follow-up is recommended to update information on the diagnosis, prognosis, treatment and variant classification [2].

8.5. Prenatal diagnosis, non-invasive fetal genotyping and preimplantation diagnosis

Invasive prenatal diagnosis is rarely requested in MODY [17]. Non-invasive fetal genotyping now has a different indication, guiding maternal GCK-MODY and K_ATP diabetes pregnancy management (section 7.6), rather than reproductive decision-making [27, 66]. In neonatal diabetes, most cases arise from heterozygous de novo variants in non-consanguineous families, recessive forms predominate in consanguineous families and X-linked inheritance is possible (IPEX); prenatal diagnosis is reserved mainly for syndromic forms potentially fatal in childhood [17]. The sources consulted do not describe preimplantation genetic diagnosis use in MODY.

9. Special Considerations: Population Diversity and Epidemiology

9.1. Prevalence and what each figure measures

MODY prevalence estimates vary by country, age group and ascertainment method, and many are minimum values dependent on testing access [1]. Table 9.1 distinguishes what each study measures.

Table 9.1. Frequency estimates for MODY and its subtypes

The 1996 article that identified HNF1A estimated that MODY accounted for 2–5% of non-insulin-dependent diabetes, a pre-molecular-era value that should not be used as current prevalence [23]. None of the sources consulted directly measures the population prevalence of isolated HNF4A-MODY or of the ABCC8, KCNJ11 and INS forms. The French PNDS states that GCK-MODY could affect 1 in every 1,000 people with diabetes in certain populations, citing the Irish study of pregnant women, whose denominator is the population of screened pregnant women and not people with diabetes; the valid estimate is that based on the original denominator [7, 9].

For differential diagnoses, neonatal diabetes incidence is 1:90,000 to 1:260,000 live births, half permanent, according to permanent neonatal diabetes GeneReviews, and 1:215,000 to 1:400,000 according to transient disease GeneReviews [5, 67]. Orphanet gives birth prevalence 1.1/100 000 in Europe for all neonatal diabetes, 0.38/100 000 in Europe and 0.4/100 000 in the USA for isolated permanent disease, 18/100 000 in Oman in a consanguineous population and 0.3/100 000 for transient disease, the last by expert opinion [98]. Wolfram syndrome point prevalence is 0.13/100 000 in the United Kingdom and Japan [69]. Congenital hyperinsulinism incidence is 1:27,000 to 1:50,000 live births in non-consanguineous Western populations and up to 1:2,500 with consanguinity [40].

9.2. Geography, ancestry and founder effect

The distribution of subtypes depends on the population and the ascertainment context. GCK is proportionally more frequent in series from the USA, Germany, Italy, France and Spain, probably because of ascertainment bias between series of children and of adults [1]. In Italy, between 2017 and 2025, GCK accounted for 54% of diagnoses and the HNF1A and HNF4A forms for only 19%, unlike the United Kingdom, France and Norway, and syndromic diabetes genes accounted for 22.8% [58]. In the international SWEET registry, which brings together 44 paediatric centres and 1,462 young people with monogenic diabetes among 115,146 patients with diabetes, GCK accounted for 59% of cases, HNF1A for 18% and HNF1B for 7%, and GCK was the most prevalent in all regions except North America [52]. At diagnosis, the mean age was 10.1 years and HbA1c 6.9%; insulin was used at diagnosis in 12.9% of patients in Europe, 50% in Asia and the Middle East and 0% in North America, and HbA1c at 1 year was lowest in Europe (6.3%) [52]. The French PNDS attributes 20–60% of MODY to GCK, 20–35% to HNF1A and 10–15% to HNF4A, and explains the variation between cohorts by the proportion of paediatric and adult cases analysed [7].

The sources consulted describe no MODY founder variants with population effects. HNF1A c.872dup (P291fsinsC) is the most frequent variant in northern European populations, present in approximately 20% of HNF1A families, but arises in a repetitive tract and may occur de novo, and was found only once in 395 referred Greek patients [32]. In hyperinsulinism, recessive ABCC8 forms have founder variants in Ashkenazi Jewish and Finnish populations [36]. In consanguineous populations, recessive neonatal and monogenic diabetes forms predominate, and panels should include all recessive genes [2, 10].

Non-European populations are underrepresented: many racial and ethnic minority cases remain undiagnosed, most studies involved White populations, the probability calculator and genetic scores were developed in Europeans, and variant databases contain less information for other ancestries [1, 2]. In China, GCK sequencing in mild hyperglycaemia had 0.5–2% yield and specific HNF1A clinical criteria were proposed; in the USA, persistent mild hyperglycaemia yield was similar in Latino and Asian people [2]. PRISM in Hong Kong found monogenic diabetes in 23 of 884 Chinese adults with young-onset diabetes [82].

Brazil and Portugal. In Brazil, a study of suspected MODY families followed in São Paulo and other centres identified GCK variants in 61% of suspected GCK probands (25 variants in 30 probands, 12 novel) and HNF1A variants in 17% of suspected HNF1A probands (7 variants in 10 probands, 2 novel) [61], and PDMI records 13% yield in clinically suspected MODY in a Brazilian series [2]. A Brazilian genetic syndromes book presents MODY1–6 classification for teaching [25]. In Portugal, a Lisbon paediatric centre participates in SWEET, which provides no country-specific data, and a pregnancy GCK-MODY case followed in Lisbon was published in 2026 (section 10) [34, 52]. The sources consulted include no national prevalence estimates for Portugal or Brazil, which does not mean none exist.

9.3. Special populations

In pregnancy, GCK-MODY represents 1–6% of gestational diabetes in unselected European series and its identification changes management (section 7.6) [2]. In children, GCK-MODY is the commonest cause of persistent incidental hyperglycaemia, and decisions to test asymptomatic relatives should respect recommendations for minors (section 8.3) [2, 10]. In adults diagnosed with diabetes after 40 years, MODY is more frequent than previously thought but difficult to identify clinically [53]. Obesity does not exclude diagnosis [10]. In Africa, transition from multiple daily insulin doses to sulfonylurea was described in a young adult with HNF1A-MODY [99].

9.4. Reference networks, centres and patient organisations

The European Reference Network for rare endocrine diseases (Endo-ERN) includes monogenic diabetes in thematic group MTG3, “Genetic disorders of Glucose & Insulin Homeostasis”, subgroup “Rare diabetes mellitus” [16]. The Monogenic Diabetes Research & Advocacy Consortium (MDRAC) brings together patients, professionals and researchers [2, 100]. Exeter Diabetes Genes provides patient and professional information, genetic testing, the probability calculator, a curated gene list and urgent cord blood testing for K_ATP diabetes [39, 101]. The University of Chicago monogenic diabetes registry is the principal United States reference [10, 102]. ISPAD also lists Norwegian testing centres (www.mody.no), EURO-WABB, Bambino Gesù Hospital in Rome and Robert-Debré Hospital in Paris [10]. EMQN maintains the European external quality assessment scheme for laboratory diagnosis [2], and the ClinGen monogenic diabetes expert panel publishes variant classifications and gene-specific specifications [28]. General diabetes organisations include the American Diabetes Association, Diabetes UK, ISPAD and, in Portugal, Associação Protectora dos Diabéticos de Portugal (APDP), which is not MODY-specific [1, 17, 103].

Organisation in France. The 2022 PNDS was coordinated by the reference centre for rare disorders of insulin secretion and insulin sensitivity (PRISIS), based at Saint-Antoine Hospital (AP-HP, Sorbonne Université, Paris), part of the FIRENDO rare endocrine diseases filière [7]. The PRISIS competence centres, published in the Journal officiel of 15 December 2017, are located in Lille, Strasbourg, Nancy, Reims, Brest, Rennes, Rouen, Nantes, Angers, Paris (Cochin, Lariboisière, Pitié-Salpêtrière, Necker-Enfants malades and Robert-Debré), Bordeaux, Clermont-Ferrand, Lyon, Saint-Étienne, Marseille, Toulouse and Saint-Pierre (Réunion) [7]. PRISIS organises multidisciplinary consultation meetings for complex diagnostic or therapeutic cases [7]. Follow-up is coordinated by the attending physician and the endocrinologist-diabetologist or paediatric diabetologist, preferably in liaison with a reference or competence centre, with involvement of ophthalmology, nephrology, cardiology, gynaecology and obstetrics, hepatogastroenterology, genetics, nursing, nutrition, psychology, podiatry and social services as needed [7]. The attending physician ensures annual follow-up of GCK-MODY and makes sure that patients with HNF1A and HNF4A benefit from long-term condition status (affection de longue durée, ALD), with exemption from co-payment, from the departmental centre for persons with disabilities (MDPH) where relevant and from at least one annual diabetology consultation [7]. The associations indicated are Aide aux Jeunes Diabétiques (AJD), the Fédération Française des Diabétiques and the Association des femmes diabétiques; the PNDS also refers to Orphanet (ORPHA:552, which lists the laboratories), FIRENDO, the French societies of endocrinology (SFE), paediatric endocrinology and diabetology (SFEDP) and diabetes (SFD) and the PRISIS website [7].

10. Illustrative Clinical Case

10.1. Maltese HNF1A-MODY family presenting as T1D, T2D and gestational diabetes

Pace and colleagues described the first confirmed Maltese HNF1A-MODY family in 2019 [32]. The proband (II.1) was diagnosed with diabetes at 14 years and labelled presumed T1D after osmotic symptoms and recurrent boils, with glycosuria without ketonuria, glucose 16 mmol/L, HbA1c 9.9%, BMI 19.56 kg/m² and negative anti-GAD and islet cell antibodies [32]. She received biphasic isophane insulin, 12–16 U/day, with poor adherence (HbA1c up to 8%) and approximately 8 years later abandoned insulin and follow-up, never experiencing ketoacidosis [32]. She returned at 30 years, 11 weeks into her first pregnancy, with hyperglycaemia and glycosuria; she was treated with diet and low-dose insulin (HbA1c up to 7.9%) and had an elective caesarean at 36 weeks for polyhydramnios and macrosomia, with a 4,600 g newborn [32].

The father (I.1) had diabetes diagnosed at 44 years and labelled T2D (BMI 25.32 kg/m²), treated with metformin 500 mg twice daily and gliclazide 80 mg three times daily, with HbA1c up to 7.2% and peripheral arterial disease requiring intervention [32]. Two sisters presented with gestational diabetes. In sister II.3, glycosuria was detected at 37 weeks when aged 16, and OGTT showed fasting glucose 5.67 mmol/L, 12.08 mmol/L at 1 hour and 14.21 mmol/L at 2 hours (BMI 23.32 kg/m²); she received insulin 15 U/day from 38 weeks, hyperglycaemia persisted postpartum and the newborn weighed 3,790 g at 40 weeks [32]. In sister II.5, glycosuria was detected at 29 weeks when aged 19, and OGTT showed 8.64, 17.22 and 17.24 mmol/L; she received low-dose insulin and the newborn weighed 3,310 g at 37 weeks [32]. Fasting C-peptide was 2.1, 1.6 and 2.8 ng/mL, sisters' islet cell antibodies were negative and the mother had no diabetes [32]. The MODY calculator indicated high probability [32].

Exome sequencing in the proband and sister II.3 found no variants in GCK, HNF4A, HNF1B, PDX1, INS, NEUROD1, CEL, KCNJ11 or ABCC8, but identified heterozygous duplication NM_000545.6:c.872dup in the cytosine tract of HNF1A exon 4, confirmed by Sanger in all four affected individuals, absent from 300 Maltese controls and gnomAD v2.1, with a CADD (Combined Annotation Dependent Depletion) pathogenicity score of 27.7 [32]. On current MANE transcript, the variant is NM_000545.8:c.872dup p.(Gly292ArgfsTer25), classified pathogenic by a ClinVar expert panel [3, 4]. Article Figure 1 shows the pedigree with carrier father and three affected daughters (https://pmc.ncbi.nlm.nih.gov/articles/PMC6383084/figure/fig1-1179547619831034/), and Figure 2 shows exome read alignment and Sanger traces of the cytosine tract duplication (https://pmc.ncbi.nlm.nih.gov/articles/PMC6383084/figure/fig2-1179547619831034/) [32]. Authors emphasise that variant description is complicated by adjacent benign synonymous polymorphism c.864G>C, varying tract cytosine number between 8 and 9 [32].

After diagnosis, both sisters successfully switched to sulfonylureas (II.3 gliclazide 80 mg twice daily and II.5 gliclazide 80 mg daily), and the proband was lost to follow-up [32]. The case combines three mispresentations of HNF1A-MODY: T1D in a lean adolescent without antibodies or ketoacidosis despite insulin omission, T2D in a mildly overweight adult and gestational diabetes not resolving postpartum. The gliclazide doses described exceed ISPAD-recommended low initial doses (approximately one-quarter of adult dose, for example 20–40 mg/day) and should not be taken as routine [10]. The proband's child's 4,600 g weight is consistent with maternal hyperglycaemia since HNF1A itself does not increase birth weight [39].

10.2. Portuguese case: GCK-MODY and pregnancy

Bolas and colleagues in Lisbon published in 2026 a 33-year-old nulliparous woman with BMI 19.2 kg/m² and diabetes since 9 years, initially labelled T2D and treated with oral antidiabetics, stopped because of ineffectiveness [34]. Referred to endocrinology at 32 years, she maintained HbA1c 6.5–6.8% without drugs, had never used insulin and had no complications; diabetes occurred in successive generations (sister, mother, maternal grandmother, maternal uncles/aunts and a maternal female cousin with confirmed GCK-MODY) [34]. Sequencing identified heterozygous NM_000162.5:c.829_830insCGG p.(Leu276_Val277insAla), a previously undescribed in-frame insertion in a conserved region classified as VUS, and presumptive GCK-MODY diagnosis was made 23 years after presentation [34].

During pregnancy the couple declined invasive fetal testing and the fetus was managed as potentially non-carrier [34]. Insulin glargine was initiated at 7 weeks (5 U/day, fasting target <110 mg/dL) and increased at 8 weeks to 12 U/day with target <95 mg/dL, causing hypoglycaemic symptoms (tremor, sweating and hunger) with fasting glucose 102–144 mg/dL; at 11 weeks, glargine was reduced to 6 U/day, CGM and premeal insulin lispro started, with fasting targets 100–120 mg/dL and postprandial <140 mg/dL [34]. Time in range was 81–95% throughout pregnancy, fetal growth remained between 44th and 48th centiles and echocardiograms were normal; at 36+2 weeks, spontaneous membrane rupture and forceps delivery occurred, with a 2,970 g boy and Apgar 9/9/9, and insulin stopped postpartum [34]. The patient's mother and sister carried the same variant, results available only after delivery and supporting classification through segregation [34]. The article has no clinical figures, only a glycaemic data table.

The case illustrates three points: the difficulty of maternal GCK-MODY insulin therapy, with hypoglycaemic symptoms at glucose levels normal for the patient; managing a VUS with strong clinical suspicion; and lack of non-invasive fetal genotyping at the time, which would have established fetal genotype from 12 weeks [34, 66]. Starting insulin at 7 weeks without fetal genotype or ultrasound growth data differs from fetal abdominal circumference or cell-free fetal DNA guidance, reserving insulin for a non-carrier fetus [27, 92]. As a contrast in variant interpretation, a family with a rare synonymous HNF1A variant (c.1533G>A p.(Gln511=)), present in a boy, his sister and normoglycaemic mother, remained classified as VUS and controlled with diet alone for 4 years [104].

11. Prognosis and Long-Term Follow-Up

11.1. Prognosis by subtype

GCK-MODY. Prognosis is that of stable mild hyperglycaemia without relevant clinical consequences: in adults with median age 48.6 years, clinically significant microvascular complications occurred in 1% of carriers, similar to family controls, and only background retinopathy was more frequent, without laser treatment requirement [50]. After 50 years, there are no significant microvascular complications and macrovascular risk appears similar to the general population [35]. HbA1c remains stable untreated for years (follow-up 3 to 126 months) [14].

HNF1A-MODY. Prognosis depends on glycaemic control: microvascular and macrovascular complications are frequent with inadequate control, and timely sulfonylurea initiation was associated with a lower microvascular complication rate than T1D [10]. Overall and cardiovascular mortality are increased (HRs 1.9 and 2.3, respectively) despite an apparently protective lipid profile [46]. Sulfonylurea efficacy may diminish with diabetes duration and weight gain, and insulin may be necessary after several years [1, 14]. These data mainly derive from British and Irish cohorts receiving heterogeneous treatments and do not rigorously separate natural history from targeted treatment effects.

HNF4A-MODY. In an Irish cohort of 42 carriers, approximately half retained sulfonylurea monotherapy response after median 6 years, and retinopathy, nephropathy and coronary disease prevalence was 24.4%, 12.9% and 22.6% [51]. Neonatal hyperinsulinism usually remits in childhood but may persist into early adulthood [10, 27].

Previous hyperinsulinism. Children with HNF1A, HNF4A, GCK, ABCC8 or KCNJ11 hyperinsulinism entering remission may later develop glucose intolerance or diabetes, warranting lifelong glycaemic follow-up [36, 40].

11.2. Surveillance

International guidelines define few MODY-specific surveillance intervals and refer to general diabetes follow-up, with regular medical and nursing supervision to prevent complications and with cardiovascular risk assessment [17]. The 2022 French PNDS is the source that sets intervals by gene: in GCK-MODY, annual follow-up only; in the HNF1A and HNF4A forms, the schedule of other diabetes, with 3–4 visits per year in children and adolescents and 2–4 in adults (at least one with a specialist), each with weight, height, blood pressure, HbA1c, review of self-monitoring data, enquiry about hypoglycaemia and assessment of adherence, and an annual review of complications [7]. Self-monitoring of blood glucose is not systematic, but is indispensable with sulfonylureas, glinides or insulin, and continuous glucose monitoring applies to patients on intensified insulin therapy; the HbA1c target is individualised [7].

Table 11.1. Recommended surveillance

12. Summary and Final Considerations

MODY-type monogenic diabetes results from dominantly acting heterozygous variants in genes controlling beta-cell glucose sensing, transcription and insulin secretion. GCK, HNF1A and HNF4A explain most cases, and identifying the gene is clinically decisive because each form has its own natural history and treatment response [10, 11]. Gene name associated with phenotype should replace historical numbering, and several of the 14 initially proposed genes, such as BLK, KLF11 and PAX4, currently lack support as monogenic diabetes causes [2].

HNF1A-related diabetes is progressive secretory-defect diabetes, frequently confused with T1D or T2D, recognised through autosomal dominant family history, absent autoimmunity, preserved C-peptide, glycosuria, large postprandial rise and low hs-CRP [10, 47]. It responds markedly to low-dose sulfonylureas, since the 2003 trial showing a 5.2-fold greater fasting glucose reduction with gliclazide than metformin, and can often dispense with insulin [14, 29]. Increased cardiovascular risk justifies statin treatment no later than 40 years, and sulfonylurea efficacy loss with disease duration and weight gain requires intensification, for which GLP-1 agonists, DPP-4 inhibitors and, based on a 4-week 2026 RCT, empagliflozin are options with limited evidence [1, 46, 78].

GCK-related hyperglycaemia is a stable shift in glucose set point without relevant complications, which should not be treated outside pregnancy; during pregnancy treatment depends on fetal genotype, now ascertainable through non-invasive cell-free fetal DNA genotyping with 100% sensitivity and 96% specificity [10, 50, 66]. HNF4A-related diabetes adds macrosomia and neonatal hyperinsulinism and responds to sulfonylureas, although approximately half eventually require additional drugs [12, 51]. ABCC8, KCNJ11 and INS forms are rare in MODY but link it to neonatal diabetes, where high-dose sulfonylurea transition and the EU-authorised paediatric formulation changed prognosis [10, 73].

Diagnosis depends on an NGS panel including copy number variants, known non-coding regions and m.3243A>G, interpreted with ClinGen gene-specific specifications and reported with gene, transcript, variant and class [2, 28]. Selection combines age, family history, autoantibodies, C-peptide and the probability calculator, whose 2026 version incorporates biomarkers, but most cases remain undiagnosed and available tools were validated mainly in young Europeans [8, 53, 59]. Penetrance of HNF1A and HNF4A variants identified outside clinical settings is lower than in referred families, requiring counselling adapted to ascertainment [30]. Genetic counselling, cascade testing, avoidance of predictive testing in young children and annual follow-up at experienced monogenic diabetes centres complete the approach [2].

The 2022 French PNDS is the guideline that translates these principles into a care pathway with explicit intervals: annual follow-up of GCK-MODY by the attending physician; 2–4 annual visits in adults and 3–4 in children with HNF1A or HNF4A, HbA1c every three to four months and an annual review of complications; periodic liver ultrasound in HNF1A because of adenomatosis; genetic testing of relatives from 10 years in HNF1A and HNF4A and after fasting glucose >5.5 mmol/L in GCK; and insulin instead of sulfonylurea in pregnancy [7]. It diverges from the other sources mainly on SGLT2 inhibitors, which it does not advocate, and on sulfonylurea in early pregnancy, which it contraindicates [7, 39, 78].

13. References

1. Naylor R, Knight Johnson A, del Gaudio D. Maturity-Onset Diabetes of the Young Overview. 2018 May 24. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK500456/ PMID:29792621.
2. Murphy R, Colclough K, Pollin TI, Ikle JM, Svalastoga P, Maloney KA, et al. The use of precision diagnostics for monogenic diabetes: a systematic review and expert opinion. Commun Med (Lond). 2023;3(1):136. doi:10.1038/s43856-023-00369-8. PMID:37794142.
3. 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. Tool: https://variantvalidator.org. PMID:28967166.
4. National Center for Biotechnology Information. ClinVar [Internet]. Bethesda (MD): National Library of Medicine. Variation IDs 14927, 14943, 14934, 135665, 9212, 156152, 447418. Available from: https://www.ncbi.nlm.nih.gov/clinvar/
5. De León DD, Pinney SE. Permanent Neonatal Diabetes Mellitus. 2008 Feb 8 [Updated 2024 Nov 14]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1447/ PMID:20301620.
6. Orphanet. MODY (ORPHA:552). Paris: INSERM; [accessed 2026 Oct 10]. Available from: https://www.orpha.net/en/disease/detail/552
7. Haute Autorité de Santé; Centre de Référence des Pathologies Rares de l’Insulino-Sécrétion et de l’Insulino-Sensibilité (PRISIS). Protocole National de Diagnostic et de Soins (PNDS): Diabètes monogéniques de type MODY. Texte du 15 novembre 2022. Saint-Denis La Plaine: HAS; 2022 Dec 12 [accessed 2026 Nov 6]. Available from: https://www.has-sante.fr/jcms/p_3389585
8. Shields BM, Hicks S, Shepherd MH, Colclough K, Hattersley AT, Ellard S. Maturity-onset diabetes of the young (MODY): how many cases are we missing?. Diabetologia. 2010;53(12):2504-2508. doi:10.1007/s00125-010-1799-4. PMID:20499044.
9. Chakera AJ, Spyer G, Vincent N, Ellard S, Hattersley AT, Dunne FP. The 0.1% of the population with glucokinase monogenic diabetes can be recognized by clinical characteristics in pregnancy: the Atlantic Diabetes in Pregnancy cohort. Diabetes Care. 2014;37(5):1230-1236. doi:10.2337/dc13-2248. PMID:24550216.
10. 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.
11. American Diabetes Association Professional Practice Committee for Diabetes. 2. Diagnosis and Classification of Diabetes: Standards of Care in Diabetes-2026. Diabetes Care. 2026;49(Supplement_1):S27-S49. doi:10.2337/dc26-s002. PMID:41358893.
12. Pearson ER, Boj SF, Steele AM, Barrett T, Stals K, Shield JP, et al. Macrosomia and hyperinsulinaemic hypoglycaemia in patients with heterozygous mutations in the HNF4A gene. PLoS Med. 2007;4(4):e118. doi:10.1371/journal.pmed.0040118. PMID:17407387.
13. Bowman P, Flanagan SE, Edghill EL, Damhuis A, Shepherd MH, Paisey R, et al. Heterozygous ABCC8 mutations are a cause of MODY. Diabetologia. 2012;55(1):123-127. doi:10.1007/s00125-011-2319-x. PMID:21989597.
14. 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(1):145. doi:10.1038/s43856-024-00556-1. PMID:39025920.
15. Chung WK, Erion K, Florez JC, Hattersley AT, Hivert MF, Lee CG, et al. Precision Medicine in Diabetes: A Consensus Report From the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD). Diabetes Care. 2020;43(7):1617-1635. doi:10.2337/dci20-0022. PMID:32561617.
16. European Reference Network on Rare Endocrine Conditions (Endo-ERN). Main Thematic Group 3: Genetic disorders of glucose & insulin homeostasis. Available from: https://endo-ern.eu/
17. Firth HV, Hurst JA. Diabetes mellitus. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 390–393.
18. Tattersall RB. Mild familial diabetes with dominant inheritance. Q J Med. 1974;43(170):339-357. doi:10.1093/oxfordjournals.qjmed.a067391. PMID:4212169.
19. Tattersall RB, Fajans SS. A difference between the inheritance of classical juvenile-onset and maturity-onset type diabetes of young people. Diabetes. 1975;24(1):44-53. doi:10.2337/diab.24.1.44. PMID:1122063.
20. Froguel P, Vaxillaire M, Sun F, Velho G, Zouali H, Butel MO, et al. Close linkage of glucokinase locus on chromosome 7p to early-onset non-insulin-dependent diabetes mellitus. Nature. 1992;356(6365):162-164. doi:10.1038/356162a0. PMID:1545870.
21. Vionnet N, Stoffel M, Takeda J, Yasuda K, Bell GI, Zouali H, et al. Nonsense mutation in the glucokinase gene causes early-onset non-insulin-dependent diabetes mellitus. Nature. 1992;356(6371):721-722. doi:10.1038/356721a0. PMID:1570017.
22. Vaxillaire M, Boccio V, Philippi A, Vigouroux C, Terwilliger J, Passa P, et al. A gene for maturity onset diabetes of the young (MODY) maps to chromosome 12q. Nat Genet. 1995;9(4):418-423. doi:10.1038/ng0495-418. PMID:7795649.
23. Yamagata K, Oda N, Kaisaki PJ, Menzel S, Furuta H, Vaxillaire M, et al. Mutations in the hepatocyte nuclear factor-1alpha gene in maturity-onset diabetes of the young (MODY3). Nature. 1996;384(6608):455-458. doi:10.1038/384455a0. PMID:8945470.
24. Yamagata K, Furuta H, Oda N, Kaisaki PJ, Menzel S, Cox NJ, et al. Mutations in the hepatocyte nuclear factor-4alpha gene in maturity-onset diabetes of the young (MODY1). Nature. 1996;384(6608):458-460. doi:10.1038/384458a0. PMID:8945471.
25. Ribeiro EM, coord. Síndromes Genéticas: Reconhecer para Tratar. Fortaleza: EdUnichristus; 2021. p. 164–165.
26. Firth HV, Hurst JA. Hypoglycaemia in the neonate and infant. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 182–183.
27. Crowley MT, Paponette B, Bacon S, Byrne MM. Management of pregnancy in women with monogenic diabetes due to mutations in GCK, HNF1A and HNF4A genes. Front Genet. 2024;15:1362977. doi:10.3389/fgene.2024.1362977. PMID:38933924.
28. ClinGen Monogenic Diabetes Expert Panel. ClinGen Monogenic Diabetes Expert Panel Specifications to the ACMG/AMP Variant Interpretation Guidelines for HNF1A, Version 3.1. Criteria Specification Registry GN017. Available from: https://cspec.genome.network/cspec/ui/svi/doc/GN017
29. Pearson ER, Starkey BJ, Powell RJ, Gribble FM, Clark PM, Hattersley AT. Genetic cause of hyperglycaemia and response to treatment in diabetes. Lancet. 2003;362(9392):1275-1281. doi:10.1016/s0140-6736(03)14571-0. PMID:14575972.
30. Mirshahi UL, Colclough K, Wright CF, Wood AR, Beaumont RN, Tyrrell J, et al. Reduced penetrance of MODY-associated HNF1A/HNF4A variants but not GCK variants in clinically unselected cohorts. Am J Hum Genet. 2022;109(11):2018-2028. doi:10.1016/j.ajhg.2022.09.014. PMID:36257325.
31. Murray Leech J, Arni AM, Chundru VK, Sharp LN, Colclough K, Hattersley AT, et al. Polygenic background contributes to GCK-MODY clinical presentation and glycaemic variability. Diabetologia. 2026;69(7):1951-1961. doi:10.1007/s00125-026-06712-7. PMID:41848901.
32. Pace NP, Rizzo C, Abela A, Gruppetta M, Fava S, Felice A, et al. Identification of an HNF1A p.Gly292fs Frameshift Mutation Presenting as Diabetes During Pregnancy in a Maltese Family. Clin Med Insights Case Rep. 2019;12:1179547619831034. doi:10.1177/1179547619831034. PMID:30814848.
33. Juszczak A, Pavić T, Vučković F, Bennett AJ, Shah N, Pape Medvidović E, et al. Plasma Fucosylated Glycans and C-Reactive Protein as Biomarkers of HNF1A-MODY in Young Adult-Onset Nonautoimmune Diabetes. Diabetes Care. 2019;42(1):17-26. doi:10.2337/dc18-0422. PMID:30455330.
34. Bolas AF, Torres JO, Bogalho P, Silva-Nunes J. Personalized management of glucokinase-related monogenic diabetes (GCK-MODY) during pregnancy: a case report. Front Endocrinol (Lausanne). 2026;17:1859312. doi:10.3389/fendo.2026.1859312. PMID:42290839.
35. Chakera AJ, Steele AM, Gloyn AL, Shepherd MH, Shields B, Ellard S, et al. Recognition and Management of Individuals With Hyperglycemia Because of a Heterozygous Glucokinase Mutation. Diabetes Care. 2015;38(7):1383-1392. doi:10.2337/dc14-2769. PMID:26106223.
36. Gillis D. Nonsyndromic Genetic Hyperinsulinism Overview. 2003 Aug 19 [Updated 2024 Oct 31]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1375/ PMID:20301549.
37. ClinGen Monogenic Diabetes Expert Panel. ClinGen Monogenic Diabetes Expert Panel Specifications to the ACMG/AMP Variant Interpretation Guidelines for HNF4A, Version 4.0. Criteria Specification Registry GN085. Available from: https://cspec.genome.network/cspec/ui/svi/doc/GN085
38. Hamilton AJ, Bingham C, McDonald TJ, Cook PR, Caswell RC, Weedon MN, et al. The HNF4A R76W mutation causes atypical dominant Fanconi syndrome in addition to a β cell phenotype. J Med Genet. 2014;51(3):165-169. doi:10.1136/jmedgenet-2013-102066. PMID:24285859.
39. Shepherd M, Brook AJ, Chakera AJ, Hattersley AT. Management of sulfonylurea-treated monogenic diabetes in pregnancy: implications of placental glibenclamide transfer. Diabet Med. 2017;34(10):1332-1339. doi:10.1111/dme.13388. PMID:28556992.
40. 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 (sections 6.1–6.6).
41. De Franco E, Saint-Martin C, Brusgaard K, Knight Johnson AE, Aguilar-Bryan L, Bowman P, et al. Update of variants identified in the pancreatic β-cell KATP channel genes KCNJ11 and ABCC8 in individuals with congenital hyperinsulinism and diabetes. Hum Mutat. 2020;41(5):884-905. doi:10.1002/humu.23995. PMID:32027066.
42. Welch CL, Aldred MA, Balachandar S, Dooijes D, Eichstaedt CA, Gräf S, et al. Defining the clinical validity of genes reported to cause pulmonary arterial hypertension. Genet Med. 2023;25(11):100925. doi:10.1016/j.gim.2023.100925. PMID:37422716.
43. Billings LK, Shi Z, Resurreccion WK, Wang CH, Wei J, Pollin TI, et al. Statistical evidence for high-penetrance MODY-causing genes in a large population-based cohort. Endocrinol Diabetes Metab. 2022;5(6):e372. doi:10.1002/edm2.372. PMID:36208030.
44. Bacq Y, Jacquemin E, Balabaud C, Jeannot E, Scotto B, Branchereau S, et al. Familial liver adenomatosis associated with hepatocyte nuclear factor 1alpha inactivation. Gastroenterology. 2003;125(5):1470-1475. doi:10.1016/j.gastro.2003.07.012. PMID:14598263.
45. Sjöholm Å, Kåks E. Concurrent hepatic adenomatosis and hemangiomatosis in a patient with HNF1A MODY. BMJ Case Rep. 2025;18(12):e269815. doi:10.1136/bcr-2025-269815. PMID:41475867.
46. Steele AM, Shields BM, Shepherd M, Ellard S, Hattersley AT, Pearson ER. Increased all-cause and cardiovascular mortality in monogenic diabetes as a result of mutations in the HNF1A gene. Diabet Med. 2010;27(2):157-161. doi:10.1111/j.1464-5491.2009.02913.x. PMID:20546258.
47. McDonald TJ, Shields BM, Lawry J, Owen KR, Gloyn AL, Ellard S, et al. High-sensitivity CRP discriminates HNF1A-MODY from other subtypes of diabetes. Diabetes Care. 2011;34(8):1860-1862. doi:10.2337/dc11-0323. PMID:21700917.
48. Pihoker C, Gilliam LK, Ellard S, Dabelea D, Davis C, Dolan LM, et al. Prevalence, characteristics and clinical diagnosis of maturity onset diabetes of the young due to mutations in HNF1A, HNF4A, and glucokinase: results from the SEARCH for Diabetes in Youth. J Clin Endocrinol Metab. 2013;98(10):4055-4062. doi:10.1210/jc.2013-1279. PMID:23771925.
49. Steele AM, Wensley KJ, Ellard S, Murphy R, Shepherd M, Colclough K, et al. Use of HbA1c in the identification of patients with hyperglycaemia caused by a glucokinase mutation: observational case control studies. PLoS One. 2013;8(6):e65326. doi:10.1371/journal.pone.0065326. PMID:23799006.
50. Steele AM, Shields BM, Wensley KJ, Colclough K, Ellard S, Hattersley AT. Prevalence of vascular complications among patients with glucokinase mutations and prolonged, mild hyperglycemia. JAMA. 2014;311(3):279-286. doi:10.1001/jama.2013.283980. PMID:24430320.
51. Crowley MT, Ng N, Goulden E, Sanchez-Lechuga B, Ludgate S, Burke M, et al. Sulphonylurea efficacy and end-organ outcomes in the management of HNF4A-MODY. Diabet Med. 2025;42(12):e70139. doi:10.1111/dme.70139. PMID:40954556.
52. Deng L, Lanzinger S, Casteels K, Cavallo F, De Sanctis L, Fitas AL, et al. A worldwide perspective on clinical characteristics and treatment of youth with monogenic diabetes in the SWEET registry. J Endocr Soc. 2026;10(4):bvag064. doi:10.1210/jendso/bvag064. PMID:41958866.
53. Sharp LN, Mirshahi UL, Colclough K, Hall TS, Haley JS, Cannon SJ, et al. MODY Is Prevalent in Later-Onset Diabetes and Has Potential for Targeted Therapy but Is Challenging to Identify. Diabetes. 2026;75(2):361-368. doi:10.2337/db25-0545. PMID:41288518.
54. Besser RE, Shepherd MH, McDonald TJ, Shields BM, Knight BA, Ellard S, et al. Urinary C-peptide creatinine ratio is a practical outpatient tool for identifying hepatocyte nuclear factor 1-{alpha}/hepatocyte nuclear factor 4-{alpha} maturity-onset diabetes of the young from long-duration type 1 diabetes. Diabetes Care. 2011;34(2):286-291. doi:10.2337/dc10-1293. PMID:21270186.
55. Thanabalasingham G, Shah N, Vaxillaire M, Hansen T, Tuomi T, Gašperíková D, et al. A large multi-centre European study validates high-sensitivity C-reactive protein (hsCRP) as a clinical biomarker for the diagnosis of diabetes subtypes. Diabetologia. 2011;54(11):2801-2810. doi:10.1007/s00125-011-2261-y. PMID:21814873.
56. Murray Leech J, Sriram A, Colclough K, Hattersley AT, Houghton JAL, Laver TW, et al. Clinical Utility of a Type 1 Diabetes Genetic Risk Score Measured as Part of MODY Genetic Testing. Diabetes Care. 2026;49(10):1842-1849. doi:10.2337/dc26-1081. PMID:42612686.
57. Shields BM, McDonald TJ, Ellard S, Campbell MJ, Hyde C, Hattersley AT. The development and validation of a clinical prediction model to determine the probability of MODY in patients with young-onset diabetes. Diabetologia. 2012;55(5):1265-1272. doi:10.1007/s00125-011-2418-8. PMID:22218698.
58. Di Paola R, Rutigliano I, Fini G, Menzaghi C, Mansueto ML, Piscitelli P, et al. The landscape of monogenic diabetes in Italy. Comparisons with other European countries. Diabetes Res Clin Pract. 2026;239:113497. doi:10.1016/j.diabres.2026.113497. PMID:42600891.
59. Knupp J, Cardoso P, Young KG, McDonald TJ, Patel KA, Colclough K, et al. Updating a Clinical Prediction Model for Identifying Monogenic Diabetes to Include Both Clinical Features and Biomarkers. Diabetes Care. 2026;49(4):589-597. doi:10.2337/dc25-1029. PMID:41086125.
60. Shepherd M, Shields B, Hammersley S, Hudson M, McDonald TJ, Colclough K, et al. Systematic Population Screening, Using Biomarkers and Genetic Testing, Identifies 2.5% of the U.K. Pediatric Diabetes Population With Monogenic Diabetes. Diabetes Care. 2016;39(11):1879-1888. doi:10.2337/dc16-0645. PMID:27271189.
61. Santana LS, Caetano LA, Costa-Riquetto AD, Quedas EPS, Nery M, Collett-Solberg P, et al. Clinical application of ACMG-AMP guidelines in HNF1A and GCK variants in a cohort of MODY families. Clin Genet. 2017;92(4):388-396. doi:10.1111/cge.12988. PMID:28170077.
62. Kalyon O, Arumugam R, Patel D, Wright S, Russ-Silsby J, Donis R, et al. Incorporating mtDNA Structural Variant Analysis Into Routine Genome Sequencing Increases Monogenic Diabetes Diagnosis at No Additional Cost. Diabetes Care. 2026:dc261211. doi:10.2337/dc26-1211. PMID:42640178.
63. ClinGen Monogenic Diabetes Expert Panel. ClinGen Monogenic Diabetes Expert Panel Specifications to the ACMG/AMP Variant Interpretation Guidelines for GCK, Version 3.1. Criteria Specification Registry GN086. Available from: https://cspec.genome.network/cspec/ui/svi/doc/GN086
64. De Sousa SMC, Phan JMN, Wells A, Wu KHC, Scott HS. Improving detection of monogenic diabetes through reanalysis of GCK variants of uncertain significance. Acta Diabetol. 2025;62(8):1283-1289. doi:10.1007/s00592-025-02449-8. PMID:39821340.
65. Ellard S, Bellanné-Chantelot C, Hattersley AT; European Molecular Genetics Quality Network (EMQN) MODY group. Best practice guidelines for the molecular genetic diagnosis of maturity-onset diabetes of the young. Diabetologia. 2008;51(4):546-553. doi:10.1007/s00125-008-0942-y. PMID:18297260.
66. Hughes AE, Houghton JAL, Bunce B, Chakera AJ, Spyer G, Shepherd MH, et al. Bringing precision medicine to the management of pregnancy in women with glucokinase-MODY: a study of diagnostic accuracy and feasibility of non-invasive prenatal testing. Diabetologia. 2023;66(11):1997-2006. doi:10.1007/s00125-023-05982-9. PMID:37653058.
67. Greeley SAW, Del Gaudio D. 6q24-Related Transient Neonatal Diabetes Mellitus. 2005 Oct 10 [Updated 2026 Aug 13]. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1534/ PMID:20301706.
68. Orphanet. Maternally-inherited diabetes and deafness (ORPHA:225). Paris: INSERM; [accessed 2026 Oct 10]. Available from: https://www.orpha.net/en/disease/detail/225
69. Orphanet. Wolfram syndrome (ORPHA:3463). Paris: INSERM; [accessed 2026 Oct 10]. Available from: https://www.orpha.net/en/disease/detail/3463
70. Teva Pharmaceuticals USA. Glyburide tablets USP [prescribing information]. NDA017532. Revised 2021 Jul. DailyMed set id 50d16f1d-c814-4411-8749-62561321ecce. Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=50d16f1d-c814-4411-8749-62561321ecce
71. A-S Medication Solutions. Glimepiride tablets [prescribing information]. ANDA (Amaryl as reference product). Revised 2024 Sep. DailyMed set id 01196f54-32c8-41a0-b0ed-7968ca5d2fa2. Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=01196f54-32c8-41a0-b0ed-7968ca5d2fa2
72. Servier Laboratories Ltd. Diamicron 30 mg MR Tablets – Summary of Product Characteristics. electronic medicines compendium (emc); last updated 2020 May 11. Available from: https://www.medicines.org.uk/emc/product/1321/smpc
73. European Medicines Agency. Amglidia (glibenclamide) oral suspension – EPAR, product information. EMEA/H/C/004379. Marketing authorisation 2018 May 24; last updated 2026 May 11. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/amglidia
74. Stride A, Shields B, Gill-Carey O, Chakera AJ, Colclough K, Ellard S, et al. Cross-sectional and longitudinal studies suggest pharmacological treatment used in patients with glucokinase mutations does not alter glycaemia. Diabetologia. 2014;57(1):54-56. doi:10.1007/s00125-013-3075-x. PMID:24092492.
75. Carmody D, Naylor RN, Bell CD, Berry S, Montgomery JT, Tadie EC, et al. GCK-MODY in the US National Monogenic Diabetes Registry: frequently misdiagnosed and unnecessarily treated. Acta Diabetol. 2016;53(5):703-708. doi:10.1007/s00592-016-0859-8. PMID:27106716.
76. Servier Laboratories Ltd. Diamicron 80 mg Tablets – Summary of Product Characteristics. emc; last updated 2022 Aug 12. Available from: https://www.medicines.org.uk/emc/product/1150/smpc
77. American Diabetes Association Professional Practice Committee for Diabetes. 9. Pharmacologic Approaches to Glycemic Treatment: Standards of Care in Diabetes-2026. Diabetes Care. 2026;49(Suppl 1):S183-S215. doi:10.2337/dc26-s009. PMID:41358900.
78. Maagensen H, Høyerup SO, Jensen JS, Thuesen ACB, Forman JL, Thomsen HH, et al. Empagliflozin in HNF1A-MODY (MODY3): A Randomized, Double-Blind, Placebo-Controlled, Crossover Trial. Diabetes Care. 2026;49(7):1166-1174. doi:10.2337/dc25-2572. PMID:41525167.
79. Shepherd M, Pearson ER, Houghton J, Salt G, Ellard S, Hattersley AT. No deterioration in glycemic control in HNF-1alpha maturity-onset diabetes of the young following transfer from long-term insulin to sulphonylureas. Diabetes Care. 2003;26(11):3191-3192. doi:10.2337/diacare.26.11.3191-a. PMID:14578267.
80. Audrain E, Bel Lassen P, Bellanné-Chantelot C, Vantyghem MC, Dupuis H, Vidal-Trecan T, et al. Glycemic and Renal Effects of SGLT2 Inhibitors in Monogenic Diabetes: A Real-World National Study. Diabetes Obes Metab. 2026;28(10):9633-9643. doi:10.1111/dom.71169. PMID:42559753.
81. Bowman P, Mathews F, Barbetti F, Shepherd MH, Sanchez J, Piccini B, 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(1):35-42. doi:10.2337/dc20-1520. PMID:33184150.
82. Luk AOY, O CK, Fan B, Lau ESH, Fan Y, Lim C, et al. Precision Medicine to Redefine Insulin Secretion and Monogenic Diabetes (PRISM): A Randomized Controlled Trial in Chinese Patients With Young-Onset Diabetes. Diabetes Care. 2026;49(7):1202-1212. doi:10.2337/dc26-0114. PMID:42118608.
83. ClinicalTrials.gov. Evaluating a Novel, Allosteric Glucokinase Activator in Monogenic Diabetes Secondary to Inactivating Glucokinase Mutations: a Randomised, Cross-over Trial. NCT06976658. Available from: https://clinicaltrials.gov/study/NCT06976658
84. ClinicalTrials.gov. Effects of Dorzagliatin on 1st Phase Insulin and Beta-cell Glucose Sensitivity in Individuals With Recent-onset Type 2 Diabetes and Monogenic Diabetes. NCT04531631. Available from: https://clinicaltrials.gov/study/NCT04531631
85. Salazar S, Delgadillo-Silva LF, Carapeto P, Kenfaoui MM, Dakessian K, Melhem R, et al. Additive Effects of Dorzagliatin and Glucagon-Like Peptide 1 Receptor Agonism in a Novel Mouse Model of GCK-MODY and in Obese db/db Mice. Diabetes. 2026;75(1):99-114. doi:10.2337/db25-0520. PMID:41196662.
86. ClinicalTrials.gov. Switching From Insulin to Sulfonylurea in Childhood and Adult Diabetes Due to Variants in the HNF1A, HNF4A, or HNF1B Genes. NCT04239586. Available from: https://clinicaltrials.gov/study/NCT04239586
87. American Diabetes Association Professional Practice Committee for Diabetes. 15. Management of Diabetes in Pregnancy: Standards of Care in Diabetes-2026. Diabetes Care. 2026;49(Suppl 1):S321-S338. doi:10.2337/dc26-s015. PMID:41358885.
88. Caswell RC, Snowsill T, Houghton JAL, Chakera AJ, Shepherd MH, Laver TW, et al. Noninvasive Fetal Genotyping by Droplet Digital PCR to Identify Maternally Inherited Monogenic Diabetes Variants. Clin Chem. 2020;66(7):958-965. doi:10.1093/clinchem/hvaa104. PMID:32533152.
89. Schwitzgebel VM, Blouin JL, Dehos B, Köhler-Ballan B, Puder JJ, Rieubland C, et al. Enhancing fetal outcomes in GCK-MODY pregnancies: a precision medicine approach via non-invasive prenatal GCK mutation detection. Front Med (Lausanne). 2024;11:1347290. doi:10.3389/fmed.2024.1347290. PMID:38745742.
90. Jones AS, Kevat D, Lee IL, Goh K, Yates CJ. Clinical and cost-effectiveness of non-invasive cell-free DNA testing to optimise foetal outcomes for women with monogenic diabetes due to inactivating glucokinase gene mutations: a case series. Intern Med J. 2025;55(12):2059-2064. doi:10.1111/imj.70262. PMID:41263552.
91. ClinicalTrials.gov. Impact on Birth Weight of Two Therapeutic Strategies in GCK-MODY pregnancies. NCT02556840. Available from: https://clinicaltrials.gov/study/NCT02556840
92. Ciangura C, Seco A, Saint-Martin C, Ancel PY, Bouvet D, Jacqueminet S, et al. Pregnancy and neonatal outcomes in women with GCK-MODY: an observational study based on standardised insulin modalities. Diabetologia. 2025;68(5):981-992. doi:10.1007/s00125-025-06363-0. PMID:39971752.
93. Ren Q, Wang Z, Yang W, Han X, Ji L. Maternal and Infant Outcomes in GCK-MODY Complicated by Pregnancy. J Clin Endocrinol Metab. 2023;108(10):2739-2746. doi:10.1210/clinem/dgad188. PMID:37011183.
94. Gardner DS, Rama Chandran S, Sng GGR, Carmody D. Continuous Glucose Monitoring in Glucokinase-maturity-onset Diabetes of the Young Pregnancy: A Treatment Paradigm Change. JCEM Case Rep. 2025;3(8):luaf147. doi:10.1210/jcemcr/luaf147. PMID:40657236.
95. Szopa M, Zawadzka K, Kania M, Witek P, Cyganek K. Management and outcomes in pregnant patients with monogenic diabetes due to pathogenic variants in GCK and HNF1A genes. Front Endocrinol (Lausanne). 2025;16:1687596. doi:10.3389/fendo.2025.1687596. PMID:41409604.
96. Bowman P, Shepherd MH, Flanagan SE, Tonks J, Salguero-Bermonth M, Letourneau-Freiberg L, et al. Improved Neurodevelopment Following In Utero Sulfonylurea Exposure in a Patient With KCNJ11 Permanent Neonatal Diabetes: Future Implications for Targeted Treatment During Pregnancy. Diabetes Care. 2025;48(2):e10-e12. doi:10.2337/dc24-1862. PMID:39688637.
97. ClinGen Actionability Working Groups. Maturity Onset Diabetes of the Young Type I & Type III (HNF1A, HNF4A). Gene-condition report AC075; adult context version 2.0.2 and pediatric context version 1.0.1, 2026 Jan 13. Available from: https://actionability.clinicalgenome.org/ac/Adult/api/sepio/doc/AC075
98. Orphanet. Neonatal diabetes mellitus (ORPHA:224). Paris: INSERM; [accessed 2026 Oct 10]. Available from: https://www.orpha.net/en/disease/detail/224
99. Katte JC, Dehayem MY, Colclough K, Sobngwi E. Treatment switch from multiple daily insulin injections to sulphonylureas in an African young adult diagnosed with HNF1A MODY: a case report. J Med Case Rep. 2024;18(1):506. doi:10.1186/s13256-024-04850-3. PMID:39420387.
100. Monogenic Diabetes Research & Advocacy Consortium (MDRAC). Available from: https://www.mdrac.org/
101. University of Exeter / Royal Devon University Healthcare NHS Foundation Trust. Diabetes Genes – information for patients and professionals on genetic types of diabetes. Available from: https://www.diabetesgenes.org/
102. University of Chicago. Monogenic Diabetes. Available from: https://monogenicdiabetes.uchicago.edu/
103. Associação Protectora dos Diabéticos de Portugal (APDP). Portal da Diabetes. Available from: https://apdp.pt/
104. Carvalho R, Moreira B, Parente S. HNF1A-Maturity-Onset Diabetes of the Young (MODY) in a Family With a Rare Silent Variant: Clinical Challenges and Variable Expressivity. Cureus. 2026;18(6):e111076. doi:10.7759/cureus.111076. PMID:42472134.