17q12 Deletion and Duplication and HNF1B-Related Disease

17q12 recurrent deletion syndrome, 17q12 recurrent duplication syndrome and HNF1B-related disease due to intragenic variants (renal cysts and diabetes syndrome, RCAD; HNF1B-MODY diabetes, formerly MODY5; autosomal dominant tubulointerstitial kidney disease ADTKD-HNF1B): altered dosage or function of the HNF1β transcription factor, with developmental kidney disease, young-onset diabetes, hypomagnesaemia, pancreatic hypoplasia, genital malformations and, in copy number alterations of the entire region, neurodevelopmental disorders.

Literature and regulatory information updated: 9 October 2026.

0. Initial Reference Table

1. Introduction

1.1. From isolated clinical descriptions to a single spectrum

HNF1B-related disease was recognised from observations scattered across different specialties, which only molecular genetics brought together into one spectrum. In 1982, Rizzoni and colleagues described two Italian sisters with “familial hypoplastic glomerulocystic kidney”: small kidneys without papillae, with renal failure from the first months of life, in a family in which the mother had similar urographic abnormalities; Kaplan and colleagues reported a mother and son with the same presentation in 1989, and five of the mother's seven siblings had renal cysts on ultrasound [1]. These families were reassessed in 2001, when all affected individuals showed small kidneys with absent calyces, cortical cysts with cystic glomeruli and diabetes beginning between 17 and 39 years [1, 36].

The gene was associated with diabetes in 1997. Horikawa and colleagues, in a letter to Nature Genetics, described a Japanese family in which three siblings developed diabetes at 14, 10 and 15 years, with non-specific nephropathy, and identified a heterozygous variant in TCF2, the hepatocyte nuclear factor 1β (HNF1β) gene, in the two siblings studied; the mother was a carrier, whereas the father, also diabetic from age 50, was not, leading the authors to consider bilineal inheritance of different susceptibilities [1, 37]. In the numerical classification then in use, the new form was called MODY5, which assigned MODY1 to HNF4A, MODY2 to GCK, MODY3 to HNF1A, MODY4 to IPF1 (PDX1) and MODY5 to HNF1B [19, 38].

In subsequent years, the phenotype broadened beyond diabetes. In 1999, Lindner and colleagues described a Norwegian family with mild diabetes, progressive non-diabetic kidney disease and, in two of four affected women, vaginal aplasia and a rudimentary uterus, caused by a 75 bp deletion in the gene's pseudo-POU domain [1]. In 2000, Bingham and colleagues reported a de novo frameshift variant in exon 4 in a woman with diabetes at 21 years, small bilateral renal cysts and creatinine clearance of 66 mL/min, whose first pregnancy was terminated at 17 weeks because of non-functioning cystic fetal kidneys and whose first child had one multicystic dysplastic kidney and one dysplastic kidney [39]. In 2001, the same group identified HNF1B variants in two of four families with hypoplastic glomerulocystic disease, placing it within the RCAD spectrum [36]; in 2003, it described a family with autosomal dominant juvenile hyperuricaemic nephropathy and gout at 21, 23 and 27 years [1]. The designation “renal cysts and diabetes syndrome” was consolidated in a 2004 review [40].

In France, Bellanné-Chantelot and colleagues demonstrated the broad clinical spectrum in 2004 (renal agenesis, dysplasia and cystic disease, pancreatic atrophy in five of six probands, genital abnormalities and abnormal liver enzymes) and concluded that gene analysis is warranted even without a family history of diabetes [1, 41]. In 2005, the same team detected alterations in 28 of 40 patients with a MODY5 phenotype (70%), comprising 18 point variants, nine deletions of at least 1.2 Mb encompassing the entire gene and one exon 5 deletion; deletions were the most frequent alteration and the phenotype did not differ from that of point variants [1, 7]. In 2006, Edghill and colleagues found variants in 23 of 160 patients with kidney disease of unknown cause and emphasised that only 6 met classical MODY criteria, so diabetes is not a mandatory feature [1].

1.2. Recurrent 17q12 deletion and duplication

The deletion was identified as a recurrent genomic disorder in 2007. Mefford and colleagues used array comparative genomic hybridisation (aCGH) to study 155 fetal autopsies and found, in a fetus with multicystic dysplastic kidneys, a 17q12 deletion including TCF2; fine mapping in eight patients known to have gene deletions showed almost identical breakpoints within flanking segmental duplication blocks and allowed identification of the reciprocal duplication, enriched in patients with epilepsy [8, 42]. The authors described it as the first example of a recurrent genomic disorder associated with diabetes [42]. The duplication had been observed in 2006 by Sharp and colleagues, who found a 1.46 Mb duplication flanked by 66 kb segmental duplications with 99.7% identity in a patient with intellectual disability and an affected sibling [13].

Between 2008 and 2010, the deletion was associated with other phenotypes. Cheroki (2008) and Bernardini (2009) identified it in women with Müllerian aplasia, some without diabetes or renal malformation [8]. Loirat and colleagues (2010) described three boys with a de novo deletion, cystic or hyperechogenic kidneys and autism, in a cohort of 86 children with HNF1B abnormalities in which 61% had whole-gene deletions [8]. In the same year, Moreno-De-Luca and colleagues found the 1.4 Mb deletion in 18 of 15,749 patients with neurodevelopmental disorders referred for microarray and none of 4,519 controls, with replication in autism spectrum disorder (ASD) and schizophrenia (2/1 182 and 4/6 340, versus 0/47 929 controls), and concluded that this was a contiguous gene syndrome extending beyond RCAD [43]. In 2016, Clissold and colleagues showed that neurodevelopmental disorders occur in patients with the deletion but not in carriers of intragenic variants, removing HNF1B from primary responsibility for this component [44].

1.3. Current nomenclature and reclassifications

Numerical nomenclature (MODY1 to MODY14) is historical; current guidelines prefer gene-based designation (HNF1B-MODY or HNF1B diabetes) and recognise that carriers rarely have isolated diabetes [27, 28]. In nephrology, the 2015 Kidney Disease: Improving Global Outcomes (KDIGO) consensus proposed the term autosomal dominant tubulointerstitial kidney disease (autosomal dominant tubulointerstitial kidney disease, ADTKD) with gene-based subclassification, replacing designations such as medullary cystic kidney disease types 1 and 2 or familial juvenile hyperuricaemic nephropathy; the form due to HNF1B haploinsufficiency became ADTKD-HNF1B, and OMIM retains the synonym ADTKD3 [1, 4, 45]. GeneReviews uses “HNF1B-related kidney disease” for the combination of CAKUT, ADTKD-HNF1B and hypomagnesaemia [2]. Deletion size deserves mention: the 2005 diabetes literature described a “≥1.2 Mb” deletion defined by markers, whereas ClinGen defines the minimum recurrent interval as 1,396,450 bp (~1.4 Mb) in GRCh38; estimates of 1.3 to 1.8 Mb reflect probe density and segmental duplications near breakpoints [2, 5, 7, 46].

1.4. Definition of the article's entities

17q12 recurrent deletion syndrome is defined by the heterozygous ~1.4 Mb deletion in the GRCh38 chr17:36458167-37854616 interval (ClinGen ISCA-37432); the definition excludes deletions outside this region and biallelic deletions, and much larger or smaller deletions may have different phenotypes [2]. It is characterised by variable combinations of structural or functional kidney and urinary tract abnormalities, HNF1B-MODY diabetes and neurodevelopmental or neuropsychiatric disorders, without consensus clinical criteria [2]. 17q12 recurrent duplication syndrome corresponds to the heterozygous duplication of the same interval (GRCh38 chr17:36459259-37832869), with reduced penetrance and a predominantly neurodevelopmental phenotype [3]. HNF1B-related disease due to an intragenic variant encompasses sequence variants and small intragenic deletions or duplications compromising one gene copy; their identification establishes HNF1B disease and excludes 17q12 deletion syndrome [2]. HNF1B-MODY diabetes is the endocrine manifestation common to intragenic variants and whole-gene deletions and is addressed here in a dedicated subsection [19, 27].

2. Molecular Genetics and Pathophysiology

2.1. The HNF1B gene and HNF1β protein

HNF1B (HNF1 homeobox B; aliases TCF2, LFB3, VHNF1, MODY5) is located at 17q12 (GRCh38 chr17:37686431-37745059) and encodes the HNF1β transcription factor (UniProt P35680), whose MANE Select reference transcript is NM_000458.4 (NP_000449.1) [5, 6, 9]. The protein has an N-terminal dimerisation domain, a bipartite DNA-binding domain formed by the POU-specific domain (POUs) and POU homeodomain (POUh), encoded predominantly by exons 2 to 4, and a C-terminal transactivation domain [10, 46]. It binds DNA as a homodimer or heterodimer with HNF1α, is coactivated by PCBD1 and regulates promoters of developmental genes: in the kidney, FXYD2 (γ-subunit of Na⁺/K⁺ ATPase) and PKHD1; in the pancreas, PDX1, SOX9 and HNF6 [6, 46, 47]. It is expressed in the kidney, pancreas, liver and biliary tract, genital tract and parathyroid, where it represses parathyroid hormone (PTH) transcription [2, 6].

2.2. Tissue and physiological mechanisms

In the developing kidney, conditional loss of Hnf1b in mice alters ureteric bud branching and nephrogenesis, explaining the CAKUT spectrum, cystic dysplasia and hypoplasia [46]. In the mature kidney, postnatal loss in the tubular epithelium causes rapid tubulointerstitial injury, cell cycle re-entry, replication stress and fibrosis; in 2026, Isnard and colleagues showed that chronic kidney disease (CKD) stresses epigenetically reduce HNF1β activity in a reinforcing cycle and that, in transcriptomes from approximately 900 patients with common CKD, lower HNF1β target activity was associated with worse kidney function and more tubular atrophy and fibrosis [46, 48]. These are preclinical and correlational data, without a derived therapy [48].

Hypomagnesaemia results from reduced FXYD2 transactivation in the distal convoluted tubule, with renal magnesium wasting and hypocalciuria; Adalat and colleagues demonstrated in 2009, using a luciferase assay, that HNF1β activates FXYD2 through conserved recognition sites [47, 49]. Distal tubular dysfunction may produce a Gitelman syndrome-like phenotype, with hypokalaemia and hypomagnesaemia [46, 50]. Hyperuricaemia and early gout reflect reduced tubular urate clearance, already documented in the 2003 family [1].

In the pancreas, conditional inactivation in progenitors causes severe hypoplasia, with reduced progenitor expansion, abnormal ducts and absence of endocrine precursors; a mouse model reproducing human intron 2 splice variants has ~30% less β-cell mass, reduced insulin content and glucose intolerance resembling adolescent-onset diabetes [46]. Human diabetes combines impaired insulin secretion due to hypoplasia and β-cell dysfunction with reduced hepatic insulin sensitivity (endogenous glucose production) and normal peripheral sensitivity, explaining the inadequate response to sulphonylureas (SU), unlike HNF1A-MODY [19, 27, 51]. Dorsal pancreatic agenesis (absence of body and tail) and reduced exocrine function involve ductal and acinar cells [27, 52]. In the liver, loss of Hnf1b in hepatobiliary progenitors causes intrahepatic bile duct paucity and ductal plate malformation [46]. In the parathyroid, loss of PTH repression causes hyperparathyroidism independent of renal failure [2, 53]. In the genital tract, involvement of the Müllerian and Wolffian ducts explains uterine and vaginal malformations and those of the epididymis and vas deferens [1, 46].

2.3. The 17q12 region and contiguous genes

The 17q12 region is a CNV hotspot: it is flanked by highly homologous segmental duplications that misalign during meiosis, and non-allelic homologous recombination (non-allelic homologous recombination, NAHR) between them generates reciprocal deletions and duplications of the interval [2, 3, 42]. The interval contains 15 unique genes (listed in section 0) [2, 33]. HNF1B haploinsufficiency is established as the cause of renal, urogenital and endocrine abnormalities, but no single gene explains the overall phenotype [2]. LHX1 (LIM-homeobox transcription factor, OMIM *601999) is essential for renal and urogenital development; likely pathogenic heterozygous variants have been described in women with Müllerian aplasia, and its role in neural development may contribute to the neurodevelopmental component [2, 18]. For ACACA (acetyl-CoA carboxylase 1, lipogenesis), haploinsufficiency is speculated to contribute to the leaner phenotype and lower risk of diabetic glomerular disease observed in the deletion [2, 54]. In duplication, the responsible gene or genes are unknown; ACACA, LHX1 and HNF1B are genes of interest, without any having been demonstrated to be independently causal [3].

ClinGen classifies the ISCA-37432 region as haploinsufficiency 3 and triplosensitivity 3 (sufficient evidence for both; assessment of 16 June 2022), and the HNF1B gene as haploinsufficiency 3 and triplosensitivity 0 (no evidence), placing duplication pathogenicity at the regional level rather than in HNF1B triplication [5]. GeneGuidance curation indicates that the ClinGen actionability report for RCAD (AC1048) assigns strong adult and moderate paediatric actionability [6]. In UK Biobank, common variants in HNF1B itself were associated with estimated glomerular filtration rate (eGFR; rs12601991, NC_000017.11:g.37741642T>G, p = 4 × 10⁻²¹) and diabetes (rs7501939, g.37741165C>T, p = 6 × 10⁻¹⁷), and a second locus in the region was associated with fluid intelligence (rs1005552) and eGFR (rs12150665), supporting contributions from several genes in the interval [21].

2.4. Variant spectrum

HNF1B alterations comprise whole-gene deletions (almost always the recurrent 17q12 deletion), small intragenic deletions of one or more exons and sequence variants [2, 19]. In the European multicentre cohort of the European Rare Kidney Disease Reference Network (ERKNet), with 508 analysable patients, 340 (67%) had the 17q12 deletion and 168 intragenic variants; among the latter, 67% were in the DNA-binding domains (POUs and POUh), 19% in the transactivation domain, 74% were nucleotide substitutions, 47% missense and 20% nonsense; deletion frequency in the literature ranges from 56 to 59% [10]. The proportion of deletions depends on the ascertainment route and on systematic use of CNV testing, which explains the differences between the estimate of “~50% deletions >1 Mb”, that of “one third gene or exon deletions” in patients with a MODY5 phenotype and that of 61% in a paediatric cystic kidney cohort [8, 19, 23]. The 2020 French PNDS records more than 220 pathogenic variants, generally private and concentrated in exons 1 to 4, attributes 50–60% of cases to the deletion and describes it as removing HNF1B and 19 other genes (GeneReviews lists 15 genes in the region, including HNF1B); the consequence common to both classes is haploinsufficiency [2, 12].

Source: Heneberg's (2026) synthesis of series from various origins; frequencies vary with referral route [46].

Proposed mechanisms include haploinsufficiency, dominant-negative interference by certain missense variants and altered nuclear localisation [46]. In transactivation assays in HeLa and MIN6 cells, Pavithram and colleagues showed that all truncating variants and the in-frame deletion p.(Arg137_Lys161del) had ≤20–30% activity, missense variants in the dimerisation or DNA-binding domains (p.Gly20Arg, p.Gly287Val, p.Asn289Lys, p.Arg295Cys) had ≤50–55%, and activity correlated inversely with clinical burden measured by the HNF1B score (ρ = −0.701; 95% CI −0.887 to −0.318; 17 variants); variants in the linker segment or transactivation domain (p.Val2Leu, p.Pro60Arg, p.Asn327Lys, p.Pro343Ser, p.Ser362Phe) had normal function and were not accompanied by HNF1B traits [22]. With these data, three variants were reclassified as likely benign (p.Pro60Arg, p.Asn327Lys, p.Pro343Ser) and three changed from uncertain significance (VUS) to likely pathogenic (p.Gly20Arg, p.Gly287Val, p.Asn289Lys) following clinical follow-up and segregation [22].

Somatic mosaicism may attenuate the phenotype or restrict it to one organ and complicates counselling [46]. Balanced structural variants also cause disease: in 2026, whole-genome sequencing identified a de novo inversion GRCh38:17:g.36934029_37729559inv interrupting HNF1B in a patient with cystic kidney disease from the fetal period [55].

The following table compiles variants cited in this article, normalised to the MANE transcript with corresponding historical nomenclature.

Sources: normalisation using VariantValidator (c. numbering matches between NM_000458.3 and NM_000458.4); ClinVar classification accessed on 9 October 2026 [9, 22, 47, 56, 57].

2.5. Genotype–phenotype correlations

In the kidney, early series found no difference between deletions, missense and truncating variants, but larger studies converge on a milder renal phenotype with deletion [2]. In the French cohort of 201 adults, stage 3–4 CKD or end-stage kidney disease (ESKD) was less frequent in deletion carriers, both at diagnosis (11/43 versus 27/35; p < 10⁻⁴) and long term (40/78, 51%, versus 71/91, 78%; p = 0.0003) [54]. In the ERKNet cohort of 521 patients, stage 3 CKD was reached by 12% of deletion carriers versus 39% of intragenic variant carriers (hazard ratio, HR, 0.29; 95% CI 0.19–0.44), with eGFR in the first week of life of 68 versus 46 mL/min/1.73 m² and no difference in subsequent slope; among intragenic variants, those in POUh (HR 0.15) and the transactivation domain (HR 0.25) progressed more slowly than those in POUs, and variant type was not associated with renal survival [10]. Proposed explanations are a dominant-negative effect of certain missense variants (p.Leu168Pro in POUs showed dominance in vitro; Hnf1b+/− mice are phenotypically normal, unlike heterozygous carriers of a splice variant) and a protective effect from loss of other regional genes [2, 10]. Two recent paediatric cohorts, in Greece (20 children) and China (26 children), found no significant associations, with trends in the same direction [58, 59].

The French PNDS (2020) states that, among point variants, type and location do not correlate with presentation or course, and that in adults the deletion is associated with slower renal progression, lower BMI and more severe diabetes, with a similar frequency of diabetes in both groups [12]. The absence of a correlation by location predates the 2024 ERKNet cohort, which found faster progression with POUs variants than with POUh and transactivation domain variants [10].

Heneberg's review (2026) classifies the 17q12 deletion as “severe” in a genotype–phenotype pattern table, referring to multisystem and neurodevelopmental burden; this classification does not contradict the renal cohorts, in which deletion has better renal survival than intragenic variants [10, 46, 54]. For hypomagnesaemia, the relationship is reversed: it was more frequent with deletion (HR 2.12; 95% CI 1.32–3.50), whereas hyperuricaemia was less frequent (HR 0.57) [10]. In diabetes, clinical burden was comparable between deletion and variants in the Norwegian registry, and hyperglycaemia was not significantly associated with genotype in the ERKNet cohort [10, 22]. For hepatic and exocrine pancreatic manifestations, there is no significant genotype–phenotype correlation [2].

In neurodevelopment, Clissold and colleagues (38 patients with HNF1B kidney disease) found a clinical diagnosis of a neurodevelopmental disorder in 8/20 (40%) deletion carriers and 0/18 intragenic variant carriers, and two geneticists predicted deletion from facial dysmorphism with 83% sensitivity and 79% specificity [44]. A 2023 review estimated neurodevelopmental disorders in ~25% of deletion carriers and 6.8% of intragenic variant carriers, so HNF1B contributes in part, but its isolated haploinsufficiency does not appear sufficient [2, 60].

2.6. Penetrance and expressivity

Deletion is described as highly penetrant for the combination of renal and multisystem manifestations, without clear reports of carrier parents lacking any manifestation, but with substantial intrafamilial variability [2]. The Bayesian estimate by Rosenfeld and colleagues (2013), of 34.4% (95% CI 13.7–70.0) for deletion and 21.1% (95% CI 10.6–39.5) for duplication, measures something different: the probability of a carrier having the paediatric phenotype (developmental delay, intellectual disability or congenital anomalies) prompting postnatal microarray, above a baseline risk of ~5% [61–63]. The two figures are not comparable. Diabetes penetrance depends on age: ~36% in paediatric populations, ~60% in mixed populations and close to 80% in adults, according to data compiled by Kołbuc and colleagues [47]. Duplication has reduced and unestablished penetrance; most are inherited from a mildly affected or asymptomatic parent [3].

Expressivity varies even within families: in a Portuguese series, a daughter with CKD from childhood and diabetes ten years later had the same deletion as her mother, who developed diabetes only at 42 years and was labelled as having type 2 diabetes [57]. Three Italian siblings with the same familial deletion ranged from mild renal and neurodevelopmental traits to multisystem involvement [64].

2.7. Allelic disorders and modifiers

Allelic HNF1B disorders are transient or, rarely, permanent neonatal diabetes (HNF1B explains less than 1% of permanent neonatal diabetes), hypoplastic glomerulocystic disease, atypical familial juvenile hyperuricaemic nephropathy and type 2 MRKH syndrome (aplasia of the uterus and upper two thirds of the vagina with another malformation), in which Orphanet lists HNF1B with a loss-of-function mechanism [1, 65, 66]. OMIM also associates HNF1B with renal cell carcinoma (somatic clear cell and chromophobe forms) and susceptibility to type 2 diabetes, the latter consistent with common variant associations in UK Biobank [6, 21]. Predicted pathogenic variants in six CAKUT genes (HNF1B, SALL1, EYA1, PBX1, GATA3, PAX2) together have a population frequency between 1/249 and 1/1 263 in the Genome Aggregation Database (gnomAD) v2.1.1, suggesting incomplete penetrance [67]. No validated genetic modifiers of HNF1B disease are known; inclusion of ACACA in the deletion and the dominant-negative effect of certain variants are hypotheses explaining variability [2, 10].

3. Clinical Spectrum and Manifestations by System

3.1. Overview and evolution with age

The presentation of HNF1B disease broadens with age. In the fetus and newborn, hyperechogenic or cystic kidneys predominate; in childhood, neurodevelopmental disorders (with deletion), hypomagnesaemia and hyperuricaemia emerge; in adolescence and adulthood, diabetes, gout, liver abnormalities, exocrine pancreatic insufficiency, hyperparathyroidism and progressive CKD appear [2, 46]. Isolated diabetes is rare and developmental kidney disease is present in almost all carriers, being the main presentation in children [27]. In the ERKNet cohort, 58% of diagnoses were prenatal (63% with deletions and 49% with intragenic variants), with median gestational age of 24 weeks, and median age at postnatal diagnosis was 3 years [10].

GeneReviews compiled 282 individuals with deletion from 42 studies, excluding cohorts ascertained through a given manifestation when calculating that manifestation's frequency, and organised frequencies into three groups [2]:

3.2. Kidneys and urinary tract

Kidney abnormalities are the most consistent trait and often the sentinel manifestation [46]. In deletion, cystic dysplasia and other structural abnormalities were described in 167/193 (87%) individuals not ascertained through renal cohorts; cystic dysplasia is the most common finding, followed by poor corticomedullary differentiation, collecting system abnormalities (duplication, hydronephrosis, pyelectasis, vesicoureteric reflux, dilated ureter), a solitary kidney due to agenesis or involution of a multicystic dysplastic kidney, and horseshoe kidney [2]. Prenatal imaging predominantly shows cysts or echogenic kidneys, but in 35% of cases findings appear only in childhood or later [2, 68]. In a series of 62 pregnancies with bilateral hyperechogenic fetal kidneys, 18 (29%) had HNF1B abnormalities, 15 from complete deletion; amniotic fluid and kidney size were generally normal, prenatal cysts were present in 11 (unilateral in 8) and, during the first year, cysts appeared in 17 of 18, increasing in number, becoming bilateral and accompanied by reduced renal growth [69]. In the ERKNet cohort, ultrasound at diagnosis showed cortical cystic disease or hyperechogenic kidneys in more than 80%, bilateral lesions in 89% and a single functioning kidney in 18% [10].

The French PNDS describes the fetal ultrasound pattern: hyperechogenic kidneys are brighter than the liver (right kidney) or the spleen (left kidney), of normal or slightly increased size, with corticomedullary differentiation generally preserved; the hyperechogenicity corresponds to cystic dilatations too small to be seen, and visible cysts measure less than 10 mm and are mainly peripheral cortical [12]. Unilateral agenesis, hypoplasia (size below −2 standard deviations or the 3rd centile), dysplasia, ectopia and urinary tract dilatation may coexist; involvement is generally bilateral and asymmetric, and unilateral hyperechogenicity is often accompanied by another abnormality in the contralateral kidney [12]. Amniotic fluid is usually normal, but may be reduced in severe forms with fetal renal failure or increased (polyhydramnios) owing to fetal polyuria of tubular origin [12]. In the most severe forms, fetal pathology has found pancreatic agenesis or hypoplasia associated with multicystic or hypoplastic kidneys with severe oligohydramnios or anhydramnios; HNF1B variants have been identified in apparently isolated severe fetal kidney disease, but never in bilateral renal agenesis [12]. After birth, cortical cysts become more numerous and bilateral and the kidneys grow little, to the point of hypoplasia; in children diagnosed only after birth, hypoplastic and microcystic kidneys predominate, and the course of renal failure in childhood appears to depend on the initial reduction in nephron number [12].

The renal phenotype also includes glomerulocystic disease, oligomeganephronia, hypoplasia and interstitial fibrosis [1, 10]. Tubulointerstitial disease manifests as reduced concentrating ability, bland urinary sediment, absent or minimal albuminuria, hyperuricaemia, hypomagnesaemia, hypokalaemia and slowly progressive CKD, with interstitial fibrosis and tubular atrophy on biopsy, which is not routinely indicated [2]. In 27 French adults from 20 families, the profile was tubulointerstitial with slow decline throughout adulthood: cysts (up to five per kidney) in 15/24, solitary kidney in 5/24, hypokalaemia in 11/24, hypomagnesaemia in 10/16, renal Fanconi syndrome in 2 and ESKD in 4 [70]. Unlike the symmetrical cystic enlargement of autosomal dominant polycystic kidney disease (ADPKD), kidneys are normal-sized or small, with hypoplasia more typical in adults; many patients have progressive CKD without significant proteinuria or hypertension [19, 46].

In adults, the French PNDS summarises the nephropathy as follows: it is present in ~98% of adults with HNF1B disease, precedes the diagnosis of diabetes in ~40% and is detected simultaneously in 25%; the most common nephrological presentation is CKD in young adults with a tubulointerstitial profile, without proteinuria, hypertension in fewer than 10%, kidneys of normal or reduced size and a mean annual GFR decline of approximately 2 mL/min/1.73 m², and the median age at ESKD is around 50 years [12]. Cysts, present in ~80% at the transition from adolescence to adulthood, are generally bilateral and few in number and, unlike ADPKD, do not increase in number or in kidney volume in young adults; additional findings include solitary kidney (12–20%), pelviureteric junction syndrome (16%) and, more rarely, horseshoe kidney [12]. Renal biopsy is not useful for diagnosis; when performed, it shows oligomeganephronia, glomerulocystic disease or tubular cystic dysplasia [12].

Severity ranges from severe prenatal renal failure to lifelong normal function, through slow progression to ESKD in adulthood; ESKD in childhood is uncommon [2]. In the ERKNet cohort (median follow-up 11 years, from birth), 21% reached stage 3 CKD; survival without stage 3 CKD at 18 years was 87% (95% CI 81–93) with deletion and 63% (95% CI 54–73) with intragenic variants, and ESKD-free survival at 18 years was 97% and 86%, respectively [10]. With deletion, a single functioning kidney accelerated progression (HR 2.32; 95% CI 1.04–5.17) [10]. In the French cohort of 201 adults, stage 3–4 CKD was present in 75/169 (44%) and ESKD in 36/169 (21%) [54]. In the Greek paediatric cohort, mean eGFR remained stable over a median of 5 years (84.5 to 84.0 mL/min/1.73 m²), with substantial individual variability; in the Chinese cohort, 3 of 5 children with stage 5 CKD underwent transplantation before 15 years [58, 59].

Source: synthesis of clinical cohorts; frequencies vary with investigation intensity [46].

3.3. Electrolytes, uric acid and parathyroid

Hypomagnesaemia from renal wasting is common and may be the initial and predominant manifestation [2, 71]. It was found in 46/107 (43%) deletion carriers and may be severe; with age-adjusted paediatric reference values, frequency in children reaches 59% [2, 72]. In a London cohort of 91 children with renal malformations, hypomagnesaemia (<1.58 mg/dL) was present in 8/18 variant carriers (44%) versus 1/48 (2%) non-carriers, with median magnesium of 1.68 versus 2.02 mg/dL, hypermagnesiuria and hypocalciuria; the case prompting the study was an adolescent with tetany [49]. In the ERKNet cohort, hypomagnesaemia (<0.6 mmol/L on at least one measurement) occurred in 29% during follow-up and was more frequent with deletion [10]. Hyperuricaemia (>320 µmol/L) occurred in 65% in the same cohort and early-onset gout, including in young women, forms part of the presentation [1, 10]. In the family with hyperuricaemic nephropathy described in 2003, all patients improved kidney function with allopurinol [1].

The French PNDS places hypomagnesaemia, the most suggestive electrolyte abnormality, at 25–44% of children and 60–75% of adults, higher frequencies than in the ERKNet cohort, which used the lower threshold of 0.6 mmol/L; it is almost always asymptomatic, only exceptionally presents with fasciculations, arrhythmia or seizures, and symptom severity does not correlate with the plasma value [10, 12]. Hypokalaemia occurs in ~40% of adults and may, paradoxically, persist in advanced or end-stage CKD; the two abnormalities frequently coexist and result from tubular wasting, with inappropriately high urinary excretion of magnesium or potassium [12]. In children, hyperuricaemia and hypocalciuria are very common from the first months of life; in adults, early gout (before the age of 40) is very rare, and a Toni-Debré-Fanconi syndrome (diffuse proximal tubular dysfunction) has been described exceptionally [12].

Hyperparathyroidism was found in 20/45 (44%) deletion carriers with PTH measured; one report described transient neonatal hypercalcaemia and hypophosphataemia [2]. Elevated PTH persisted in some patients after kidney transplantation and magnesium normalisation; although hyperparathyroidism persists in 20–50% of transplant recipients from any cause, the authors concluded that HNF1B haploinsufficiency causes hyperparathyroidism independent of renal failure [2, 53].

3.4. HNF1B-MODY diabetes

Diabetes is the most common extrarenal manifestation and generally develops after identification of childhood kidney disease [19]. Mean age at onset is 24 years according to one series (reported range 0 to 61 years) and 26 years according to another, with typical onset in the second or third decade, later than HNF1A and HNF4A forms [19, 23, 24, 28]. In deletion, diabetes or abnormalities of blood glucose or insulin response were present in 58/155 (37%) individuals not ascertained through diabetes cohorts, a value underestimating lifetime prevalence because it includes many children; including diabetes cohorts, prevalence was 50%; diabetes is generally diagnosed before 25 years (range 10–50) [2]. In UK Biobank, deletions were strongly associated with diabetes (p = 2 × 10⁻⁷), unlike duplications [21].

The classical phenotype is non-autoimmune diabetes without obesity, with residual insulin secretion at diagnosis but progressive β-cell failure, relatively stable glycaemic control and low ketoacidosis risk [2, 46]. Birth weight is reduced (median 2.6 kg), contrasting with GCK-MODY (median 3 kg) and HNF4A-MODY (median 4.3 kg), and intrauterine growth restriction is among the distinguishing traits [19, 23]. In the French cohort of 201 adults, 159 had diabetes, 67/144 had symptoms at diagnosis, 64/102 already had known HNF1B kidney disease when diabetes was diagnosed, 29/51 tested responded to SU or repaglinide, but 111/140 (79%) were receiving insulin at follow-up (median diabetes duration 12 years); retinopathy or neuropathy were present in 46/114, and deletion carriers were leaner and more frequently insulin-treated [28, 54]. Microvascular complications are common and relate to overall glycaemic control [19].

The French PNDS, based mainly on French cohorts, adds presentation data. In series recruited through nephrology, diabetes affects 1–25% of adolescents and 40–50% of adults; among patients tested for diabetes suggestive of MODY without known kidney disease, HNF1B accounts for 1–6% of genetically confirmed MODY [12]. The mean age at diagnosis is 24–28 years, with 75% of cases after the age of 20; in patients with both manifestations, diabetes is diagnosed after the kidney disease in 35%, before it in 40% and simultaneously in 25% [12]. At diagnosis, 80% have normal weight, only half have symptoms (polyuria, weight loss), ketoacidosis is rare (5%) and initial glycaemic control is highly variable: one third have HbA1c ≤7% and another third ≥13% [12]. Insulin secretion is reduced to a degree similar to type 2 diabetes and persists in the long term in most; the hepatic insulin resistance described in small groups is not reflected in insulin requirements [12]. Diabetes may become apparent after kidney transplantation or decompensate under corticosteroid therapy, and the absence of a family history does not exclude the diagnosis, because the alteration arises de novo in more than half of cases [12]. In the differential diagnosis, the PNDS excludes type 1 diabetes by negative anti-GAD, anti-IA-2 and anti-ZnT8 antibodies, and type 2 diabetes by normal weight and the other manifestations, especially renal, although there are cases of HNF1B-MODY without kidney disease [12]; the Norwegian cases with positive autoantibodies show that the first criterion is not absolute [22].

The French PNDS for MODY diabetes (PRISIS reference centre, 2022) places MODY at 2–3% of all diabetes and attributes approximately 75% of identified cases to GCK, HNF1A and HNF4A; it leaves the HNF1B form outside its scope, as it has its own protocol, but summarises it in the differential diagnosis [73]. According to that summary, HNF1B diabetes is accompanied by renal involvement in approximately 90% of cases (cysts, malformations, renal failure, hypokalaemia, hypomagnesaemia), together with genitourinary malformations, pancreatic atrophy with exocrine insufficiency, fluctuating liver test abnormalities, early-onset gout and, in approximately 10%, intellectual delay; the alteration arises de novo in half of cases, so the absence of a family history does not exclude the diagnosis [73]. The 90% proportion refers to patients ascertained through diabetes and is lower than the ~98% nephropathy in adults reported by the dedicated protocol, which is built on series recruited through nephrology [12, 73].

The Norwegian MODY registry and Norwegian childhood diabetes registry allow comparison of deletion and sequence variants [22]:

IQR, interquartile range. Denominators for extrarenal manifestations are small because data were available only for a subset; absence of a trait may reflect lack of screening [22].

Islet autoantibodies do not exclude the diagnosis: in the Norwegian childhood registry, 3 of 7 children with deletion were positive for glutamic acid decarboxylase antibodies (GADA) or anti-IA-2, with an autoimmune phenotype and absent or low insulin production, and one adult with deletion was positive for GADA, anti-IA-2 and anti-ZnT8 [22]. In a Portuguese series of 10 patients, one was anti-GAD positive at diagnosis, with ketoacidosis at 9 years; presenting ketoacidosis occurred in 2 of 10 [57]. Transient neonatal diabetes is rare: by 2026, four cases had been described, all due to de novo point variants (c.1045+1G>A, p.Ser148Trp, p.Ala167fs, p.Ser148Leu), three small for gestational age with pancreatic hypoplasia or atrophy, one of whom progressed to permanent diabetes; Kołbuc and colleagues added the first two cases with whole-gene deletion, both appropriate for gestational age [27, 47]. HNF1B accounts for less than 1% of permanent neonatal diabetes [65].

3.5. Exocrine pancreas and pancreatic morphology

Pancreatic hypoplasia is common: 39/113 (~1/3) deletion carriers had abnormal imaging (hypoplasia, atrophy or agenesis of the body and tail) [2]. In five Norwegian carriers from two families, computed tomography (CT) and magnetic resonance cholangiopancreatography (MRCP) showed absent body and tail in all, with a slightly atrophic head; all had low faecal elastase, three vitamin D deficiency and two vitamin E deficiency, but only one had abdominal symptoms [52]. Faecal elastase-1 was below the control 2.5th percentile (410 µg/g) in 18/29 (62%) patients with a variant and below 200 µg/g, a value suggestive of exocrine insufficiency, in 8/29 (28%); the three with abdominal pain, loose stools or weight loss improved and gained weight with enzyme replacement therapy, and low elastase was more frequent with diabetes (79% versus 47%; p = 0.1) [74]. In a combined analysis of deletions and variants, exocrine insufficiency (elastase <200 µg/g) was present in 37/67 (55%) [2]. The 2022 International Society for Paediatric and Adolescent Diabetes (ISPAD) guideline states that faecal elastase is “always” abnormal in HNF1B-MODY, but the cited study describes only moderately reduced exocrine function, and published series show deficiency in 50–62% of carriers [27, 46, 74]. The absence of pancreatic malformation on CT in all patients in a Portuguese series illustrates the variability [57]. The French PNDS emphasises that abdominal ultrasound has low sensitivity for these abnormalities, that CT or MRI shows agenesis or hypoplasia of the body and tail in 40–100% of patients without specific symptoms and pancreatic calcifications in 25–50%, especially in adults, and that dorsal pancreatic agenesis, being very rare in other settings, is a strong argument in favour of the diagnosis [12]. Exocrine deficiency is very common but rarely causes steatorrhoea; deficiencies of vitamins A, E and D, calcium and zinc may occur, and rare cases of acute pancreatitis have been described [12].

3.6. Liver and biliary tract

Liver abnormalities range from asymptomatic fluctuating transaminase elevation to adult cholestasis [2]. In cohorts ascertained through kidney disease, diabetes or uterine malformation, 74/171 (43%) deletion carriers had elevated liver enzymes or structural abnormalities; in a mixed cohort, liver tests were abnormal in 71%, without genotype correlation, so the true frequency probably exceeds 50% [2]. Several infants had neonatal cholestasis with paucity of interlobular bile ducts and variable periportal fibrosis; one required portoenterostomy and another developed hepatocellular carcinoma with liver transplantation [2]. Choledochal cysts, hepatomegaly, steatohepatitis and hypoplasia with portal vein thrombosis have been described [2]. Gallbladder hypoplasia and duodenal atresia are rare and more often reported with deletion [46]. In pregnancy, cholestasis may be severe and recurrent [75, 76]. According to the French PNDS, neonatal cholestasis due to paucity of the intrahepatic bile ducts is exceptional and jaundice may regress in the first year; in adults, 40–80% have moderate, fluctuating elevation of transaminases or gamma-glutamyltransferase (up to 5 times the upper limit of normal), with no known significant consequence, which frequently worsens after kidney transplantation; liver biopsy is generally normal or shows absence of the primary cilium at the apical pole of cholangiocytes, and the very long-term prognosis is unknown [12].

3.7. Genital tract and fertility

Approximately one third of women and one quarter of men with deletion have genital abnormalities [2]. Other sources estimate genital tract defects in ~20% of women with HNF1B disease, and in the 2005 French series genital abnormalities were found in 80% of examined carriers [1, 77]. In women, Müllerian aplasia or MRKH predominates (partial or complete absence of the upper vagina, cervix and uterus), followed by bicornuate, didelphic or hypoplastic uterus and ovarian cysts; in men, cryptorchidism, shawl scrotum, phimosis, urethral stenosis or obstruction, hypospadias, epididymal cysts, prostatic cyst, vas deferens atresia and infertility [1, 2, 70]. In a series of 108 women with congenital uterine abnormalities, 9 (8%) had an HNF1B alteration: 9 of 50 (18%) with uterine and renal abnormalities, and none of 58 with isolated uterine abnormalities [78]. In a French pilot study of 26 carrier women (deletion in half; age 29 ± 13 years), anti-Müllerian hormone (AMH) was below the age-specific 25th centile in 38%, without association with genotype, renal abnormality or diabetes; uterine malformation was present in 42% and obstetric complications included threatened preterm labour and cholestasis of pregnancy [76]. The French PNDS estimates genital abnormalities in up to 20% of women, with a predominance of bicornuate uterus and, more rarely, abnormalities of the fallopian tubes or vagina, generally with little or no impact on fertility; in men, it considers that hypospadias, cryptorchidism, vas deferens agenesis, epididymal cysts and asthenospermia cannot yet be reliably attributed to HNF1B [12].

3.8. Neurodevelopment and psychiatry in 17q12 deletion

In the Danish population-based iPSYCH2015 study, which started from phenotype and analysed CNV in neonatal blood, 17q12 deletion showed the strongest association among the recurrent deletions studied with ASD (HR 7.79; 95% CI 2.71–22.43) and attention deficit hyperactivity disorder (ADHD; HR 4.24; 95% CI 1.29–13.92), and a trend towards the schizophrenia spectrum (HR 4.84; 95% CI 0.81–28.85); these estimates are less biased towards severe phenotypes than those from clinical cohorts [2, 79]. In genotype-ascertained series, ASD or autistic traits were described in 13% of individuals ascertained through other findings (without systematic assessment), special education needs in 14/110 (13%) children, speech delay in 56%, motor delay in 64% and some degree of learning difficulty, generally mild, in 41/89 [2]. Global delay, intellectual disability, ADHD, bipolar disorder and schizophrenia have been described [2, 43]. Schizophrenia risk was estimated at approximately 30 times baseline from Moreno-De-Luca's data [46]. In UK Biobank, deletions were not significantly associated with fluid intelligence, unlike duplications [21]. Cognitive and psychiatric traits frequently precede renal and metabolic manifestations, and neurodevelopmental referral does not always trigger systemic investigation [46].

The French PNDS treats the deletion as a susceptibility factor whose strength varies with the ascertainment route, the age of the cohorts and the definition of the phenotypes, and gathers the available studies in an annex [12]. In a paediatric series with nephrological entry, 12.7% of deletion carriers did not attend regular mainstream schooling, versus 3.6% of point variant carriers; in 107 French adults with HNF1B diabetes (54 variants, 53 deletions), intellectual disability affected 14%, more than in 339 patients with type 1 diabetes (2.4%) and in 227 with MODY of other causes (2.6%), with no difference between deletion (17%) and variant (11%); in 39 French children (26 with deletion), most had no neuropsychological disorders or severe learning difficulties, although deletion carriers tended to have a lower intelligence quotient; and in a Japanese series with renal entry, 5 of 14 patients had neuropsychological abnormalities [12]. At the other extreme, a review of 361 published cases (Roehlen, 2018) estimated intellectual disability or learning disorder in 58% and neuropsychiatric disorders in 27%, in a sample limited by the content of the publications, and in a Danish cohort of 12 carriers ascertained mainly through microarray requested for developmental delay, 88% had language delay [12]. The PNDS concludes that the deficits generally have a moderate impact on daily life, that they require surveillance in childhood and that specific preventive measures are not justified [12].

3.9. Other deletion manifestations

Dysmorphism is subtle: high forehead, frontal bossing, depressed nasal bridge, deep-set eyes, full cheeks, downslanting palpebral fissures, high palate and arched eyebrows; hypoplastic nails and 2–3 syndactyly of the hands or feet are common [2]. Macrocephaly is among the shared traits in Moreno-De-Luca's series [8]. Ocular findings were described in 24/66 (36%): strabismus, esotropia, nystagmus, posterior embryotoxon, hypermetropia, high myopia, cataracts, coloboma, parapapillary dystrophy, Purtscher-like retinopathy, diabetic retinopathy and keratoconus with corneal transplantation [2]. Prematurity (<37 weeks) occurred in 16/55 (29%) [2]. Without systematic neuroimaging studies, 10/37 (27%) had non-specific abnormalities: ventricular dilatation with or without diffuse atrophy, mild cerebellar atrophy, white matter abnormalities and hippocampal atrophy [2].

Among less common manifestations, congenital heart disease occurred in 10/52 (19%) (ventricular septal defect, aortic coarctation, bicuspid aortic valve, transposition of the great arteries, patent ductus arteriosus, patent foramen ovale, mitral prolapse, aortic root dilatation, among others), short stature in 11/47 (23%) and seizures in 10/84 (12%), including febrile, complex focal and tonic-clonic seizures and mesial temporal epilepsy treated by lobectomy [2]. Joint laxity, long hands and feet, pectus, fifth-finger clinodactyly, gastro-oesophageal reflux, congenital diaphragmatic hernia, duodenal atresia, oesophageal abnormalities, hypotonia, oligohydramnios, sensorineural deafness, deep vein thrombosis or vascular calcifications, ovarian carcinoma and dyslipidaemia have also been described [2].

A cohort of 37 deletion carriers recruited through the 17q12 Foundation (Yale, 2020–2024; self-reported data) described atopy (asthma, eczema or food allergy) in 30–35%, radiographically confirmed pneumonia in 40%, recurrent otitis in 52% and sinusitis in 43%; in 11 patients, the frequency of interferon-γ-producing CD4+ T lymphocytes (6.81% versus 21.91%) and tumour necrosis factor α-producing cells (13.60% versus 34.17%) was lower than in 13 controls (p < 0.001), without differences in Th2 and Th17 cells [33]. This is an exploratory study with a small sample and no established genetic mechanism; the authors propose more than four otitis or sinusitis episodes annually or one to two radiographic pneumonias annually as warning signs of an inborn error of immunity [33].

The oncological risk is discussed in the French PNDS: isolated observations, supported by epidemiological and onco-molecular arguments, have linked loss of HNF1B expression to chromophobe renal cell carcinoma, and adenocarcinomas and, more rarely, pancreatic neuroendocrine tumours have been described in patients with dorsal pancreatic agenesis without genetic characterisation, as well as rare cases of intraductal papillary mucinous neoplasm of the pancreas in HNF1B-MODY [12]. There is no estimate of the absolute risk. The PNDS proposes regular renal ultrasound because of chromophobe carcinoma and considers morphological pancreatic surveillance by CT or MRI justified, without setting the frequency, but in the section dedicated to follow-up it states that there is no established benefit in systematic screening for pancreatic carcinoma [12]. The metastatic appendiceal neuroendocrine tumour in the Lisbon series was an incidental finding [57].

3.10. Recurrent 17q12 duplication

Duplication has a predominantly neurodevelopmental phenotype, ranging from normal intellectual abilities to severe intellectual disability, with reduced penetrance and variable findings [3]. The table compiles frequencies in 86 published individuals, mostly referred for developmental delay or intellectual disability, making these higher than in the carrier population [3]:

P, percentile. The deletion GeneReviews (2025) attributes seizures in 75% to duplication; the primary duplication source gives 36%, the value adopted here [2, 3].

Brain magnetic resonance imaging (MRI) occasionally showed focal cortical dysplasia, agenesis of the corpus callosum, periventricular leukomalacia and schizencephaly [3]. Ocular abnormalities include strabismus, astigmatism, amblyopia, cataract, coloboma and microphthalmia; endocrine abnormalities include diabetes insipidus, type 2 diabetes, growth hormone deficiency, hypoglycaemia, hyponatraemia and hyperkalaemia with pseudohypoaldosteronism [3]. The most common cardiac defect is a small ventricular septal defect; kidney abnormalities (horseshoe kidney, cysts, hypoplasia) affect approximately 20%; tracheo-oesophageal fistula and duodenal atresia have been described, and there is no consistent dysmorphism [3]. In a series of five duplication carriers, all but one had cognitive impairment; the remaining individual had isolated oesophageal atresia with normal development [13]. A boy with a 1.8 Mb duplication had XX sex reversal, Peters anomaly, microphthalmia and glaucoma, and his healthy father was a carrier [13]. In UK Biobank, the 106 duplications were associated with lower eGFR (p = 6 × 10⁻²⁵), eGFR <60 mL/min/1.73 m² including ESKD, and lower fluid intelligence (p = 3 × 10⁻⁴), but not diabetes [21].

3.11. Comparison of the three entities

Sources: [2, 3, 10, 21, 22, 60].

4. Clinical Recognition and Warning Signs

4.1. When to suspect, by age and setting

Prenatally, bilateral hyperechogenic kidneys, with or without cysts, dysplasia and reduced corticomedullary differentiation are the findings most suggestive of HNF1B disease: gene alterations are the most common genetic cause of hyperechogenic fetal kidneys and are identified in 20–30% of fetuses with renal abnormalities [10, 46, 69]. Normal amniotic fluid and normal-sized kidneys do not rule out the diagnosis; a family history of renal cysts or early diabetes increases the probability [69, 80].

In newborns and infants, suspicion arises with CAKUT (hypoplasia, dysplasia, multicystic dysplastic kidney, agenesis, horseshoe kidney, duplication, pelviureteric junction obstruction, isolated hydronephrosis or hydroureter), particularly when associated with growth restriction, transient neonatal hyperglycaemia, neonatal cholestasis, hypomagnesaemia or duodenal atresia [2, 46, 47]. Postnatally, HNF1B alterations are the most common cause of isolated renal hypodysplasia [10]. In children, speech or developmental delay, ASD or mild dysmorphism combined with any kidney abnormality point to 17q12 deletion [2, 44]. In adolescents and young adults, suspicion arises with non-autoimmune diabetes without obesity and developmental kidney disease, early gout or hyperuricaemia, unexplained hypomagnesaemia, unexplained liver enzyme elevation, exocrine pancreatic insufficiency or uterine malformation associated with a kidney abnormality, including primary amenorrhoea due to Müllerian aplasia [19, 34, 78]. In adults with CKD, suspicion arises with tubulointerstitial disease without proteinuria, few cysts in non-enlarged kidneys and a dominant family history of kidney disease or diabetes [2, 46]. Absence of family history does not exclude the diagnosis because de novo alterations and deletions are common [27].

The Exeter monogenic diabetes centre (DiabetesGenes.org) lists unexplained cystic kidney disease (including that detected in utero), solitary kidney, early-onset non-insulin-dependent diabetes, genital tract malformations, hyperuricaemia with early gout, hypomagnesaemia, abnormal liver tests, exocrine pancreatic dysfunction and neurodevelopmental disorders, including ASD, as traits warranting genetic testing [34]. Heneberg proposes suspicion in any patient with CAKUT and at least one extrarenal trait (persistent hypomagnesaemia, pancreatic hypoplasia or low faecal elastase, non-autoimmune diabetes, unexplained liver enzyme elevation or genital malformation), emphasising that restricting suspicion to “diabetes with renal cysts” misses children, in whom renal findings precede diabetes [46].

4.2. Pitfalls delaying diagnosis

Diagnostic delay is the rule. In a Lisbon series, the mean interval between diabetes onset and HNF1B-MODY diagnosis was 16.5 years, and the ten patients had initially been classified as type 1 (six) or type 2 (four) diabetes [46, 57]. Described pitfalls are classification as type 1 or type 2 diabetes, CAKUT assessment in nephrology without genetic investigation, magnesium not measured or attributed to drugs or diet, omission of CNV testing in adults (sequencing panels without deletion detection), underuse of the HNF1B score and compartmentalisation of neurodevelopmental care [46]. A Spanish case illustrates the problem: a woman with diabetes from 10 years, insulinopenia at 19, renal cysts, cholestasis, hypomagnesaemia and a bicornuate uterus was diagnosed with 17q12 deletion only at 49 years, incidentally on a liver disease panel [81].

4.3. Warning signs and urgent situations

Ketoacidosis is rare in HNF1B-MODY but occurred at diagnosis in 2 of 10 Portuguese patients, and insulinopenic patients are at risk of euglycaemic ketoacidosis with sodium–glucose cotransporter 2 inhibitors (SGLT2i) [57, 84, 85].

4.4. Newborn screening

There is no newborn screening for HNF1B disease or 17q12 alterations. Early diagnosis depends on prenatal ultrasound and postnatal investigation of kidney abnormalities [2, 46].

5. Diagnosis: Criteria, Tests and Algorithms

5.1. Criteria and diagnostic definition

There are no consensus clinical criteria for 17q12 deletion, duplication or HNF1B disease; diagnosis is molecular [2, 3]. Deletion is established by identification of the heterozygous ~1.4 Mb deletion in the ClinGen ISCA-37432 interval; minimum coordinates may vary with microarray design, and calculated size may differ because of segmental duplications near breakpoints [2]. Duplication is established by heterozygous duplication of the same interval [3]. HNF1B disease due to an intragenic variant is established by identification of a heterozygous pathogenic or likely pathogenic HNF1B variant, excluding deletion syndrome [2]. A whole-gene deletion detected by MLPA or targeted analysis almost always corresponds to the recurrent 17q12 deletion, but its size should be confirmed by microarray [2, 19].

5.2. HNF1B score

Faguer's HNF1B score (2014) weights 17 clinical and imaging items (prenatal discovery, family history, kidney, pancreas, liver and genital tract), with points assigned according to each trait's frequency and specificity, and was proposed to select patients for genetic testing [11, 25]. Performance differs between derivation and validation cohorts:

Source: [11, 25]. The area under the curve in derivation was 0.78 [11].

The items and points of the score, in the version reproduced in the French PNDS, are as follows; the renal items are added for each kidney [11, 12]:

Source: [11, 12]. The PNDS recommends requesting genetic confirmation when the score is ≥8 [12].

In independent validation, the threshold of 8 would miss one in five cases, reducing clinical utility [25]. In children, traits not yet expressed make the score less useful, so prioritising renal triggers (CAKUT pattern, hyperechogenic kidneys, cysts or dysplasia), biochemical magnesium screening and genetic testing with CNV analysis is recommended [46]. In the Norwegian study, a score ≥8 was used as a supporting phenotypic criterion (PP4) in variant classification [22].

5.3. Biomarkers and ancillary investigations

The pancreatic imaging figure in Heneberg's review, reproduced from Haldorsen's study, shows on CT, MRI and MRCP an atrophic pancreatic head and absent body and tail ventral to the splenic vein, with a short pancreatic duct limited to the head (Figure 1 in Heneberg 2026) [46, 52].

5.4. Genetic tests, sensitivity and limitations

Limitations have practical consequences. In 2005, 60% of patients with a MODY5 phenotype had no point variant on sequencing, and one third of cases were due to large deletions [7]. In the Norwegian registry, MLPA became routine only in 2016, and approximately 200 undiagnosed probands predating that time were never tested for CNV [22]. In monogenic diabetes, CNV explain ~1% of cases (63/5,051 in 11 studies), with HNF1B deletion the most common; NGS increased yield by ~30% compared with Sanger sequencing of GCK, HNF1A and HNF4A and revealed unexpected syndromic forms [25]. An inversion interrupting the gene escapes exon sequencing, MLPA and microarray [55]. Low-level mosaicism may not be detected in leucocyte DNA [2, 46]. A negative result with one method does not exclude disease if the method does not cover the suspected alteration class. The French PNDS describes diagnosis as the combination of techniques for detecting CNV and sequence variants in exonic regions, allows the laboratory to request confirmation on a second sample and distinguishes three results: pathogenic variant (diagnosis established), no variant in the regions explored (diagnosis not confirmed) and VUS (no conclusion, with possible segregation analysis), noting that interpretation may change with family studies and with advances in knowledge [12]. The PNDS for MODY diabetes describes the same pathway for the monogenic diabetes NGS panel, which includes the genes of the syndromic forms and detection of deletions of one or more exons or of the whole gene; in the GCK, HNF1A and HNF4A forms large deletions are very rare (<0.1% of cases), whereas in HNF1B whole-gene deletion accounts for approximately half of the alterations, which makes dosage analysis mandatory when this form is suspected [7, 46, 73].

The recommendations of the American Diabetes Association and European Association for the Study of Diabetes precision medicine initiative (ADA/EASD, Precision Medicine in Diabetes Initiative, PMDI) are: targeted NGS as the preferred method (grade A); panels including all known causes, mitochondrial diabetes, known non-coding variants and CNV detection (grade A); separate MLPA acceptable, although more expensive (grade A) [25]. The reference panel for non-neonatal monogenic diabetes includes GCK, HNF1A, HNF4A, HNF1B, ABCC8, KCNJ11, INS and m.3243A>G [25]. The 2025 Australia and New Zealand (ANZ) consensus recommends considering targeted microarray or MLPA if CNV analysis by NGS is limited or unavailable, and reporting assay limitations (strong recommendation) [26]. ISPAD recommends NGS for suspected monogenic diabetes, except in scenarios highly suggestive of GCK (grade B) [27]. In CKD, the KDIGO genetics consensus prefers phenotype-directed panels, exome or genome sequencing over sequential single-gene analyses [45].

When developmental delay or syndromic traits are present, CMA should be prioritised to define deletion extent and investigate other CNV; microarray yield in children with renal agenesis, hypoplasia or dysplasia is ~10% (10.5% of 522 children), with the most common imbalances at 17q12 and 22q11.2, half inherited from apparently unaffected parents; pathogenic CNV were found in 8% of children with CKD of all causes [46, 86, 87]. Monogenic variants explain ~30–50% of G3b–G5 CKD in children and 10–30% in adults, with yields of 12–65% depending on patient selection [45].

5.5. Who to test: guideline recommendations

The MODY probability calculator applies to people under 35 years and was developed in white Europeans [19]. The Norwegian registry referral criteria (diabetes in a 1st-degree relative, onset before ~35 years, negative autoantibodies, BMI ≤30 kg/m², atypical type 1 diabetes) do not all need to be met [22]. For children with structural renal or pancreatic abnormalities or electrolyte disturbances, the Norwegian authors recommend considering testing even without diabetes or with a single positive autoantibody [22].

5.6. Practical algorithm

For the initial evaluation of a patient with suspected disease, Heneberg proposes renal ultrasound, serum magnesium, potassium and sodium, autoantibodies and C-peptide if hyperglycaemia is present, and faecal elastase-1 if there are gastrointestinal symptoms, poor weight gain or nutritional deficiencies [46]. The decision tree of the French PNDS for MODY diabetes separates three entry points: diabetes before the age of 40 with an atypical type 1 or type 2 phenotype (gene panel), syndromic diabetes (HNF1B variant or deletion, or mitochondrial DNA variant if transmission is maternal) and moderate, stable fasting hyperglycaemia (GCK) [73].

5.7. Genetic reporting, variant classification and VUS

The report should be one page, with the reason for testing and phenotype, the main result without a “positive/negative” dichotomy, and a table giving the gene (HUGO Gene Nomenclature Committee nomenclature), zygosity, nucleotide and protein HGVS description with transcript and version (NM_000458.4), genomic coordinates and American College of Medical Genetics and Genomics and Association for Molecular Pathology (ACMG/AMP) pathogenicity class; benign and likely benign variants are not reported, VUS depend on professional judgement, and offspring risk should be included [25]. For a CNV, the report should state International System for Human Cytogenomic Nomenclature (ISCN) nomenclature, coordinates, genome version and gene content; in recurrent deletion, “arr[GRCh37] 17q12(34856055_36248918)x1” is a published example [47]. Special care should be taken when reporting variants in syndromic diabetes genes in patients with isolated diabetes [25].

There are no HNF1B-specific interpretation rules from the ClinGen Monogenic Diabetes Expert Panel (MDEP), which has published them only for HNF1A, HNF4A and GCK; laboratories adapt the HNF1A specifications, with functional thresholds of ≤50% of normal activity for functional evidence of pathogenicity (PS3 supporting) and ≥85% for benignity (BS3 supporting) [22, 25]. Phenotypic heterogeneity, the high de novo rate (limiting segregation), incomplete penetrance and scarce functional data generate interlaboratory variability [22]. VUS should not guide management and should be reviewed periodically or before pregnancy (strong ANZ recommendation); the KDIGO consensus recommends careful communication of VUS following interdisciplinary discussion, with segregation analysis and periodic reclassification [26, 45]. Pretest and post-test counselling, with written consent, form part of the process [26]. Both French PNDS emphasise that a VUS cannot serve as the basis for presymptomatic diagnosis in relatives [12, 73].

5.8. Prenatal diagnosis

Prenatal CMA requested for fetal renal abnormalities or other indications may detect the deletion, and the correlation between hyperechogenic fetal kidneys and 17q12 deletion is high; several authors recommend investigating deletion in unexplained prenatal hyperechogenic kidneys [2]. A 2026 case illustrates this scenario: a 21-year-old pregnant woman at 19 weeks, with bilateral polycystic hyperechogenic fetal kidneys and severe oligohydramnios, required amnioinfusion to allow amniocentesis; microarray showed the 1.4 Mb 17q12 deletion, with maternal renal cysts and early paternal diabetes in the family history, and pregnancy ended in fetal death following intrauterine infection [80]. If a pregnancy with bright kidneys ends without diagnosis, DNA should be banked and renal biopsy or complete autopsy obtained, and parental renal ultrasound may be useful [86]. The 2018 European recommendation on perinatal cystic kidney disorders addresses imaging, genetic testing and prenatal counselling [88]. A prenatal result does not predict phenotype [2, 3].

The French PNDS adopts a more restrictive position: molecular testing or testing for the 17q12 deletion in the fetus is discussed case by case and is not advised as a matter of course, even for highly suggestive isolated hyperechogenic kidneys, for three reasons: renal severity, which may weigh in a decision on termination, is assessed by imaging and does not correlate with the molecular alteration; the pancreatic prognosis cannot be established in utero; and the neurocognitive risk of the deletion remains under discussion and is highly variable in severity [12]. It recommends instead confirmation of the abnormalities by a reference sonographer and discussion in a multidisciplinary prenatal diagnosis centre (CPDPN), which may have to decide when a very unfavourable renal prognosis leads the parents to request termination of pregnancy for medical reasons [12]. This position differs from that of GeneReviews, which cites authors recommending testing for the deletion in prenatal hyperechogenic kidneys of unknown cause; the difference lies in the value attributed to the neurodevelopmental information provided by microarray [2, 12].

6. Differential Diagnosis

6.1. Cystic kidney disorders and CAKUT

The renal differential depends on age and pattern. In the fetus, hyperechogenic kidneys (“bright”, more echogenic than the liver or spleen) place HNF1B disease alongside autosomal recessive polycystic kidney disease (ARPKD; ~1/20,000 live births), fetal-onset ADPKD, ciliopathies (Bardet-Biedl, Meckel-Gruber, orofaciodigital syndrome type I), Beckwith-Wiedemann syndrome, trisomy 13 and Zellweger syndrome [86]. In children, idiopathic cystic dysplasia and obstructive dysplasia are common differential diagnoses; in adults, acquired cysts of CKD and dialysis and simple cortical cysts [2]. In adults with nephropathy of undetermined cause, HNF1B disease forms part of the extended genetic approach. The 2025 KDIGO ADPKD guideline includes HNF1B-related disease among ADPKD mimics, notes that up to 50% of cases appear de novo and recommends genetic testing when there are few cysts, variable intrafamilial severity, very early ADPKD, imaging discordant with GFR or a negative family history; the Mayo imaging classification does not apply to carriers of variants in genes other than PKD1 or PKD2 [89].

Sources: GeneReviews table adapted from 2019 reviews, Oxford Desk Reference, KDIGO 2025 [2, 86, 89, 90]. AD, autosomal dominant; AR, autosomal recessive.

ADTKD-HNF1B differs from other ADTKD forms through extrarenal manifestations (diabetes, abnormal liver enzymes, CAKUT), although incomplete penetrance of renal involvement and absence of the other manifestations may complicate the distinction [4, 91]. HNF1B variants are present in 3–5% of patients with renal hypoplasia, aplasia or dysplasia [86]. Among CAKUT, molecular diagnosis is possible in 15–20% of patients with renal agenesis, hypoplasia or dysplasia: 5–10% through variants in PAX2, HNF1B, DSTYK, SALL1, SIX1 or EYA1, and ~10% through genomic imbalances [86]. In Japan, 5 of 50 children with CAKUT (10%) had an HNF1B alteration, none with diabetes; in the United States, 4 of 73 children in the CKiD (Chronic Kidney Disease in Children) cohort with renal hypodysplasia [1].

6.2. Tubular hypomagnesaemia

Hypomagnesaemia with hypocalciuria and sometimes hypokalaemia brings HNF1B disease close to Gitelman syndrome (SLC12A3, recessive, with hypokalaemic metabolic alkalosis) and classical Bartter syndrome (CLCNKB) [50]. The KDIGO Gitelman consensus indicates that MODY-type diabetes, early CKD, dominant family history, abnormal liver enzymes, renal or urogenital malformations or renal cysts should point towards HNF1B disease, in which approximately half of patients have renal wasting hypomagnesaemia [50]. EAST syndrome (KCNJ10) is distinguished by epilepsy, ataxia and deafness [50]. Hypomagnesaemia also accompanies renal malformations or MODY caused by PCBD1, an HNF1β coactivator [6, 92].

6.3. Diabetes

Sources: [19, 23, 27, 82].

HNF1B-MODY differs from other MODY forms through developmental kidney disease, the presence of gout and a family history in which some members have kidney disease and others diabetes; no other physical finding distinguishes the causes [19]. Unlike HNF1A-MODY, the SU response is inconsistent [19, 27]. Obesity does not exclude MODY (≥4.5% of overweight adolescents in a type 2 diabetes trial had MODY) and positive autoantibodies do not exclude HNF1B disease [19, 22]. Diabetes developing after kidney transplantation in a patient with nephropathy of unknown cause should prompt consideration of HNF1B disease [83]. The French PNDS for MODY diabetes adds as a guiding sign kidney disease without diabetic retinopathy, which rules out diabetic nephropathy as the explanation; mitochondrial diabetes shares nephropathy unrelated to diabetes, but is distinguished by maternal transmission, sensorineural deafness, macular dystrophy and neurological, muscular or cardiac involvement [73].

6.4. Müllerian aplasia and MRKH syndrome

Müllerian aplasia has a prevalence of 1/4,000 to 1/5,000 women, with renal malformations in 30% [77]. Type 2 MRKH syndrome, with associated renal, vertebral, auditory or cardiac malformations, has an estimated birth prevalence of 1–9/100,000 in Europe, and Orphanet includes HNF1B among the associated genes [66]. 16p11.2 and 17q12 microdeletions were found in 4/38 (10.5%) isolated cases, and 16p11.2, 17q12 and distal 22q11.2 deletions in 5/25 (20%) syndromic cases [77, 93]. Point variants in WNT4 and HNF1B have rarely been documented, and CNV of 17q12, 16p11.2 and 1q21 were identified in the MURCS association [94]. A 2024 review highlights 17q12 (LHX1, HNF1B) and 16p11.2 (TBX6) deletions and variants in GREB1L and PAX8 as promising alterations, and lists other CNV and more than 20 genes with variants in at least two patients [2, 95]. The combination of uterine and renal abnormalities warrants HNF1B investigation; isolated uterine abnormality does not [78].

6.5. Neurodevelopmental disorders and 17q12 duplication

For the neurodevelopmental component, the differential includes hundreds of CNV and intellectual disability genes, and 17q12 duplication enters the differential diagnosis of any intellectual disability, epilepsy or ASD [2, 3]. In practice, first-line CMA resolves the distinction between deletion, duplication and other CNV; the presence of a kidney abnormality, hypomagnesaemia or diabetes in a child with delay points towards 17q12 deletion [2, 46].

7. Treatments and Risk Management Strategies

7.1. Principles and available guidelines

There is no disease-modifying therapy: management is supportive, organ-directed and multidisciplinary (nephrology, endocrinology, gastroenterology, gynaecology, developmental medicine and psychiatry, genetics) [2, 46]. There are no evidence-based guidelines specific to HNF1B disease or the 17q12 deletion; GeneReviews recommendations derive from the authors' experience, and the UK Kidney Association states that the literature does not allow evidence-based guidelines [2, 24]. The French national diagnosis and care protocol (PNDS) «Maladie liée à HNF1β», drawn up in March 2020 by the rare kidney disease reference centres, under the aegis of the ORKiD filière, according to the Haute Autorité de Santé (HAS) method for rare diseases, and published by the HAS in April 2020, is the only national guidance document dedicated to HNF1B disease identified; it is based on a literature review and the consensus of a multidisciplinary group, without grading of recommendations [12]. Applicable recommendations come from monogenic diabetes, CKD and tubular disorder guidelines and diverge on specific points:

7.2. Initial evaluation after diagnosis

The initial work-up in the French PNDS includes creatinine with estimated GFR, magnesium, potassium, uric acid and liver tests, calcium, phosphate and PTH, fasting glucose and HbA1c, faecal elastase on a single sample (in the presence of growth abnormalities in children, maldigestion or weight loss), assessment of psychomotor development and abdominopelvic ultrasound, to which it adds assessment of the educational, socio-occupational and family impact and information on patient associations [12]. Unlike GeneReviews, which requests PTH in the initial work-up and, during follow-up, only if calcium or phosphate is abnormal, the PNDS also includes PTH in the follow-up work-up [2, 12].

7.3. Kidney disease, transplantation and immunosuppression

Kidney disease follows general CKD guidelines for CAKUT and ADTKD in children and adults [2]. Monitoring includes kidney function, blood pressure and proteinuria, with nephrology referral if eGFR falls or proteinuria increases; renoprotection with angiotensin-converting enzyme inhibitors (ACEi) or angiotensin receptor blockers (ARB) is extrapolated from CKD guidelines, without trials in HNF1B disease [46]. Nephrotoxins, particularly non-steroidal anti-inflammatory drugs (NSAID), should be avoided [2]. The French PNDS recognises no preventive measure for the nephropathy; in children it refers to the French PNDS for paediatric CKD (2018) and, in adults, to the HAS reference documents on CKD, and adds avoidance of tobacco to that of NSAID [12]. In hypertensive adults with proteinuric nephropathy, it sets a blood pressure target <130/80 mmHg and the use of ACEi or ARB to reduce albuminuria as far as possible, with a urinary protein/creatinine target <0.5 g/g [12]. This blood pressure target corresponds to earlier recommendations: the 2021 KDIGO guideline on blood pressure in CKD suggests, in adults, a systolic pressure <120 mmHg measured in the office in a standardised manner, when tolerated, and, in children, a 24-hour mean arterial pressure by ABPM at or below the 50th centile for age, sex and height, with ACEi or ARB as first line in the presence of albuminuria; the PNDS target stands as the French position of 2020 [12, 96].

In ESKD, dialysis and kidney transplantation are options; transplantation is a good option because the disease does not recur in the graft, and simultaneous pancreas–kidney transplantation has been successful in patients with diabetes [2, 99]. In the Lisbon series, 5 of 10 patients underwent transplantation (two kidney alone, two kidney–pancreas, one kidney–liver) at a median age of 31 years and with mean renal graft survival of 10.2 years, all functioning; both pancreatic grafts were lost (arterial thrombosis on day 3 in one case; progressive failure in the other, with an incidental metastatic appendiceal neuroendocrine tumour) [57]. The KDIGO consensus recommends early testing of related living donors for a recipient with a known genetic disease, which, in HNF1B disease with variable expressivity, prevents donation by a mildly affected carrier [45]. According to the French PNDS, haemodialysis and peritoneal dialysis are both usually possible; the living donor may be a first-degree relative once it has been excluded that they carry the recipient's alteration; the patient without diabetes should be informed of the high risk (~50%) of early post-transplant diabetes; calcineurin inhibitors worsen the liver abnormalities, so immunosuppression without this class should be considered; and combined kidney–pancreas transplantation, in principle reserved for type 1 diabetes, may be considered close to ESKD, especially if the diabetes is insulin-dependent [12]. The PNDS recommendation covers the whole class of calcineurin inhibitors, whereas that of GeneReviews targets tacrolimus and mTOR inhibitors; neither is based on comparative studies [2, 12].

Post-transplant diabetes risk is increased. A 37-year-old man with gout and normal blood glucose and HbA1c developed cholestatic pruritus and hyperglycaemia of 23 mmol/L requiring insulin weeks after kidney transplantation; HNF1B deletion was identified, tacrolimus minimisation and corticosteroid withdrawal greatly reduced insulin requirements, and ursodeoxycholic acid improved pruritus [83]. In a Chinese cohort, one child had reversible tacrolimus-induced hyperglycaemia [59]. GeneReviews suggests that an immunosuppressive regimen avoiding tacrolimus and mammalian target of rapamycin (mTOR) inhibitors and reducing corticosteroids may benefit these patients, even without prior diabetes; the recommendation rests on clinical reasoning rather than efficacy studies [2]. After transplantation, patients in the Lisbon series used morning isophane insulin (NPH), with or without prandial insulin, with requirements increased by corticosteroids and immunosuppressants [57].

7.4. Hypomagnesaemia and hypokalaemia

Sources differ on the need for treatment. The UK Kidney Association considers hypomagnesaemia generally asymptomatic and rarely requiring treatment; Heneberg's review proposes oral supplementation, or occasionally intravenous supplementation in symptomatic patients; GeneReviews includes hypomagnesaemia in its treatment table; the 2026 Polish case received oral magnesium from the age of 6 [2, 24, 46, 47]. These positions can be reconciled by severity: mild asymptomatic hypomagnesaemia may be monitored, whereas symptomatic (tetany, cramps, arrhythmia) or severe hypomagnesaemia requires correction, and hypomagnesaemia may resist oral supplementation [46, 49]. The French PNDS, the only guidance dedicated to the disease, aligns with the restrictive position: hypomagnesaemia and hypokalaemia are generally well tolerated, respond poorly or incompletely to oral supplementation, and supplementation should be reserved for symptomatic forms, with particular surveillance in pregnancy [12].

The PNDS gives the following doses, which are the reference for HNF1B disease [12]:

Source: [12]. Compared with the KDIGO Gitelman consensus, the adult initial doses coincide (300 mg/day elemental magnesium; ≥40 mmol/day potassium chloride), neither source defines a maximum dose and the targets are the same, but the PNDS paediatric magnesium dose (10–20 mg/kg/day) is two to four times that of the consensus (5 mg/kg/day), and the potassium dose lies at the lower limit of the consensus range (1–2 mmol/kg/day); the consensus weight-based intravenous dose (0.1 mmol/kg per dose) allows the dose to be calculated in children, which the PNDS prescription in ampoules does not [12, 50].

None of these doses has been studied in trials in HNF1B disease. GeneReviews indicates that organic salts (aspartate, citrate, gluconate) may have higher bioavailability than inorganic salts (oxide, sulphate, glycerophosphate) [2]. For comparison, the KDIGO consensus doses for Gitelman syndrome, a distal tubulopathy with a similar electrolyte phenotype, are as follows; their application to HNF1B disease is an extrapolation rather than a studied indication [50]:

Potassium-sparing diuretics (amiloride, spironolactone, eplerenone) and SGLT2i have been proposed as adjuncts for refractory hypomagnesaemia [2]. In Gitelman syndrome, an open-label crossover trial in 30 patients compared indometacin 75 mg, eplerenone 150 mg and amiloride 20 mg daily for 6 weeks, and each drug increased potassium by ~0.3 mmol/L, with sodium depletion (amiloride, eplerenone) or reduced eGFR and gastrointestinal intolerance (indometacin); spironolactone has antiandrogenic effects [50]. In HNF1B disease, NSAID should be avoided because of nephrotoxicity [2]. Evidence for SGLT2i in hypomagnesaemia is a series of four patients without diabetes or HNF1B disease (calcineurin inhibitor- or platinum-associated hypomagnesaemia), in which one stopped high-dose oral supplementation, two stopped intravenous magnesium and one reduced it; use in HNF1B disease is off-label and extrapolated [100]. The Gitelman consensus advises avoiding or cautiously using drugs worsening hypomagnesaemia, such as proton pump inhibitors and gentamicin, and those prolonging the QT interval [50].

7.5. Hyperuricaemia and gout

The sources consulted do not include specific recommendations for treating hyperuricaemia in HNF1B disease. In the family with hyperuricaemic nephropathy described in 2003, all patients improved kidney function with allopurinol [1]. Early gout warrants serum urate measurement in surveillance [24]. High-dose pancrelipase may cause hyperuricaemia, and the US label recommends considering uric acid monitoring in patients with gout, renal impairment or hyperuricaemia, common situations in HNF1B disease [29].

7.6. HNF1B-MODY diabetes

HNF1B diabetes has no defined precision treatment. The PMDI systematic review included 293 cases in 18 articles, without any controlled trial: of 168 with data after genetic diagnosis, 132 (79%) used insulin, without it being known whether other drugs had been tried, and at least one fifth did not use insulin; the authors concluded that data are insufficient to recommend or refute preferential use of insulin or any other drug and that, in patients not receiving insulin, deterioration in secretion should be assessed if control worsens [28]. ISPAD indicates insulin as the HNF1B-MODY treatment; the ANZ consensus indicates SU as initial treatment, with usual progression to insulin, because some patients respond to standard-dose SU (rather than low doses as in HNF1A-MODY); the 2026 systematic review retains insulin as the mainstay and advises a lower threshold for starting it [26, 27, 98]. The French PNDS comes close to the ANZ consensus: while renal function allows, oral treatment is acceptable, including metformin and/or an SU; repaglinide can be used even in severe renal impairment; withdrawal of insulin under an SU or glinide is sometimes successful; and most eventually require insulin, alone or combined with oral antidiabetic agents, owing to loss of secretion or progression of CKD, with good glycaemic control in most cases [12]. HbA1c targets (<7%, or 6.5%), therapeutic education, diabetological follow-up and control of other cardiovascular risk factors are those of diabetes in general; the only prevention proposed is lifestyle and dietary, to avoid excess weight, by analogy with what is suggested in HNF1A diabetes [12]. The PNDS does not give doses of antidiabetic agents; those of the respective labels apply, adjusted to GFR.

SU response data are observational:

Source: PMDI review study table [28].

In practice, insulin is frequently required and should be initiated for symptomatic hyperglycaemia, inadequate control or insulinopenia; a cautious SU trial may be justified near diagnosis, with reassessment of insulin secretion if control worsens [26–28]. The sources consulted do not describe insulin regimens specific to HNF1B disease; in the Lisbon series, insulin-treated patients used basal–bolus regimens [57]. Continuous glucose monitoring (CGM) and individual dose adjustment are recommended in a review to reduce hypoglycaemia (expert opinion) [46]. In a Brazilian case, control remained difficult (HbA1c 12.1%) despite very high-dose insulin (1.7 IU/kg/day), showing that hepatic insulin resistance may be relevant [98]. Metformin and SU tend to be replaced by insulin as kidney function declines; gliquidone, a short-acting SU authorised in advanced renal impairment in some countries, is unavailable in France [84]. In a Portuguese patient diagnosed as having gestational diabetes in the first trimester, with childhood CKD and a bicornuate uterus, control was good with metformin 2 g/day and gliclazide 30 mg/day after 1–2 years of disease [57].

SGLT2i were assessed in a French retrospective study of 40 adults with monogenic diabetes (nine with HNF1B-MODY; age 55.2 ± 16.3 years; HbA1c 7.9 ± 1.9%; eGFR 61.7 ± 22.6 mL/min/1.73 m²; nephropathy in 7/9) followed for 23.9 ± 5.9 months [84]. In HNF1B-MODY, HbA1c increased during the first year (+0.7 ± 2.5%) and returned to baseline at 2 years (7.9%), without glycaemic benefit, unlike HNF1A/HNF4A forms (−0.7 ± 1.3% at year 1); in the entire cohort, urinary albumin/creatinine ratio fell 43.5% at 2 years (geometric mean; 95% CI −64.8 to −9.3%), and in the combined mitochondrial diabetes and HNF1B group it fell 50.2% (95% CI −75.4 to +1.0%), without significant interaction; eGFR in HNF1B fell 6.0 ± 10.5 mL/min/1.73 m² during the first year, attributed by the authors to intrinsic progression, and two HNF1B patients with heart failure with reduced ejection fraction improved their ejection fraction to >40% [84]. Six of 40 patients (15%) experienced non-serious adverse effects (genital infections, urinary tract infection, ketosis, polyuria) and three discontinued, without ketoacidosis, acute kidney injury or deaths [84]. The authors consider SGLT2i a potential option requiring confirmation in controlled studies, do not recommend routine use and highlight euglycaemic ketoacidosis risk in insulinopenic patients [28, 84]. The US dapagliflozin label does not recommend it for CKD treatment in patients with polycystic kidney disease or recent immunosuppressive treatment for kidney disease; HNF1B disease is not a polycystic kidney disease, but no data allow this caution to be dismissed, particularly in transplant recipients [85].

GLP-1 RA have been used as insulin-sparing agents in case reports: a 17-year-old adolescent reduced insulin from 1.0 to 0.4 IU/kg/day with semaglutide 0.5 mg weekly, with HbA1c from 7.0 to 5.9%, time in range from 70 to 88% and time below range from 5 to 1%, but stopped after 4 months because of gastrointestinal effects; in another case, liraglutide added to basal insulin increased time in range from 39 to 65% in 6 days [98]. The 2026 review considers GLP-1 RA and DPP-4i possible adjuncts, with limited strength of evidence [98]. All these uses are off-label because the labels refer to type 2 diabetes [85, 101, 102].

The sources consulted do not include labelled doses for insulin, SU, metformin, empagliflozin and GLP-1 RA; the summary of product characteristics for each product applies, adjusted to eGFR. Microvascular and macrovascular complications depend on the degree and duration of hyperglycaemia rather than aetiology, so general diabetes glycaemic targets apply [19, 98].

7.7. Exocrine pancreatic insufficiency

Pancreatic enzyme replacement therapy is indicated for exocrine insufficiency with clinically significant malabsorption, alongside avoiding smoking and alcohol, small frequent meals with healthy fats and sometimes fat-soluble vitamins [2, 24]. The three symptomatic patients in the Exeter series improved and gained weight [74]. Pancrelipase (Creon®) is indicated in the USA for exocrine pancreatic insufficiency in adults and children, including HNF1B disease; in the EU, Creon is authorised through national procedures [29].

Source: United States Food and Drug Administration (FDA) label for Creon® (pancrelipase), version of 28 February 2024 [29]. Snacks receive approximately half a meal dose; capsules are swallowed whole or opened onto soft acidic food, without crushing or chewing; there are no contraindications; warnings include fibrosing colonopathy with prolonged high doses (colonic stricture in children <12 years with >6,000 units/kg per meal), oral mucosal irritation, hyperuricaemia, porcine protein hypersensitivity and theoretical viral transmission risk [29].

The French PNDS restricts treatment to patients with clinical steatorrhoea and gives fixed doses of enteric-coated pancreatic extracts taken during the meal: 40,000–50,000 lipase units with main meals and 20,000–25,000 units with snacks, with screening for and correction of deficiencies of vitamins A, E, D and K, calcium and zinc when faecal elastase is low [12]. In a 70 kg adult, the PNDS dose for a main meal corresponds to ~570–710 units/kg, close to the initial dose in the US label (500 units/kg per meal) and well below the maximum of 2,500 units/kg per meal; the proportion of half the dose for snacks is the same in both sources [12, 29]. The PNDS gives no doses for children, in whom the weight-based dosing of the label applies.

7.8. Liver, parathyroid, genital tract and other organs

Liver abnormalities are treated according to gastroenterological practice after excluding other causes; a minority of infants with neonatal cholestasis required surgery [2, 46]. Colestyramine and ursodeoxycholic acid partially improved cholestatic pruritus in an adult case, and ursodeoxycholic acid dramatically improved pruritus in the post-transplant case [81, 83]. Hyperparathyroidism follows standard treatment [2]. Hypercholesterolaemia was treated with rosuvastatin in an adolescent with deletion [47]. The French PNDS gives no treatment for the liver abnormalities, except in cholestasis, for which it proposes ursodeoxycholic acid 20 mg/kg/day in two doses with meals [12]. The UK label of Ursofalk® does not include this indication: it provides 14 ± 2 mg/kg/day in primary biliary cholangitis and 20 mg/kg/day in 2–3 doses, which may be increased to 30 mg/kg/day, in hepatobiliary disorders of cystic fibrosis between 1 month and 18 years; the PNDS dose coincides with that of the latter indication, so use in the cholestasis of HNF1B disease is off-label and the dose extrapolated [12, 105].

In Müllerian aplasia, options include non-surgical and surgical intervention, counselling and peer groups; primary vaginal dilatation succeeds in >90–96%, uterine transplantation may restore fertility and surrogacy allows genetic motherhood [2]. Assessment of uterine abnormalities is more important in infertility, and urology guides the management of cryptorchidism and male infertility [46]. Based on a pilot study, the French authors suggest monitoring ovarian reserve and discussing fertility preservation case by case [76]. Refractive errors are corrected with lenses, heart disease is managed by cardiology and deafness with hearing aids [2]. For the genital tract, the French PNDS advises pelvic MRI in girls after menarche and before a first pregnancy plan, with referral of abnormalities to a gynaecology team experienced in malformations and, sometimes, to an assisted reproduction centre; GeneReviews reserves MRI for pubertal girls with primary amenorrhoea [2, 12].

7.9. Neurodevelopment, behaviour, psychiatry and epilepsy

With deletion, treatment includes special education, occupational therapy, physiotherapy, speech and language therapy and psychosocial interventions from early childhood, with developmental paediatrics, child and adolescent psychiatry and psychology, and psychiatry for mood disorders, anxiety and psychosis [2]. GeneReviews refers to American Academy of Child and Adolescent Psychiatry guidelines on psychiatric disorders in intellectual disability and American Academy of Pediatrics and Lancet Commission guidelines on ASD [2]. With duplication, recommendations include early intervention with occupational, physical, speech and feeding therapies, augmentative and alternative communication (which does not impair speech), applied behaviour analysis for ASD, ADHD medication when needed, child and adolescent psychiatry for severe aggression, swallowing studies and thickened feeds for dysphagia, and nasogastric tube or gastrostomy in severe cases [3]. Seizures are treated with standard antiepileptics; none has demonstrated specific efficacy [2, 3]. Nephrologists should refer symptomatic deletion carriers to psychiatry [24, 44]. The French PNDS adds educational and psychosocial organisation: psychological support for the patient, parents and siblings from the time of disclosure; regular assessment of developmental milestones in infants and young children, with speech and language and psychomotor assessment as appropriate; rapid multidisciplinary intervention for a learning disorder, for example in an early medico-social action centre (CAMPS) before the age of 6; a school support assistant, a personalised schooling project or individualised reception plan (PPS/PAI) and, sometimes, adapted schooling; adaptation of care to school, and not the reverse; and therapeutic education programmes authorised by the regional health agency [12].

7.10. Agents and circumstances to avoid

These recommendations derive from clinical reasoning, without efficacy studies [2].

7.11. Comparative regulatory status

No medicine is authorised by the FDA or European Commission specifically for HNF1B disease, RCAD, HNF1B-MODY or 17q12 alterations, and searching the official EMA orphan designation register (3,310 records) found no designation for these conditions; the FDA orphan designation database was not consulted [106]. All drugs are used under a general indication (diabetes, CKD, exocrine pancreatic insufficiency) or off-label.

National access (INFARMED in Portugal, ANVISA and CONITEC in Brazil) to SGLT2i, GLP-1 RA and pancreatic enzymes was not checked in the sources consulted. In France, SGLT2i have been reimbursed since April 2020 (dapagliflozin) and March 2021 (empagliflozin) [84].

7.12. Clinical evidence and registered studies

There are no randomised trials in HNF1B disease; studies are observational, small and heterogeneous and do not allow indirect drug comparisons [28, 98]. Relative albuminuria reductions with SGLT2i were measured in the entire monogenic diabetes cohort rather than HNF1B alone, and the confidence interval in the group including HNF1B included no effect [84]. The ClinicalTrials.gov search found no interventional therapeutic trial specific to HNF1B disease or 17q12 alterations; the European clinical trials register (EU CTR) was not consulted, and a negative search does not prove an absence of studies [35]. Other related registries include Simons Searchlight (NCT01238250, United States, genetic alterations with autism traits, including recurrent CNV), a Norwegian study switching insulin to SU in HNF1A/HNF4A variants (NCT04239586, not applicable to HNF1B) and a four-gene MODY genetic testing study in the United Arab Emirates (NCT05586594) [35].

7.13. Developments in the last 12 months

No antisense oligonucleotide trial in HNF1B-MODY has been published, and there are no disease-specific announcements, opinions or regulatory authorisations during the period [35, 46].

7.14. Pregnancy and breastfeeding

Pregnancy data in monogenic diabetes are mainly general. The baseline birth defect risk is 3–4%, rises to ~6–8% in pregestational insulin-dependent diabetes and lies between these values in non-insulin-dependent diabetes; good glycaemic control reduces, without eliminating, the risk of malformations and neonatal complications; one meta-analysis found no difference in malformations between oral hypoglycaemic agents and insulin, and another found more macrosomia and neonatal hypoglycaemia with glibenclamide than with insulin; high-resolution ultrasound with fetal echocardiography and maternal–fetal medicine follow-up are recommended [19]. HNF1B disease adds specific risks: uterine malformation (bicornuate uterus) was associated with preterm births, cholestasis of pregnancy may be severe and recurrent, and hypomagnesaemia may worsen to the point of requiring continuous magnesium infusions [75, 76]. In a 30-year-old pregnant woman with two previous preterm births, bicornuate uterus and severe cholestasis in previous pregnancies, critical hypomagnesaemia with urinary wasting, severe cholestasis and gestational diabetes led to induction at 33 weeks, and the 17q12 deletion was diagnosed on maternal microarray; the authors advise informing affected women about the risk of fetal death associated with severe cholestasis and renal dysfunction [75]. The Gitelman consensus notes that ACEi and ARB should be stopped in pregnancy because of fetal risk and that electrolytes are monitored during labour [50]. SGLT2i are not recommended during the second and third trimesters or during breastfeeding, according to the US label [85]. Pancrelipase has minimal systemic absorption and no identified fetal risk [29]. The French PNDS calls for screening for diabetes before a planned pregnancy, with HbA1c at target (5.5%) at conception because of the malformative and placental risks, a value more demanding than the non-pregnancy targets in the same document (<7% or 6.5%) and than the preconception target of the 2026 American Diabetes Association standards, which recommend glucose as close to normal as possible without excessive hypoglycaemia, ideally HbA1c <6.5% (<48 mmol/mol) [97]; regular glucose monitoring during pregnancy because of the risk of decompensation with physiological insulin resistance; particular surveillance of electrolyte supplementation; and follow-up by a reference sonographer and an informed obstetrician, for the maternal risk of diabetes and for fetal growth [12].

7.15. Perioperative period and transition

SGLT2i should be stopped at least 3 days before surgery or prolonged fasting [85]. Hypokalaemia and hypomagnesaemia may potentiate anaesthetics and neuromuscular blockers; in the general population, British guidance suggests potassium ≥3.0 mmol/L and magnesium ≥0.5 mmol/L before anaesthesia [50]. Transition from paediatric to adult care requires continuity of renal and metabolic surveillance, screening for diabetes emerging in adolescence or young adulthood, adolescent psychiatric assessment in deletion carriers and reproductive counselling before pregnancy [2, 24]. In the United Kingdom, a national multidisciplinary Exeter clinic (diabetology, paediatric and adult nephrology, genetics, dietetics and research) follows children and adults, with neurodevelopmental and psychiatric screening [34].

8. Reproductive Options and Family Counselling

8.1. Inheritance and recurrence risks

The three entities are transmitted in an autosomal dominant manner, with highly variable expressivity and, for duplication, reduced penetrance [2, 3, 25]. The proportion of de novo cases differs: for deletion, 82 of 109 individuals with tested parents in 26 studies (75%) had a de novo deletion; in the ERKNet cohort, 53% of deletions and 26% of intragenic variants were de novo; in the Yale cohort, 55%, with 21% confirmed inheritance and 24% unknown origin; in Norwegian newborns, all three deletions found were de novo [2, 10, 20, 33]. Duplication is inherited in ~90% of cases [3]. In the ERKNet cohort, transmission was maternal in 64% and paternal in 35% of inherited cases; no parental transmission bias for 17q12 has been described in recurrent CNV analyses [10, 61].

Sources: [2, 3, 19, 25, 61].

Parents should be tested even without apparent manifestations for two reasons: the recurrence risk and their own risk of kidney disease and diabetes, which may emerge only in adulthood and go unnoticed in a mildly affected parent [2]. Leucocyte DNA may fail to detect somatic mosaicism or exclusively germline mosaicism [2]. In HNF1B disease, mosaicism may produce attenuated or single-organ phenotypes in the parent, which complicates counselling [46]. The French PNDS estimates de novo occurrence in ~70% of deletions and ~50% of point variants, values different from those of the ERKNet cohort (53% and 26%); it recommends a genetics clinic referral and family studies in all cases, with targeted testing systematically offered to parents, and records the exceptional observation of germline mosaicism undetected on parental testing, with recurrence among siblings [10, 12]. A published example is the mother of a patient with a complex duplication, who was mosaic (28% of nuclei) for an LHX1 duplication present in 98% of her son's cells [13].

8.2. Cascade testing and asymptomatic relatives

Following identification in the proband, cascade testing and counselling are standard practice [45]. If a parent is a carrier, the status of older and younger siblings and other at-risk relatives should be clarified for renal, diabetes and neurodevelopmental surveillance; targeted testing (FISH, qPCR or MLPA) suffices for relatives, because almost all whole-gene deletions detected by targeted analysis include the entire recurrent region, which may be confirmed by CMA [2]. The ANZ consensus recommends cascade testing by familial variant genotyping (strong recommendation) [26]. For duplication, GeneReviews suggests testing the proband's siblings for developmental and psychological surveillance into adulthood [3]. Relatives of people with MODY generally favour early predictive testing, allowing targeted surveillance and early intervention to reduce microvascular and macrovascular complications; carriers without diabetes should have annual HbA1c [19, 23]. In the Portuguese series, three relatives previously labelled as having type 2 diabetes were reclassified after family screening, and in the post-transplant case relatives were offered renal ultrasound and testing [57, 83].

Testing asymptomatic minors raises consent and autonomy issues [23]. In HNF1B disease, renal manifestations may be present from childhood and surveillance has clinical benefit, justifying paediatric testing of at-risk relatives; for duplication, the benefit is mainly developmental surveillance [2, 3]. For CNV carriers with incomplete penetrance, there is no consensus on cascade testing because outcomes are uncertain [61].

8.3. Genotype-based counselling

The ERKNet cohort provides data for personalised counselling: in a child with 17q12 deletion and two functioning kidneys, the probability of stage 3 CKD before 18 years is low, but the increased neurodevelopmental disorder risk should be explained; in a child with an intragenic variant, particularly in POUs, earlier and more intensive renal monitoring is warranted [10]. Deletion confers risk of ASD, ADHD and schizophrenia, and duplication a risk of intellectual disability with reduced penetrance [2, 3, 79]. Variable expressivity should be explained with examples, and the patient should receive a copy of the report [25].

8.4. Prenatal diagnosis and preimplantation genetic testing

With a known familial alteration, prenatal diagnosis and preimplantation genetic testing (PGT) are possible [2, 3]. The prenatal result does not predict the phenotype, and perspectives on the use of these tests differ; the decision is personal [2]. In France, when prenatal genetic analysis is performed, the result is disclosed in a joint consultation with a geneticist and a paediatrician, who may propose complementary analyses (biochemical or proteomic) to refine the renal functional prognosis, support for the couple and a care plan for the child; postnatal disclosure takes place in a dedicated consultation, by a physician experienced in the disease, if possible with a psychologist, and in several stages [12]. The ANZ consensus recommends information on all reproductive options, including in vitro fertilisation with PGT, with non-directive counselling, and considers that PGT may be appropriate in subtypes with potentially severe manifestations, such as HNF1B syndrome (strong recommendation) [26]. Some prenatal CMA laboratories do not routinely report CNV with substantial incomplete penetrance, which should be discussed before testing [61]. Without a molecular diagnosis, empirical CAKUT recurrence risks are 3–8% for siblings in bilateral renal agenesis and 2–3% in non-lethal multicystic dysplastic kidney (15–20% if one parent is affected) [86].

The ideal time for counselling is before pregnancy, and it should be offered to affected or at-risk young adults [2]. In affected women, preconception counselling includes uterine malformation, preterm birth and cholestasis risks, hypomagnesaemia, kidney function and diabetes control [19, 75, 76]. In Müllerian aplasia, uterine transplantation and surrogacy make recurrence counselling relevant; in an older study, none of 17 daughters born through surrogacy to women with MRKH was affected [77, 95].

9. Special Considerations: Population Diversity and Epidemiology

9.1. Prevalence and what each number measures

Estimates vary by two orders of magnitude because they measure different things: genotype frequency in unselected newborns or adults, frequency in clinical cohorts and numbers of diagnosed cases.

The Norwegian birth estimate is much higher than that from UK Biobank (1 in ~156,000 participants according to the original analysis cited by the Norwegian authors), which, according to the authors, suggests that many carriers remain undiagnosed [20]. The French PNDS estimate of ~1/67,000, which covers deletions and variants, is more than 15 times lower than the birth frequency of the deletion alone in Norway, which is compatible with a large fraction of undiagnosed or mildly affected carriers [12, 20]. Orphanet estimates refer to diagnosed cases and are not comparable with genotype frequencies [2, 14]. No population estimate exists for intragenic HNF1B variants, so the total frequency of HNF1B disease exceeds that of the deletion [6].

9.2. United States, Europe and founder populations

Most data come from Europe: France (201-adult cohort; Toulouse and Limoges paediatric cohorts), United Kingdom (Exeter, UK Biobank), Norway (national MODY and childhood diabetes registries, MoBa), Denmark (iPSYCH2015), Iceland, Greece, Poland and the ERKNet cohort from 14 European countries [10, 20–22, 47, 54, 58, 74, 79]. In the United States, data mainly come from clinical microarray cohorts, the CKiD cohort and the Yale cohort recruited through the 17q12 Foundation [1, 33, 43]. Outside these regions, cohorts exist in Japan and China [1, 59]. The sources consulted do not describe HNF1B founder variants, which is consistent with the high proportion of de novo alterations and the recurrent NAHR mechanism of the deletion [2, 10]. In a gnomAD study, predicted pathogenic variants across the set of six CAKUT genes were more frequent in African Americans (1/149) and less frequent in Ashkenazi Jews (1/864), data not specific to HNF1B [67]. In France, the data underpinning the 2020 PNDS come mainly from the cohorts of the rare kidney disease reference centres (Toulouse, Limoges, Paris) and the PRISIS centre for diabetes [12].

Clinical tools for selecting patients for testing were derived in European populations: the MODY probability calculator and type 1 diabetes genetic risk scores were developed in white Europeans, ethnic minorities are under-represented and many MODY cases remain undiagnosed [19]. Most HNF1B diabetes cases included in the PMDI systematic review were of European ancestry [28]. BMI thresholds used to select patients for testing may need adjustment for ancestry [25].

9.3. Portugal and Brazil

In Portugal, the only series identified in the sources consulted is from Centro Hospitalar Universitário Lisboa Central: 10 patients (7 index cases and 3 relatives) followed between 2013 and 2021, with median age at diabetes diagnosis of 28 years, at HNF1B-MODY diagnosis of 40.5 years and a mean interval of 16.5 years between them; diabetes was the first manifestation in half, and in the other half the first manifestation was renal malformation or paediatric CKD, with kidney transplantation in all these patients [57]. The authors describe it as the first Portuguese series, single-centre, without national prevalence value [57]. In Brazil, a national genetic syndrome manual includes HNF1B in the historical MODY classification, without epidemiological data, and the sources consulted include only one Brazilian case with difficult control despite very high-dose insulin, cited by a systematic review [38, 98]. Absence of data in the sources consulted does not demonstrate an absence of published data in these countries.

9.4. Special populations

In children, the presentation is dominated by kidney abnormalities and, with deletion, neurodevelopment; diabetes has not yet manifested in many young carriers, and adult-derived scores perform less well [22, 46]. In pregnancy, uterine malformation, cholestasis and hypomagnesaemia combine [75, 76]. In kidney transplant recipients, post-transplant diabetes risk influences immunosuppression choice [2, 83]. In adults with deletion and cognitive impairment or psychiatric illness, adherence to surveillance and selection of psychotropic drugs with less metabolic and renal impact become important [2].

9.5. European reference networks, resources and patient organisations

The relevant European reference networks are ERKNet, for HNF1B kidney disease, CAKUT and ADTKD, which coordinated the European cohort of 521 patients; Endo-ERN, through its rare genetic glucose–insulin homeostasis disorder group, for monogenic diabetes; and ERN ITHACA, for microdeletion syndromes and neurodevelopmental disorders [10, 30–32]. No HNF1B-specific ERKNet guideline was found. EURORDIS represents European rare disease patient organisations [110].

Resources for patients and clinicians include the 17q12 Foundation (United States; chromo17q12.org), Unique (United Kingdom; rarechromo.org, with guides on 17q12 microdeletions and microduplications), Chromosome Disorder Outreach (United States), the Simons Searchlight registry for autism-associated CNV, Exeter's DiabetesGenes.org, with patient and clinician information and a national genetic testing service, the University of Chicago monogenic diabetes registry [111] and the UK Kidney Association RaDaR registry, including an HNF1B kidney disease group [2, 3, 24, 33–35]. The American Diabetes Association, Diabetes UK and ISPAD provide MODY information [19]. The PRISMA research group (University of Alberta) studies autism-associated CNV, including 17q12 deletion [2].

In France, the 2020 PNDS organises care around the ORKiD rare kidney disease filière and the SORARE (Toulouse, with paediatric teams in Bordeaux, Montpellier and Limoges), MARHEA (Hôpital Necker-Enfants Malades, Paris) and NEPHROGONES (Lyon) reference centres, with the PRISIS reference centre (rare disorders of insulin secretion and insulin sensitivity, Hôpital Saint-Antoine, Paris) of the FIRENDO filière for diabetes [12]. Testing for the 17q12 deletion is performed by all laboratories that carry out microarray (AChro-Puce network), and targeted HNF1B sequencing was performed in 2020 by laboratories in Lyon, Brest, Lille, Paris (Necker, Hôpital Européen Georges Pompidou, Robert Debré, Bicêtre and Pitié-Salpêtrière), Rouen and Toulouse [12]. The disease is recognised as an off-list long-term condition (ALD), with 100% coverage that may be based on the items for severe chronic nephropathy or diabetes; the patient associations listed are AIRG-France (information and research on genetic kidney diseases) and France Rein, and the Maladies Rares Info Services service provides telephone information [12]. For the diabetological aspect, the 2022 PNDS for MODY diabetes lists the PRISIS network of reference and competence centres, the FIRENDO filière and the Orphanet entry ORPHA:552 (MODY), and the associations Aide aux Jeunes Diabétiques (AJD), Fédération Française des Diabétiques and Association des femmes diabétiques [73].

10. Illustrative Clinical Case

10.1. Transient neonatal diabetes as the first clue to HNF1B deletion (Poland, 2026)

Kołbuc and colleagues described two Polish girls with whole-gene deletion presenting with transient neonatal diabetes in 2026 [47]. The first was born from a first pregnancy at 35 weeks after premature rupture of membranes, weighing 2,025 g (appropriate for gestational age), and required non-invasive ventilation until day two [47]. She had hyperglycaemia from day one (maximum blood glucose 298 mg/dL, serum insulin 35 µIU/mL), treated with intensive insulin until day five, with subsequent normalisation, and transient hypomagnesaemia (0.65 mmol/L); ultrasound showed bilateral renal dysplasia and only the pancreatic head, radiography showed absence of the 12th pair of ribs, and family history was negative [47]. She was discharged from diabetes follow-up at one year, without recurrence [47].

At 4 years, investigation for abdominal pain revealed mild hyperuricaemia (5.5 mg/dL) and aspartate aminotransferase (AST) of 79.8 U/L, with normal magnesium (0.79 mmol/L), blood glucose (83 mg/dL), HbA1c (5.07%), C-peptide (1.53 ng/mL) and PTH; ultrasound showed the pancreatic head and body without a tail, prompting suspicion of HNF1B disease [47]. MLPA (SALSA P241-E1) showed heterozygous deletion of HNF1B exons 1 to 9, absent in both parents (de novo alteration) [47]. Oral glucose tolerance testing (OGTT) was normal at 4 years; at 6 years, glucose exceeded 200 mg/dL at 30 and 60 minutes (maximum 210 mg/dL), and at 8 years the maximum was 175 mg/dL at 30 minutes, with consistently normal values at 0 and 120 minutes, normal HbA1c (maximum 5.32%) and C-peptide at least doubling after glucagon [47]. OGTT and insulin curves at different ages appear in Figure 1 of Kołbuc et al., with an insulin peak of ~88 µU/mL at 60 minutes at 8 years [47]. Ultrasound remained stable, with slightly increased bilateral echogenicity and small left-sided cysts; MRI at 8 years showed bilateral cysts with preserved corticomedullary differentiation and a morphologically normal pancreas, discordant with ultrasound [47]. She has received oral magnesium since 6 years; at 8 years, on supplementation, she had magnesium of 0.79 mmol/L with FEMg 7.24%, AST 81.2 U/L, fasting blood glucose 62 mg/dL, HbA1c 4.8% and eGFR 136.8 mL/min/1.73 m² [47].

The second was born at 31 weeks weighing 1,320 g (appropriate for gestational age), with cardiorespiratory failure and necrotising enterocolitis; she had hyperglycaemia on day three (maximum 249 mg/dL), treated with continuous insulin infusion for 5 days, a non-visualised pancreas and hypomagnesaemia (0.72 mmol/L) at discharge [47]. She developed global developmental delay (intellectual, motor and speech) from 2 years, suspected epilepsy at 5, an isolated fasting glucose of 121 mg/dL at 6 years that was not investigated and, from 7 years, hypothyroidism, hypercholesterolaemia (maximum 312 mg/dL), persistent metopic suture and scoliosis [47]. Bilateral renal cysts were an incidental finding on spinal MRI at 10 years, and at 11 ultrasound showed cystic disease resembling medullary sponge kidney; the cystic kidney disease NGS panel showed complete HNF1B deletion, and aCGH confirmed the 1.4 Mb 17q12 deletion, arr[GRCh37] 17q12(34856055_36248918)x1; the mother was not a carrier and the father was not tested [47]. Between 12 and 15 years, she maintained hypomagnesaemia (0.56–0.66 mmol/L) despite supplementation, hyperparathyroidism (PTH up to 131 pg/mL) with hypophosphataemia due to renal phosphate wasting, hyperuricaemia, mild facial dysmorphism (high forehead, narrow palpebral fissures), clinodactyly and low-set thumbs, EEG without epileptiform discharges and bilateral medullary nephrocalcinosis on CT; MRI showed a partially developed pancreas (head and rudimentary body) [47]. OGTT remained normal, but the HOMA-IR (homeostasis model assessment) insulin resistance index increased from 2.49 to 4.83 in 2 years, with fasting insulin of 22 µU/mL and C-peptide from 3.54 to 8.19 ng/mL after glucagon, and metformin was recommended; hypercholesterolaemia was treated with rosuvastatin (low-density lipoprotein cholesterol, LDL, from 186 to 88 mg/dL) [47].

The cases illustrate several points: transient neonatal diabetes may be the first manifestation even with appropriate birth weight for gestational age; diagnosis emerged years later through extrarenal traits (liver enzymes, incomplete pancreas) or a renal panel; whole-gene deletion was accompanied in the second patient by the 17q12 deletion neurodevelopmental phenotype; and ultrasound, MRI and CT gave discordant pancreatic information [47]. The authors did not exclude other neonatal diabetes causes by testing [47]. Continuous magnesium supplementation differs from the UK Kidney Association position that treating hypomagnesaemia is rarely necessary, and metformin for insulin resistance in an adolescent without diabetes is an individual author decision rather than a guideline recommendation [24, 47].

10.2. “Type 1” diabetes for 39 years (Spain, 2025)

Sager-La Ganga and colleagues described a woman with diabetes onset at 10 years in 1985, with polyuria, polydipsia, blood glucose 120–230 mg/dL, negative ketonaemia, HbA1c 6.0% (42 mmol/mol) and negative islet antibodies on two measurements, who was enrolled in a randomised immunosuppression trial for type 1 diabetes (azathioprine 75 mg/day and monthly intramuscular thymostimulin for 4 years) and remained untreated until 19 years [81]. At 19 years she had insulinopenia (C-peptide <0.8 ng/mL after glucagon), HbA1c 9.1% and BMI 20.2 kg/m², with elevated transaminases and two simple renal cysts, and started insulin, without microvascular complications 30 years later [81]. At 35 years she developed cholestatic pruritus (alkaline phosphatase 413 U/L) with negative liver autoimmunity and biopsy showing minimal steatosis, treated with colestyramine and ursodeoxycholic acid with partial improvement; she had persistent hypomagnesaemia (1.0–1.2 mg/dL) with preserved kidney function and a bicornuate uterus with fibroids [81]. Diagnosis occurred incidentally at 49 years on an exome-based liver disease panel: a heterozygous ~1.3 Mb deletion in 17q12 (including ACACA, HNF1B and PIGW), confirmed by aCGH; there was no cascade testing because there were no siblings or descendants [81]. The article's figure is a clinical timeline [81]. The case shows how diabetes with negative autoantibodies, liver abnormalities, hypomagnesaemia and uterine malformation may remain labelled as type 1 diabetes for decades, and why guidelines recommend including HNF1B in panels even for apparently isolated diabetes [26, 27, 81].

11. Prognosis and Long-Term Follow-up

11.1. Prognosis

Prognosis is determined mainly by kidney function, diabetes and its complications and, with deletion, neurodevelopment. In the ERKNet cohort, survival without stage 3 CKD at 18 years was 87% with deletion and 63% with intragenic variants, and ESKD-free survival at 18 years was 97% and 86%, respectively; eGFR was lower with variants from the first week of life, without a difference in slope [10]. In adults, kidney disease is slowly progressive: in 17 adults without ESKD, the mean annual eGFR decline was 2.45 mL/min, ranging from stable to 19 mL/min/1.73 m² per year; in the French cohort of 201 adults, 44% had stage 3–4 CKD and 21% ESKD, more frequently with intragenic variants (78% versus 51% for stage 3–4 CKD or ESKD in the long term) [54, 84]. A single functioning kidney accelerates progression with deletion [10]. In Greek children, mean eGFR remained stable over 5 years, with substantial individual variability [58]. The French PNDS places the median age at ESKD at approximately 50 years, with a mean annual GFR decline of ~2 mL/min/1.73 m² in adults, and states that the disease does not recur after transplantation [12].

Diabetes affects ~50% of carriers over time, with penetrance increasing with age to almost 80% in adults; most eventually require insulin [2, 47, 54]. Microvascular complications are common: retinopathy or neuropathy in 46 of 114 French adults, and complications in 5 of 10 Portuguese patients, in those with longer diabetes duration (median 21 years) and poorer control (median HbA1c 9.3%) [54, 57]. Cardiovascular risk factors are highly prevalent in adults [54]. Kidney transplantation has good outcomes without disease recurrence; pancreatic graft outcomes were more variable in a small series [2, 57]. With deletion, neurodevelopment influences independence and schooling, and the risk of psychosis or bipolar illness persists in adolescence and adulthood [2, 79]. With duplication, prognosis depends on the degree of intellectual disability and epilepsy, and many carriers of inherited duplications have a mild phenotype [3]. The sources consulted provide no mortality or life expectancy data.

11.2. Surveillance

Surveillance recommendations are expert-based and differ in intervals; the table indicates the source of each.

Heneberg's review proposes annual reviews when renal and endocrine function are stable and six-monthly reviews for abnormalities in several domains; with intragenic variants, surveillance centres on endocrinology and nephrology, and with deletion neurodevelopmental and psychiatric assessment, reproductive tract imaging and counselling are added [46]. The ANZ consensus recommends surveillance of urogenital, hepatic and pancreatic manifestations in addition to the microvascular complications of diabetes (strong recommendation) [26]. For hypomagnesaemia, the UK Kidney Association considers that it rarely requires treatment and the French PNDS reserves supplementation for symptomatic forms, whereas other sources propose broader supplementation, so the treatment threshold is determined by severity and symptoms [12, 24, 46]. Periodic renal ultrasound in adults is proposed only by the PNDS, on the basis of the risk of chromophobe carcinoma; the UK Kidney Association considers creatinine and eGFR sufficient in adults [12, 24].

12. Summary and Final Considerations

HNF1B-related disease results from loss of function of a transcription factor organising kidney, pancreatic, biliary tract and genital tract development and, in adults, maintaining tubular epithelial differentiation and repressing PTH. Approximately half of cases are due to the recurrent ~1.4 Mb deletion at 17q12, generated by recombination between segmental duplications and frequently de novo; the remainder are due to intragenic variants. Reciprocal duplication, with reduced penetrance and mostly inherited, has a predominantly neurodevelopmental phenotype and a modest renal effect in the population, without association with diabetes.

The presentation changes with age: hyperechogenic or cystic fetal kidneys, CAKUT, hypomagnesaemia and hyperuricaemia in childhood, non-autoimmune diabetes with progressive insulinopenia in adolescents and young adults, and slowly progressive tubulointerstitial CKD, liver abnormalities, exocrine pancreatic insufficiency, hyperparathyroidism and genital malformations throughout life. European cohorts show that deletion has a better renal prognosis than intragenic variants, particularly those in the POU-specific domain, and conversely adds risk of developmental delay, ASD, ADHD and psychosis. High penetrance for the multisystem combination and ~34% penetrance for the paediatric phenotype prompting microarray measure different realities.

Diagnosis is molecular and requires testing detecting both sequence variants and CNV: sequencing alone misses deletions, microarray misses sequence variants, and only genome sequencing detects inversions. Diabetes guidelines recommend testing for atypical diabetes with developmental kidney disease or pancreatic or uterine abnormalities, even without family history and even with apparently isolated diabetes; positive autoantibodies do not exclude diagnosis. The HNF1B score has a high negative predictive value in the derivation cohort, but 80% sensitivity in independent validation and worse performance in children.

There is no disease-modifying therapy or authorised medicine for the disease. Diabetes is treated mainly with insulin, although guidelines differ: ISPAD indicates insulin, the Australian and New Zealand consensus accepts sulphonylureas as initial treatment, and the ADA/EASD initiative systematic review concludes that the evidence does not allow any drug to be recommended; SGLT2i and GLP-1 receptor agonists are exploratory. Hypomagnesaemia is treated according to severity, with doses extrapolated from Gitelman syndrome; symptomatic exocrine insufficiency with pancreatic enzymes; ESKD with kidney transplantation and immunosuppression limiting tacrolimus and corticosteroids. Follow-up is multidisciplinary and organised by the organs involved, with renal and metabolic surveillance intervals varying from one to five years according to the source and the patient's status. The 2020 French PNDS, the only national guidance dedicated to the disease, reserves electrolyte supplementation for symptomatic forms, accepts initial oral treatment of diabetes, does not advise prenatal testing as a matter of course for isolated hyperechogenic kidneys and adds periodic renal ultrasound in adults because of the risk of chromophobe carcinoma. Recurrence is 50% for a carrier's children and approximately 1% or less for siblings when the parents are not carriers, and unpredictable expressivity influences prenatal and preimplantation counselling.

13. References

1. Online Mendelian Inheritance in Man, OMIM®. #137920 Renal cysts and diabetes syndrome; RCAD [Internet]. Baltimore (MD): Johns Hopkins University; last edited 2025 Apr 21 [cited 2026 Oct 9]. Available from: https://www.omim.org/entry/137920
2. Mitchel MW, Moreno-De-Luca D, Myers SM, Levy RV, Turner S, Ledbetter DH, Martin CL. 17q12 Recurrent Deletion Syndrome. 2016 Dec 8 [Updated 2025 Feb 6]. In: Adam MP, Feldman J, Mirzaa GM, Pagon RA, Wallace SE, Amemiya A, editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993-2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK401562/
3. Mefford H. 17q12 Recurrent Duplication. 2016 Feb 25 [Updated 2021 Apr 29]. 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/NBK344340/
4. Eckardt KU, Alper SL, Antignac C, Bleyer AJ, Chauveau D, Dahan K, et al. Autosomal dominant tubulointerstitial kidney disease: diagnosis, classification, and management--A KDIGO consensus report. Kidney Int. 2015;88(4):676-683. doi:10.1038/ki.2015.28. PMID: 25738250.
5. ClinGen Dosage Sensitivity Curation. 17q12 recurrent (RCAD syndrome) region (includes HNF1B), ISCA-37432; and HNF1B gene [Internet]. Clinical Genome Resource; region curated 2022 Jun 16, gene curated 2020 Dec 21 [cited 2026 Oct 9]. Available from: https://www.ncbi.nlm.nih.gov/projects/dbvar/clingen/clingen_region.cgi?id=ISCA-37432
6. GeneGuidance. HNF1B — clinical gene curation (version of 2026-09-26). Local project database; unpublished.
7. Bellanné-Chantelot C, Clauin S, Chauveau D, Collin P, Daumont M, Douillard C, et al. Large genomic rearrangements in the hepatocyte nuclear factor-1beta (TCF2) gene are the most frequent cause of maturity-onset diabetes of the young type 5. Diabetes. 2005;54(11):3126-3132. doi:10.2337/diabetes.54.11.3126. PMID: 16249435.
8. Online Mendelian Inheritance in Man, OMIM®. #614527 Chromosome 17q12 deletion syndrome [Internet]. Baltimore (MD): Johns Hopkins University; last edited 2013 Nov 26 [cited 2026 Oct 9]. Available from: https://www.omim.org/entry/614527
9. Freeman PJ, Hart RK, Gretton LJ, Brookes AJ, Dalgleish R. VariantValidator: accurate validation, mapping, and formatting of sequence variation descriptions. Hum Mutat. 2018;39(1):61-68. doi:10.1002/humu.23348. PMID: 28967166. PMCID: PMC5765404.
10. Buffin-Meyer B, Richard J, Guigonis V, Weber S, König J, Heidet L, et al. Renal and Extrarenal Phenotypes in Patients With HNF1B Variants and Chromosome 17q12 Microdeletions. Kidney Int Rep. 2024;9(8):2514-2526. doi:10.1016/j.ekir.2024.05.007. PMID: 39156164. PMCID: PMC11328578.
11. Faguer S, Chassaing N, Bandin F, Prouheze C, Garnier A, Casemayou A, et al. The HNF1B score is a simple tool to select patients for HNF1B gene analysis. Kidney Int. 2014;86(5):1007-1015. doi:10.1038/ki.2014.202. PMID: 24897035.
12. Haute Autorité de Santé. Protocole National de Diagnostic et de Soins (PNDS): Maladie liée à HNF1β. Centres de Référence des Maladies Rénales Rares (SORARE, MARHEA, NEPHROGONES), filière ORKiD; coord. Decramer S, Guigonis V. Saint-Denis La Plaine: HAS; 2020 Mar (published 2020 Apr 7) [cited 2026 Nov 6]. Available from: https://www.has-sante.fr/jcms/p_3167157
13. Online Mendelian Inheritance in Man, OMIM®. #614526 Chromosome 17q12 duplication syndrome [Internet]. Baltimore (MD): Johns Hopkins University; last edited 2012 Oct 4 [cited 2026 Oct 9]. Available from: https://www.omim.org/entry/614526
14. Orphanet. 17q12 microdeletion syndrome (ORPHA:261265) [Internet]. Paris: INSERM; [cited 2026 Oct 9]. Available from: https://www.orpha.net/en/disease/detail/261265
15. Orphanet. 17q12 microduplication syndrome (ORPHA:261272) [Internet]. Paris: INSERM; [cited 2026 Oct 9]. Available from: https://www.orpha.net/en/disease/detail/261272
16. Orphanet. HNF1B-related autosomal dominant tubulointerstitial kidney disease (ORPHA:93111) [Internet]. Paris: INSERM; [cited 2026 Oct 9]. Available from: https://www.orpha.net/en/disease/detail/93111
17. Orphanet. MODY (ORPHA:552) [Internet]. Paris: INSERM; [cited 2026 Oct 9]. Available from: https://www.orpha.net/en/disease/detail/552
18. GeneGuidance. LHX1 — identifiers and entry 17q12 microdeletion syndrome (version of 2026-09-26). Local project database.
19. 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/
20. Smajlagić D, Lavrichenko K, Berland S, Helgeland Ø, Knudsen GP, Vaudel M, et al. Population prevalence and inheritance pattern of recurrent CNVs associated with neurodevelopmental disorders in 12,252 newborns and their parents. Eur J Hum Genet. 2021;29(1):205-215. doi:10.1038/s41431-020-00707-7. PMID: 32778765. PMCID: PMC7852900.
21. Cannon S, Clissold R, Sukcharoen K, Tuke M, Hawkes G, Beaumont RN, et al. Recurrent 17q12 microduplications contribute to renal disease but not diabetes. J Med Genet. 2023;60(5):491-497. doi:10.1136/jmg-2022-108615. PMID: 36109160. PMCID: PMC10176419.
22. Pavithram A, Johansson BB, Tjora E, Svalastoga P, Mohamed KA, Koløen IL, et al. HNF1B-MODY in the Norwegian MODY Registry and the Norwegian Childhood Diabetes Registry: Clinical Insights and Prevalence Informed by Genetic and Functional Evaluation. Int J Mol Sci. 2026;27(11):5067. doi:10.3390/ijms27115067. PMID: 42278591. PMCID: PMC13257315.
23. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 390–393 (Diabetes mellitus).
24. UK Kidney Association. HNF1B Clinician Information [Internet]. Bristol: UK Kidney Association (RaDaR Rare Disease Groups); [cited 2026 Oct 9]. Available from: https://www.ukkidney.org/rare-renal/clinician-information/hepatic-nuclear-factor-1b-mutation
25. 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. PMCID: PMC10550998.
26. De Sousa SM, Davis TM, Harraway J, Greenslade M, Wu KH, Paul RG, et al. Australian and New Zealand joint society consensus statement on genetic testing for monogenic diabetes in adults. Med J Aust. 2025;223(9):484-491. doi:10.5694/mja2.52717. PMID: 40640995. PMCID: PMC12579922.
27. 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. PMCID: PMC10107883.
28. 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. PMCID: PMC11258280.
29. AbbVie Inc. CREON (pancrelipase) delayed-release capsules. Prescribing information. BLA 020725 [Internet]. Silver Spring (MD): U.S. Food and Drug Administration / DailyMed; label effective 2024 Feb 28 [cited 2026 Oct 9]. Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=073201aa-556d-4a70-918e-84e9616fd88d
30. European Rare Kidney Disease Reference Network (ERKNet) [Internet]. [cited 2026 Oct 9]. Available from: https://www.erknet.org
31. European Reference Network on Rare Endocrine Conditions (Endo-ERN). Rare genetic disorders of glucose-insulin homeostasis [Internet]. [cited 2026 Oct 9]. Available from: https://endo-ern.eu/rare-genetic-disorders-of-glucose-insulin-homeostasis/
32. European Reference Network on Rare Malformation Syndromes, Intellectual and Other Neurodevelopmental Disorders (ERN ITHACA) [Internet]. [cited 2026 Oct 9]. Available from: https://ern-ithaca.eu
33. Shin JJ, Shin HJ, Gutierrez A, Yoo N, Par-Young J, Osmani L, et al. Decreased T helper 1 cell function underlies recurrent sinopulmonary infections in the 17q12 deletion syndrome. EBioMedicine. 2025;112:105578. doi:10.1016/j.ebiom.2025.105578. PMID: 39891996. PMCID: PMC11840234.
34. University of Exeter / Royal Devon University Healthcare NHS Foundation Trust. HNF1B-MODY / RCAD [Internet]. Exeter: DiabetesGenes.org; [cited 2026 Oct 9]. Available from: https://www.diabetesgenes.org/what-is-mody/hnf1b-mody-rcad/
35. ClinicalTrials.gov [Internet]. Bethesda (MD): National Library of Medicine; [cited 2026 Oct 9]. NCT00760331, NCT04746053, NCT06065852, NCT01238250, NCT04239586, NCT06976658, NCT05586594. Available from: https://clinicaltrials.gov/
36. Bingham C, Bulman MP, Ellard S, Allen LI, Lipkin GW, Hoff WG, et al. Mutations in the hepatocyte nuclear factor-1beta gene are associated with familial hypoplastic glomerulocystic kidney disease. Am J Hum Genet. 2001;68(1):219-224. doi:10.1086/316945. PMID: 11085914. PMCID: PMC1234916.
37. Horikawa Y, Iwasaki N, Hara M, Furuta H, Hinokio Y, Cockburn BN, et al. Mutation in hepatocyte nuclear factor-1 beta gene (TCF2) associated with MODY. Nat Genet. 1997;17(4):384-385. doi:10.1038/ng1297-384. PMID: 9398836.
38. Ribeiro EM, coord. Síndromes Genéticas: Reconhecer para Tratar. Fortaleza: EdUnichristus; 2021. p. 163–167 (7.7 Doenças de adultos; Table 7.7.3 p. 165).
39. Bingham C, Ellard S, Allen L, Bulman M, Shepherd M, Frayling T, et al. Abnormal nephron development associated with a frameshift mutation in the transcription factor hepatocyte nuclear factor-1 beta. Kidney Int. 2000;57(3):898-907. doi:10.1046/j.1523-1755.2000.057003898.x. PMID: 10720943.
40. Bingham C, Hattersley AT. Renal cysts and diabetes syndrome resulting from mutations in hepatocyte nuclear factor-1beta. Nephrol Dial Transplant. 2004;19(11):2703-2708. doi:10.1093/ndt/gfh348. PMID: 15496559.
41. Bellanné-Chantelot C, Chauveau D, Gautier JF, et al. Clinical spectrum associated with hepatocyte nuclear factor-1beta mutations. Ann Intern Med. 2004;140:510-7.
42. Mefford HC, Clauin S, Sharp AJ, Moller RS, Ullmann R, Kapur R, et al. Recurrent reciprocal genomic rearrangements of 17q12 are associated with renal disease, diabetes, and epilepsy. Am J Hum Genet. 2007;81(5):1057-1069. doi:10.1086/522591. PMID: 17924346. PMCID: PMC2265663.
43. Moreno-De-Luca D, SGENE Consortium, Mulle JG, Simons Simplex Collection Genetics Consortium, Kaminsky EB, Sanders SJ, et al. Deletion 17q12 is a recurrent copy number variant that confers high risk of autism and schizophrenia. Am J Hum Genet. 2010;87(5):618-630. doi:10.1016/j.ajhg.2010.10.004. PMID: 21055719. PMCID: PMC2978962.
44. Clissold RL, Shaw-Smith C, Turnpenny P, Bunce B, Bockenhauer D, Kerecuk L, et al. Chromosome 17q12 microdeletions but not intragenic HNF1B mutations link developmental kidney disease and psychiatric disorder. Kidney Int. 2016;90(1):203-211. doi:10.1016/j.kint.2016.03.027. PMID: 27234567. PMCID: PMC4915913.
45. KDIGO Conference Participants. Genetics in chronic kidney disease: conclusions from a Kidney Disease: Improving Global Outcomes (KDIGO) Controversies Conference. Kidney Int. 2022;101(6):1126-1141. doi:10.1016/j.kint.2022.03.019. PMID: 35460632. PMCID: PMC9922534.
46. Heneberg P. HNF1B-related disease: developmental origins, molecular mechanisms, and multisystem clinical challenges. Clin Kidney J. 2026;19(7):sfag185. doi:10.1093/ckj/sfag185. PMID: 42441133. PMCID: PMC13335107.
47. Kołbuc M, Bednarek P, Motyka R, Jarmoliński T, Michalak-Kloc M, Beck BB, et al. Transient neonatal diabetes mellitus as an early diagnostic clue to HNF1B-related disease - two case reports and a literature review. Mol Cell Pediatr. 2026;13(1):20. doi:10.1186/s40348-026-00234-3. PMID: 41986760. PMCID: PMC13083689.
48. Isnard P, Makinistoglu MP, Leibovici M, Levinsohn J, Zimmermann N, Cohen C, et al. HNF1B integrates signals in a feed-forward loop driving kidney disease progression. Science. 2026;392(6795):eaea3219. doi:10.1126/science.aea3219. PMID: 41990178.
49. Adalat S, Woolf AS, Johnstone KA, Wirsing A, Harries LW, Long DA, et al. HNF1B mutations associate with hypomagnesemia and renal magnesium wasting. J Am Soc Nephrol. 2009;20(5):1123-1131. doi:10.1681/asn.2008060633. PMID: 19389850. PMCID: PMC2678044.
50. Blanchard A, Bockenhauer D, Bolignano D, Calò LA, Cosyns E, Devuyst O, et al. Gitelman syndrome: consensus and guidance from a Kidney Disease: Improving Global Outcomes (KDIGO) Controversies Conference. Kidney Int. 2017;91(1):24-33. doi:10.1016/j.kint.2016.09.046. PMID: 28003083.
51. Brackenridge A, Pearson ER, Shojaee-Moradie F, et al. Contrasting insulin sensitivity of endogenous glucose production rate in subjects with hepatocyte nuclear factor-1beta and -1alpha mutations. Diabetes. 2006;55:405-11.
52. Haldorsen IS, Vesterhus M, Raeder H, Jensen DK, Søvik O, Molven A, et al. Lack of pancreatic body and tail in HNF1B mutation carriers. Diabet Med. 2008;25(7):782-787. doi:10.1111/j.1464-5491.2008.02460.x. PMID: 18644064.
53. Ferrè S, Bongers EM, Sonneveld R, et al. Early development of hyperparathyroidism due to loss of PTH transcriptional repression in patients with HNF1β mutations? J Clin Endocrinol Metab. 2013;98:4089-96.
54. Dubois-Laforgue D, Cornu E, Saint-Martin C, Coste J, Bellanné-Chantelot C, Timsit J, et al. Diabetes, Associated Clinical Spectrum, Long-term Prognosis, and Genotype/Phenotype Correlations in 201 Adult Patients With Hepatocyte Nuclear Factor 1B (HNF1B) Molecular Defects. Diabetes Care. 2017;40(11):1436-1443. doi:10.2337/dc16-2462. PMID: 28420700.
55. Wang Y, Wang Q, Guan L, Jiang T, Wang D, Wu R, et al. A novel inversion at 17q12 disrupting HNF1B gene in a patient with renal cysts and diabetes syndrome. BMC Med Genomics. 2026;19(1):48. doi:10.1186/s12920-026-02325-6. PMID: 41703530. PMCID: PMC13014950.
56. National Center for Biotechnology Information. ClinVar [Internet]. Bethesda (MD): NCBI; [cited 2026 Oct 9]. VCV000012635, VCV000012647, VCV000635684, VCV000635698, VCV000635641, VCV000635675. Available from: https://www.ncbi.nlm.nih.gov/clinvar/
57. Amaral S, Palha A, Bogalho P, Silva-Nunes J. Maturity-onset diabetes of the young secondary to HNF1B variants (HNF1B-MODY): a series of 10 patients from a single diabetes center. Diabetol Metab Syndr. 2023;15(1):21. doi:10.1186/s13098-022-00964-0. PMID: 36793123. PMCID: PMC9930356.
58. Dotis J, Zampetoglou A, Askiti V, Bitsori M, Siomou E, Kosmeri C, et al. Clinical Spectrum and Early Renal Functional Variability in a Nationwide Greek Pediatric HNF1B Multicenter Cohort. J Clin Med. 2026;15(17):6590. doi:10.3390/jcm15176590. PMID: 42739594. PMCID: PMC13566293.
59. Zhang H, Wang C, Jiang X, Gao X, Tang X, Liu J, et al. From mutation to symptoms: a multi-center study on HNF1B-related nephropathy in Chinese children. BMC Nephrol. 2025;26(1):701. doi:10.1186/s12882-025-04616-z. PMID: 41437330. PMCID: PMC12723943.
60. Nittel CM, Dobelke F, König J, et al.; NEOCYST consortium. Review of neurodevelopmental disorders in patients with HNF1B gene variations. Front Pediatr. 2023;11:1149875.
61. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 660–665 (Deletions and duplications — microdeletions; Table 5.5 p. 663).
62. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 188–193 (Intellectual disability; Table 2.16 p. 189).
63. Rosenfeld JA, Coe BP, Eichler EE, Cuckle H, Shaffer LG. Estimates of penetrance for recurrent pathogenic copy-number variations. Genet Med. 2013;15(6):478-481. doi:10.1038/gim.2012.164. PMID: 23258348. PMCID: PMC3664238.
64. Ceravolo G, Mollica S, Cavallaro M, Ceravolo I, Sica G, Granata F, et al. 17q12 Recurrent Deletion Syndrome in Childhood. Genes (Basel). 2025;16(12):1499. doi:10.3390/genes16121499. PMID: 41465172. PMCID: PMC12733229.
65. 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/
66. Orphanet. Mayer-Rokitansky-Küster-Hauser syndrome type 2 (ORPHA:2578) [Internet]. Paris: INSERM; [cited 2026 Oct 9]. Available from: https://www.orpha.net/en/disease/detail/2578
67. Huang M, Savige J. The population frequency of predicted pathogenic genetic variants in commonly affected CAKUT genes in the general population. Pediatr Nephrol. 2026;41(10):3519-3530. doi:10.1007/s00467-026-07324-4. PMID: 42142172. PMCID: PMC13481559.
68. Verscaj CP, Velez-Bartolomei F, Bodle E, et al. Characterization of the prenatal renal phenotype associated with 17q12, HNF1B, microdeletions. Prenat Diagn. 2024;44:237-46.
69. Decramer S, Parant O, Beaufils S, Clauin S, Guillou C, Kessler S, et al. Anomalies of the TCF2 gene are the main cause of fetal bilateral hyperechogenic kidneys. J Am Soc Nephrol. 2007;18(3):923-933. doi:10.1681/asn.2006091057. PMID: 17267738.
70. Faguer S, Decramer S, Chassaing N, Bellanné-Chantelot C, Calvas P, Beaufils S, et al. Diagnosis, management, and prognosis of HNF1B nephropathy in adulthood. Kidney Int. 2011;80(7):768-776. doi:10.1038/ki.2011.225. PMID: 21775974.
71. van der Made CI, Hoorn EJ, de la Faille R, et al. Hypomagnesemia as first clinical manifestation of ADTKD-HNF1B: a case series and literature review. Am J Nephrol. 2015;42:85-90.
72. Kołbuc M, Leßmeier L, Salamon-Słowińska D, et al. Hypomagnesemia is underestimated in children with HNF1B mutations. Pediatr Nephrol. 2020;35:1877-86.
73. 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 [cited 2026 Nov 6]. Available from: https://www.has-sante.fr/jcms/p_3389585
74. Clissold RL, Fulford J, Hudson M, Shields BM, McDonald TJ, Ellard S, et al. Exocrine pancreatic dysfunction is common in hepatocyte nuclear factor 1β-associated renal disease and can be symptomatic. Clin Kidney J. 2018;11(4):453-458. doi:10.1093/ckj/sfx150. PMID: 30094008. PMCID: PMC6070112.
75. Boe N, Hedriana H, Singh K. Chromosome 17q12 deletion, including HNF1B presenting with recurrent severe cholestasis and profound hypomagnesemia during pregnancy: a case report. AJOG Glob Rep. 2026;6(2):100635. doi:10.1016/j.xagr.2026.100635. PMID: 42016181. PMCID: PMC13094637.
76. Cartault A, Paret C, Garczynski C, Costa SD, Chakhtoura Z, Ernoult P, et al. Study of the ovarian function and gyneco-obstetrical profile of patients with an HNF1B abnormality. Eur J Obstet Gynecol Reprod Biol. 2026;318:114961. doi:10.1016/j.ejogrb.2026.114961. PMID: 41570463.
77. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 722–723 (Female infertility and amenorrhoea).
78. Oram RA, Edghill EL, Blackman J, Taylor MJ, Kay T, Flanagan SE, et al. Mutations in the hepatocyte nuclear factor-1β (HNF1B) gene are common with combined uterine and renal malformations but are not found with isolated uterine malformations. Am J Obstet Gynecol. 2010;203(4):364.e1-5. doi:10.1016/j.ajog.2010.05.022. PMID: 20633866.
79. Vaez M, Montalbano S, Calle Sánchez X, Georgii Hellberg KL, Dehkordi SR, Krebs MD, et al. Population-Based Risk of Psychiatric Disorders Associated With Recurrent Copy Number Variants. JAMA Psychiatry. 2024;81(10):957-966. doi:10.1001/jamapsychiatry.2024.1453. PMID: 38922630. PMCID: PMC11209205.
80. Manasar-Dyrbuś M, Stojko R, Paul-Samojedny M, Winkowska E, Staniczek J. Prenatal Diagnosis of Renal Cysts and Diabetes Syndrome (RCAD): A Case Report. Am J Case Rep. 2026;27:e951250. doi:10.12659/ajcr.951250. PMID: 41999034. PMCID: PMC13101561.
81. Sager-La Ganga C, Solà C, Castillo K, Figueredo C, Conget I. Diabetes associated with HNF1B: beyond Occam's razor-A case report. Acta Diabetol. 2025;62(8):1347-1350. doi:10.1007/s00592-025-02472-9. PMID: 39976628. PMCID: PMC12364728.
82. American Diabetes Association Professional Practice Committee for Diabetes. 2. Diagnosis and Classification of Diabetes: Standards of Care in Diabetes—2026. Diabetes Care. 2026;49(Suppl 1):S27-S49. doi:10.2337/dc26-S002. PMID: 41358893. PMCID: PMC12690183.
83. Zuber J, Bellanné-Chantelot C, Carette C, Canaud G, Gobrecht S, Gaha K, et al. HNF1B-related diabetes triggered by renal transplantation. Nat Rev Nephrol. 2009;5(8):480-484. doi:10.1038/nrneph.2009.98. PMID: 19639018.
84. 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. PMCID: PMC13538749.
85. FARXIGA (dapagliflozin) tablets. Prescribing information. NDA 202293 [Internet]. Silver Spring (MD): U.S. Food and Drug Administration / DailyMed; label effective 2026 Jul 10 [cited 2026 Oct 9]. Available from: https://dailymed.nlm.nih.gov/dailymed/lookup.cfm?setid=01f90c94-71cb-4a1f-81ff-8004b850529b
86. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 796–799 (Urinary tract and renal anomalies — CAKUT).
87. Sanna-Cherchi S, Kiryluk K, Burgess KE, et al. Copy-number disorders are a common cause of congenital kidney malformations. Am J Hum Genet. 2012;91:987-97.
88. Gimpel C, Avni FE, Bergmann C, Cetiner M, Habbig S, Haffner D, et al. Perinatal Diagnosis, Management, and Follow-up of Cystic Renal Diseases: A Clinical Practice Recommendation With Systematic Literature Reviews. JAMA Pediatr. 2018;172(1):74-86. doi:10.1001/jamapediatrics.2017.3938. PMID: 29181500.
89. Torres VE, Ahn C, Barten TRM, Brosnahan G, Cadnapaphornchai MA, Chapman AB, et al. KDIGO 2025 clinical practice guideline for the evaluation, management, and treatment of autosomal dominant polycystic kidney disease (ADPKD): executive summary. Kidney Int. 2025;107(2):234-254. doi:10.1016/j.kint.2024.07.010. PMID: 39848746.
90. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 354–357 (Autosomal dominant polycystic kidney disease — differential diagnosis).
91. Bleyer AJ, Kidd K, Živná M, Kmoch S. Autosomal Dominant Tubulointerstitial Kidney Disease – UMOD. 2007 Jan 12 [Updated 2021 Apr 8]. 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/NBK1356/
92. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 34, section 34.3 Magnesium (34.3.4).
93. Nik-Zainal S, Strick R, Storer M, et al. High incidence of recurrent copy number variants in patients with isolated and syndromic Müllerian aplasia. J Med Genet. 2011;48:197-204.
94. Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 926–928 (MURCS association).
95. Herlin MK. Genetics of Mayer-Rokitansky-Küster-Hauser (MRKH) syndrome: advancements and implications. Front Endocrinol (Lausanne). 2024;15:1368990. doi:10.3389/fendo.2024.1368990. PMID: 38699388. PMCID: PMC11063329.
96. Kidney Disease: Improving Global Outcomes (KDIGO) Blood Pressure Work Group. KDIGO 2021 Clinical Practice Guideline for the Management of Blood Pressure in Chronic Kidney Disease. Kidney Int. 2021;99(3S):S1-S87. doi:10.1016/j.kint.2020.11.003. PMID: 33637192.
97. 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. PMCID: PMC12690181.
98. Zagaroli L, Di Berardino A, Petragnano F, Lustri S, Fasciani I, Rossi M, et al. Treatment Options for Patients with Maturity-Onset Diabetes of the Young (MODY): A Systematic Review of Literature: 2026 Update. Diabetes Ther. 2026;17(7):959-983. doi:10.1007/s13300-026-01884-2. PMID: 42295651. PMCID: PMC13388877.
99. Poitou C, Francois H, Bellanné-Chantelot C, et al. Maturity-onset diabetes of the young: clinical characteristics and outcome after kidney and pancreas transplantation in MODY3 and RCAD patients: a single center experience. Transpl Int. 2012;25:564-72.
100. Shah CV, Hammad N, Bhasin-Chhabra B, Rashidi A. SGLT2 Inhibitors in Management of Severe Hypomagnesemia in Patients Without Diabetes: A Report of 4 Cases. Kidney Med. 2023;5(9):100697. doi:10.1016/j.xkme.2023.100697. PMID: 37602145. PMCID: PMC10432792.
101. European Medicines Agency. Ozempic (semaglutide): EPAR [Internet]. Amsterdam: EMA; authorised 2018 Feb 8; last updated 2026 Aug 25 [cited 2026 Oct 9]. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/ozempic
102. European Medicines Agency. Victoza (liraglutide): EPAR [Internet]. Amsterdam: EMA; authorised 2009 Jun 30; last updated 2026 Jan 16 [cited 2026 Oct 9]. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/victoza
103. European Medicines Agency. Forxiga (dapagliflozin): EPAR [Internet]. Amsterdam: EMA; authorised 2012 Nov 11; last updated 2026 Jun 11 [cited 2026 Oct 9]. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/forxiga
104. European Medicines Agency. Jardiance (empagliflozin): EPAR [Internet]. Amsterdam: EMA; authorised 2014 May 22; last updated 2026 Mar 24 [cited 2026 Oct 9]. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/jardiance
105. Dr. Falk Pharma UK. Ursofalk 250 mg/5 ml Suspension — Summary of Product Characteristics [Internet]. Text revised 11/2024; updated on the eMC 2025 Feb 3 [cited 2026 Nov 6]. Available from: https://www.medicines.org.uk/emc/product/146/smpc
106. European Medicines Agency. Medicines data download: orphan designations [Internet]. Amsterdam: EMA; [cited 2026 Oct 9]. Available from: https://www.ema.europa.eu/en/medicines/download-medicine-data
107. O'Gorman JR, Tse Y, Reid J, Bingham C, Mohamed S. Fifteen minute consultation: managing children with HNF1B pathogenic variants. Arch Dis Child Educ Pract Ed. 2026 Aug 14 [Epub ahead of print]:edpract-2026-330293. doi:10.1136/archdischild-2026-330293. PMID: 42601188.
108. Rasmussen M, Vestergaard EM, Graakjaer J, et al. 17q12 deletion and duplication syndrome in Denmark—a clinical cohort of 38 patients and review of the literature. Am J Med Genet A. 2016;170:2934-42.
109. Stefansson H, Meyer-Lindenberg A, Steinberg S, et al. CNVs conferring risk of autism or schizophrenia affect cognition in controls. Nature. 2014;505:361-6.
110. EURORDIS – Rare Diseases Europe [Internet]. Paris: EURORDIS; [cited 2026 Oct 9]. Available from: https://www.eurordis.org
111. University of Chicago. Monogenic Diabetes Registry [Internet]. Chicago (IL): University of Chicago; [cited 2026 Oct 9]. Available from: https://monogenicdiabetes.uchicago.edu