How do I diagnose Maturity Onset Diabetes of the Young in my patients?

Abstract Maturity Onset Diabetes of the Young (MODY) is a monogenic form of diabetes diagnosed in young individuals that lack the typical features of type 1 and type 2 diabetes. The genetic subtype of MODY determines the most effective treatment and this is the driver for MODY genetic testing in diabetes populations. Despite the obvious clinical and health economic benefits, MODY is significantly underdiagnosed with the majority of patients being inappropriately managed as having type 1 or type 2 diabetes. Low detection rates result from the difficulty in identifying patients with a likely diagnosis of MODY from the high background population of young onset type 1 and type 2 diabetes, compounded by the lack of MODY awareness and education in diabetes care physicians. MODY diagnosis can be improved through (1) access to education and training, (2) the use of sensitive and specific selection criteria based on accurate prediction models and biomarkers to identify patients for testing, (3) the development and mainstream implementation of simple criteria‐based selection pathways applicable across a range of healthcare settings and ethnicities to select the most appropriate patients for genetic testing and (4) the correct use of next generation sequencing technology to provide accurate and comprehensive testing of all known MODY and monogenic diabetes genes. The creation and public sharing of educational materials, clinical and scientific best practice guidelines and genetic variants will help identify the missing patients so they can benefit from the more effective clinical care that a genetic diagnosis brings.

MODY is typically diagnosed before the age of 30 years. In contrast to type 1 diabetes, MODY patients rarely have a severe presentation with ketoacidosis and weight loss, do not have islet autoantibodies, have significant endogenous insulin production outside of the honeymoon period and are very likely to have a parent affected with diabetes. 2 They differ from the majority of type 2 diabetes patients by their young age of diagnosis, nonobesity and absence of insulin resistance/dyslipidaemia. 3 MODY is both clinically and genetically heterogeneous, with 10 different genetic subtypes identified to date (Table 1). 31 Genetic testing of all MODY genes is essential to determine the genetic subtype as this in turn determines the patient's clinical features, clinical course and response to treatment. The majority of patients with a genetic a diagnosis of MODY (>80%) will have mutations in the HNF1A, HNF4A or GCK genes, 32 and a genetic diagnosis of these subtypes has important implications for precision medicine-based clinical care. 33 This article will therefore focus on these three subtypes.
HNF1A and HNF4A encode transcription factors that regulate the expression of genes involved in insulin secretion in the beta cell. Loss of HNF1A or HNF4A results in progressive young-onset diabetes requiring treatment to prevent complications. These patients are highly sensitive to the glucose lowering effects of sulphonylureas and may experience hypoglycaemia even at low doses. 34 Foetuses that inherit an HNF4A mutation have an increased risk of macrosomia and hyperinsulinaemic hypoglycaemia at birth. 35 In contrast, GCK mutations result in a glucose sensing defect in the beta cell that results in mild, persistent fasting hyperglycaemia from birth. Since the defect is in glucose sensing rather than regulation, the beta call counterregulates any attempts to reduce glucose levels and a stable raised glucose and HbA1c level is maintained. The GCK phenotype is strikingly similar between patients; fasting hyperglycaemia in the range of 5.5-8 mmol/L, HbA1c 40-60 mmol/L and a small postprandial increase in glucose (typically <4 mmol/L difference between fasting and 2 h values on OGTT) because the expression of the unaffected gene copy is upregulated to compensate for the loss the activity of the defective gene. 36 Because the hyperglycaemia is asymptomatic, the majority of patients are identified only when fasting blood glucose is measured, most often during pregnancy.
Patients referred for MODY testing may also have mutations in genes that cause syndromic diabetes, but the diabetes phenotype is similar to MODY and they either lack the clinical features of the syndrome because of variable expressivity, severity and penetrance, or the clinician is unaware of the significance of the features and they are not reported to the testing laboratory. 37 Mutations in the HNF1B gene (associated with a syndrome of diabetes and structural kidney disease), or the mitochondrial MT-TL1 gene mutation m.3243A>G (associated with a syndrome of diabetes and sensorineural deafness) account for~15% of the patients referred for MODY testing that receive a genetic diagnosis. 38 Patients can also be unexpectedly diagnosed with rarer conditions such as Wolfram syndrome where diabetes can be the presenting feature and the patient is tested before the onset of syndromic features. 37 Estimates of MODY prevalence will vary depending on the clinical selection criteria, population tested and the test methodology used. The most accurate estimates are those in paediatric/young adult populations where comprehensive testing of all known MODY genes has been undertaken in patients with negative islet autoantibodies and detectable C-peptide. It is thought that MODY accounts for about 2%-4% of all diabetes in children and young adults. 40 patients that have discontinued insulin treatment after many years describe feeling like they no longer have diabetes. 43 Transfer from insulin to a sulphonylurea should only be attempted once a genetic diagnosis has been confirmed and Cpeptide testing has been undertaken to provide evidence of endogenous insulin production. The patient should be advised this is a 'trial' off insulin and if unsuccessful, insulin will need to be recommenced (particularly in older patients with long duration of diabetes). HbA1c should be measured, then insulin should be stopped and a low dose started (e.g., Gliclazide 40 mg once daily) and increased as necessary. The patient should be made aware of the signs of hypoglycaemia since this can be triggered even with very low doses.

NEUROD1
Diabetes with reduced/variable penetrance that is rarely treated with insulin.
Unclear-possible haploinsufficiency although specific missense mutations in the DNA binding domain may act in a dominant negative manner. [19,20] Genes associated with syndromic subtypes of monogenic diabetes that may result in referral for MODY genetic testing in the absence of typical characteristic nondiabetic features HNF1B Renal structural disease, urogenital tract malformations, hypomagnesaemia, gout, abnormal liver function, pancreatic hypoplasia, autism (HNF1Bdeletion cases). Severity of the renal phenotype is highly variable and patients may present with isolated MODY.
Loss of function (haploinsufficiency). Whole gene deletions of HNF1Baccount for~50% of all cases. [21] MT-TL1 Bilateral sensorineural deafness. Penetrance of deafness is highly variable due to variation in heteroplasmy in specific tissues and patients may present with isolated MODY.

<1%
Autosomal dominant. Toxic gain of function from misfolded proteins caused by frameshift variants in the first 1-4 repeats of the VNTR region. [23]

WFS1
Optic atrophy, deafness, bladder dysfunction, neurological problems. Islet autoantibody negative diabetes is usually the first presenting feature in childhood and this may trigger referral for MODY testing in the absence of any other features.
Loss of function. [25,26] GATA6 Structural heart defects, pancreatic agenesis and neonatal diabetes. Very rarely patients can present with diabetes in childhood without structural defects of the heart and pancreas.

<1%
Autosomal dominant. Loss of function (haploinsufficiency). Typically null mutations are identified and can arise de novo. [27] Putative MODY genes with limited evidence for gene-disease association APPL1 Associated with a later age of onset, less severe disease and reduced penetrance for diabetes. Potentially risk variants for type 2 diabetes in combination with obesity rather than highly penetrant monogenic mutations.
Loss of function (haploinsufficiency). [28] Genes with refuted evidence for gene-disease association PAX4 Potentially risk variants for type 2 diabetes in combination with obesity rather than highly penetrant monogenic mutations.

BLK KLF11
Note: Genes are categorized into those that cause the classic MODY phenotype of isolated diabetes, those where the diabetes can resemble MODY but is part of a syndrome with characteristic nondiabetic features, and putative genes published as likely to cause MODY but require more evidence to definitively assign MODY gene status. HNF4A In addition to hyperglycaemia in childhood/young adulthood, foetuses that inherit an HNF4A mutation from either parent develop hyperinsulinism in utero. This results in an increase in birth weight by an average of 800 g and 10% of babies will have transient hyperinsulinaemic hypoglycaemia at birth. 35 Knowledge of an HNF4A mutation in either parent therefore has implications for the management of the pregnancy; 50% of foetuses will inherit the mutation and be at risk of macrosomia and hyperinsulinism.
Monitoring of foetal growth and checking for hypoglycaemia at birth is essential to prevent complications in HNF4A pregnancies.
In contrast to the severe and progressive beta cell defect seen in HNF1A and HNF4A MODY, patients with GCK-related fasting hyperglycaemia retain their counter-regulatory response to an increase in blood glucose concentration. Therefore they do not require treatment, and do not respond to oral or low dose insulin therapies; any attempt to lower blood glucose is counter-regulated by the beta cell. 51 The lifelong stable hyperglycaemia in these patients does not increase the risk of severe microvascular and macrovascular disease. 52 Therefore, it is safe to discontinue all glucose lowering therapies, blood glucose monitoring and screening for complications. In essence these patients do not have diabetes and the GCK MODY diagnosis prevents unnecessary clinical intervention.
The exception is when GCK MODY coexists with polygenic type 1 or type 2 diabetes; treatment will be required to normalize glucose and reduce complication risk, but if counter-regulatory mechanisms are still intact then glycaemic targets should take into account the raised set point for glucose homoeostasis.
The management of GCK MODY in pregnancy is more complex and is dependent on foetal genotype status. 53 If the foetus does not inherit the mutation, the foetal pancreas will sense the maternal hyperglycaemia and increase insulin secretion, leading to increased foetal growth and birth weight. If the foetus inherits the mutation, it will sense the maternal hyperglycaemia as normal (since the foetal pancreas will have the same glucose sensing defect) and there will be no increased foetal insulin secretion. Therefore, pregnancies where the foetus has inherited the mutation can be discharged from high- Laboratories should determine the thresholds for antibody positivity based on centiles derived from nondiabetic and type 1 diabetes populations to avoid false-positive results.
C-peptide testing to identify and exclude those with absolute insulin deficiency (i.e., urine C-peptide <0.2 nmol/mmol or serum/ plasma <200 pmol/L) will further improve selection, but the test has limited use at diagnosis due to the preserved insulin secretion during the honeymoon period and is most useful 3-5 years postdiagnosis. 63 The practicalities of urine C-peptide are straightforward-a random nonfasting urine sample taken into a boric acid tube will be stable for 3 days, can be taken at home and is comparable to gold standard serum testing. 64 Where sufficient resources are available, testing all children with negative antibodies and detectable C-peptide will be a highly  69 and an annual 2 day face to face and virtual training course is run by the University of Exeter. 70 MODY genetic testing is amenable to mainstreaming since a simple set of selection criteria can be used to identify patients for testing using a single assay that will test all known MODY genes. An example of a pathway for patient selection and testing is shown in Figure 1. This pathway suggests C-peptide testing before antibodies since it is less expensive and noninvasive compared to antibody testing, but for practical reasons and patient convenience it may be preferable to take samples and test for C-peptide and antibodies at the same time (particularly for newly diagnosed patients that are more likely to have detectable C-peptide during the honeymoon period). The preference for sensitivity or specificity will depend upon the capacity to perform the tests-sacrificing sensitivity (i.e., more likely to miss some cases) for increased specificity (higher proportion of tests positive due to a higher positive predictive value) may be required if only limited numbers of tests can be performed. The high sensitivity option in Figure 1 will likely detect MODY in about one in six patients tested and identify~99% of cases, whereas a high specificity pathway will make a diagnosis in about one in three patients but only identify~50% of cases. Patients with a strong clinical suspicion of MODY but a negative genetic test may be recruited into whole genome sequencing (WGS) research studies to identify novel MODY genes or mutations in noncoding regions of known MODY genes. Measuring polygenic risk scores (PRS) for type 1 and type 2 diabetes will help to further refine the selection of patients most likely to have a monogenic cause for further study. 65 Separate criteria based on FBG and HbA1c levels could be employed for selecting patients with a suspicion of GCK MODY for Sanger sequencing. Diabetes and paediatric teams might start by testing the very high probability cases to ensure a diagnosis which will lead to increased interest, motivation and enthusiasm for identifying more MODY cases once the positive impact on clinical management is recognized. It is important for clinicians to appreciate that identifying MODY is challenging due to the high prevalence of polygenic disease and it is acceptable for the majority their referrals to not receive a diagnosis (even in those with a high prior probability) and not to be disheartened by this.
Clinicians should be aware that patients with a diagnosis of MODY are still at population risk of developing polygenic autoimmune type 1 and insulin resistant type 2 diabetes. The incidence of childhood type 1 diabetes is increasing significantly in European populations, 71