The effect of early, comprehensive genomic testing on clinical care in neonatal diabetes: an international cohort study

Abstract

Background: Traditional genetic testing focusses on analysis of one or a few genes according to clinical features; this approach is changing as improved sequencing methods enable simultaneous analysis of several genes. Neonatal diabetes is the presenting feature of many discrete clinical phenotypes defined by different genetic causes. Genetic subtype defines treatment, with improved glycaemic control on sulfonylurea treatment for most patients with potassium channel mutations. We investigated the effect of early, comprehensive testing of all known genetic causes of neonatal diabetes.

Methods: In this large, international, cohort study, we studied patients with neonatal diabetes diagnosed with diabetes before 6 months of age who were referred from 79 countries. We identified mutations by comprehensive genetic testing including Sanger sequencing, 6q24 methylation analysis, and targeted next-generation sequencing of all known neonatal diabetes genes.

Findings: Between January, 2000, and August, 2013, genetic testing was done in 1020 patients (571 boys, 449 girls). Mutations in the potassium channel genes were the most common cause (n=390) of neonatal diabetes, but were identified less frequently in consanguineous families (12% in consanguineous families vs 46% in non-consanguineous families; p<0·0001). Median duration of diabetes at the time of genetic testing decreased from more than 4 years before 2005 to less than 3 months after 2012. Earlier referral for genetic testing affected the clinical phenotype. In patients with genetically diagnosed Wolcott-Rallison syndrome, 23 (88%) of 26 patients tested within 3 months from diagnosis had isolated diabetes, compared with three (17%) of 18 patients referred later (>4 years; p<0·0001), in whom skeletal and liver involvement was common. Similarly, for patients with genetically diagnosed transient neonatal diabetes, the diabetes had remitted in only ten (10%) of 101 patients tested early (<3 months) compared with 60 (100%) of the 60 later referrals (p<0·0001).

Interpretation: Patients are now referred for genetic testing closer to their presentation with neonatal diabetes. Comprehensive testing of all causes identified causal mutations in more than 80% of cases. The genetic result predicts the best diabetes treatment and development of related features. This model represents a new framework for clinical care with genetic diagnosis preceding development of clinical features and guiding clinical management.

Funding: Wellcome Trust and Diabetes UK.

Figure 1

The paradigm shift of genetic testing Schematic representation of the steps involved in genetic testing before and after the introduction of next-generation sequencing. The red boxes indicate the role of genetic testing.

Figure 2

Different genetic causes of neonatal diabetes in patients born to non-consanguineous and consanguineous parents Comparison of genetic causes of neonatal diabetes in non-consanguineous (n=790) and consanguineous groups (n=230). Consanguinity is defined by parents being second cousins or more closely related or by the presence of 1·56% or higher total homozygosity. Genes mutated in fewer than 2·5% of patients in both cohorts were grouped in the other category (appendix).

Figure 3

A genetic diagnosis guides clinical management Schematic representation of genetic causes of neonatal diabetes and the implications of this genetic diagnosis. N indicates the number of patients identified with mutations in each of the genes in the 1020 neonatal diabetes patient cohort. Solid arrows indicate implications for most mutations in the genes. Dashed arrows indicate the implications for specific mutations.

Figure 4

Genetic diagnosis precedes development of additional clinical features defining the neonatal diabetes subtype (A) Effect of early genetic diagnosis in transient neonatal diabetes caused by 6q24 methylation defects or potassium channel gene mutations. Bar chart representing clinical features at the time of genetic testing for neonatal diabetes. Orange=diabetes, purple=diabetes remitted. (B) The effect of age at genetic testing on whether patients have non-diabetes features of Wolcott-Rallison Syndrome at the time of referral for genetic testing. Bar chart representing clinical features at the time of genetic testing for neonatal diabetes. Orange=diabetes only, light blue=diabetes and either skeletal abnormalities or liver dysfunction, dark blue=diabetes, skeletal abnormalities, and liver dysfunction. (C) The effect of age at genetic testing on whether patients with a KCNJ11 p.Val59Met mutation have neurological features at the time of referral for genetic testing. Bar chart representing clinical features at the time of genetic testing for neonatal diabetes. Orange=diabetes only, green=diabetes and neurological features.