Evidence now suggested that Jennifer did not have type 1 diabetes. We then considered whether she had evidence for type 2 diabetes. This seemed unlikely as she was lean at onset and had no other features of insulin resistance (i.e., was normotensive, no dyslipidaemia or symptoms of polycystic ovarian syndrome). The clinical picture of familial young-onset diabetes with negative β-cell antibodies, C-peptide positivity several years after diagnosis and no signs of insulin resistance was very suggestive of a diagnosis of MODY (Table 20.1).

A mutation in the gene hepatocyte nuclear factor 1-alpha (HNF1A), the commonest form of MODY in adults [1, 2], was confirmed in Jennifer, her brother and her mother. All three were reviewed in the genetic diabetes clinic and the diagnosis explained.

Those with HNF1A-MODY have been shown in an RCT [3] to be very sensitive to low-dose sulphonylurea (SU) treatment and this is the first line recommended treatment for diabetes due to HNF1A (and HNF4A) mutations [4]. Patients treated with insulin from diagnosis (on the assumption that they have type 1 diabetes) have successfully transferred to SU even many years after diagnosis [5].

Jennifer agreed to a trial of SU. Her basal bolus insulin was stopped and gliclazide 40 mg commenced (see http://​www.​diabetesgenes.​org/​content/​guidance-transferring-hnf1a-or-hnf4a-patients-insulin-sulphonylureas for advice on how to do this). Stopping insulin can cause great anxiety in those previously told they have type 1 diabetes, so daily contact with our local Genetic Diabetes Nurse [6] was offered during the transition period. Happily she was able to transfer to gliclazide with no problems. Over the next few months Jennifer lost all the weight she had gained on insulin and her HbA1c fell to 40 mmol/mol (5.8 %). She reported very infrequent hypoglycaemia.

Jennifer’s brother, Stephen, had been commenced on metformin shortly after diagnosis in the assumption that he had type 2 diabetes. He had a suboptimal HbA1c on metformin treatment (60 mmol/mol, 7.6 %) and his GP had then added gliclazide 80 mg daily. This had caused severe hypoglycaemic episodes and the gliclazide had been stopped. When he was diagnosed with HNF1A-MODY, we suggested stopping metformin and commencing a very low dose of gliclazide (20 mg increasing to 40 mg). This had the effect of controlling his diabetes without hypoglycaemia and HbA1c improved to 48 mmol/mol, 6.5 %.

Their mother, Marian, was already treated with metformin 1 g and gliclazide 40 mg daily with a good HbA1c (45 mmol/mol, 6 %), so no treatment changes were suggested.

Melanie is at 50 % risk of inheriting the HNF1A mutation from her mother. Melanie did not have diabetes, so we also discussed the pros and cons of predictive genetic testing with her. Predictive testing is a diagnostic genetic test in someone who has not yet inherited the disease. An informed decision should be made to have predictive test: clinical genetic services can be involved if necessary and we always recommend this for those under age 18. Melanie decided to go ahead with predictive testing to inform her risk of developing diabetes in the future. She was not found to carry the mutation and so her lifetime risk of developing diabetes is the same as the general population.

In non-diabetic family members who either carry the mutation or do not know their mutation status, we recommend annual screening for diabetes using either oral glucose tolerance test (OGTT; this tends to become abnormal first in HNF1A-MODY) or a combination of fasting blood glucose (FPG) and HbA1c.

Marian’s brother also has diabetes, diagnosed in his late 40s. He declined further investigation.

Five years later, Jennifer remains on gliclazide 40 mg, with no deterioration in HbA1c. She married recently and attended the diabetes clinic for pre-conception advice. She was anxious about the prospect of restarting insulin for pregnancy.

There are no studies of pregnancy outcomes in those with HNF1A-MODY, or specific advice in national guidelines, so advice was given based on evidence in type 1 and type 2 diabetes and expert opinion. Glibenclamide is considered safe in pregnancy in type 2 diabetes, although it is not currently recommended by NICE guidelines. Given that Jennifer had maintained much better control on low dose gliclazide than she did on insulin, we were happy to advise that the benefits of transferring to glibenclamide pre-conceptionally, and for as long as good control was maintained during pregnancy, would outweigh any risks associated with glibenclamide treatment. This needs to be an informed decision, made on a one to one basis with patients with HNF1A (and HNF4A-MODY), until definite outcome data is available. Jennifer remained on a low dose of glibenclamide until week 16 of her pregnancy, when her blood glucoses were rising and she agreed to go back onto basal bolus insulin. She delivered a healthy baby boy at 38 weeks and returned to low-dose gliclazide treatment.