Providing information for patients and professionals on research and clinical care in genetic types of diabetes.

Transferring Patients with Diabetes due to a KIR6.2 Mutation from Insulin to Sulphonylureas

Andrew Hattersley ( & Ewan Pearson   

The finding that an activating mutation in the Kir6.2 subunit immediately led to the idea that sulphonylurea treatment might be successful in these patients as it binds to the same channel and would open the KATP channel by a non-ATP dependent route.  In Gloyn et al (NEJM 2004)  we supported this idea with functional work by showing there was no response to IV glucose in patients but there was a response to IV tolbutamide (a sulphonylurea).  Since then there have been over 150 patients who have come off insulin and gone onto sulphonylureas with over 60 published (see Pearson E et al NEJM 2006 for main description and also , Sagen et al, Diabetes 2004Zung A et al JCEM 2004Codner E, et al Diabetes Care 2005Klupa T,et al Diabetologia 2005). The hypothesis that patients could stop insulin and go onto sulphonylureas is now being tested by many clinicians – this is unpublished work which is being made available to help other clinicians make this transfer. We would hope that many people will follow previously used protocols to allow the data to be compared between patients.  We also would request as much feedback as possible to help, especially if there are any problems.



61 patients with Kir6.2 mutations are not on insulin, 60 transferred from insulin with the longest being 38 months to date but one has been on sulphonylureas for 46 years from diagnosis (see Gloyn et al NEJM 04).

8 Patients have not managed to transfer to sulphonylureas - 3 had DEND syndrome with Developmental delay (severe), Epilepsy as well as Neonatal Diabetes, and 2 were adults in whom the children with the same mutation have transferred (although other adults have transferred).  The chances of transferring are reduced in adults especially if over 30 years where there has been poor glycaemic control.


Remarkably in a large multi-national study of the first 44 patients who discontinued insulin and went on to sulphonylureas, HbA1c improved in all patients (8.1% before, 6.4% after 12 weeks, p<10-11) without an increase in severe hypoglycaemia.  (see Pearson et al NEJM 2006). This supports individual patients 24 hour glucose monitoring that shows a marked reduction in the fluctuations as well as an overall lower level of glycaemic control (Sagen et al, Diabetes 2004, Zung A et al JCEM 2004). This HbA1c reduction was sustained at one year despite a reduced sulphonylurea dose. This is maintained for over 48 months and the reduction of dose with time suggests that this will be long lasting.  One patient who was on sulphonylureas since diagnosis at 3 months still has excellent control aged 46 years (Gloyn et al NEJM 2004)



Some patients show improvement in their moderate neurological / developmental delay with sulphonylureas but in no cases has this returned to normal and motor function is more improved than higher mental function (See Slingerland A et al Diabetologia 2006).  Further study of this is needed



Most have been transferred to glibenclamide where the median dose is 0.5mg/kg (range 0.13 – 1.2 mg/kg ).  After initial control is achieved glycaemic control may improve with time and the dose need to be decreased.

This is a high dose as an adult dose in type 2 diabetes for a 60kg person is 5-15mg glibenclamide = 0.08 to 0.25 mg/kg

Other sulphonylureas tried include gliclazide but this was less effective than glibenclamide in some patients.

A successful dose is likely to range from 0.1mg/kg to 1.0 mg/kg of glibenclamide (i.e. from normal adult introductory dose to up to 4 times the maximum dose). There are patients in the literature who have  been treated successfully with up to 2mg/kg/day but these doses are not recommended unless there is a clear reduction of the insulin dose after 4 weeks of 1mg/kg/day glibenclamide. Any decision to go to these higher doses is a clinical decision of the local clinician after discussion with the patient and/or their parents.


Theoretically any SU should be as effective as any other in treating the diabetes.  Gliclazide only binds to SUR1 (pancreas/neurons) whereas Glibenclamide binds to cardiac and muscle (SUR2A) as well.  We have used glibenclamide in most cases and so have the most experience of this but other sulphonylureas have been successfully used.



The doses required exceed the maximum adult dose, and glibenclamide is not licensed in children.  Reported side effects to Daonil (generic name glibenclamide) are listed on the drug information.

The commonest known side effects are skin allergies (1.5%) which may resolve.  GI side effects including diarrhoea in 1-2%  and haematological – anaemia, leucopoenia and thrombocytopaemia . 

To date, in the Kir6.2 patients patients, six patients have had diarrhoea which resolved without dose reduction, and two patients had transient mild leukopenia which resolved on re-test with no dose reduction. Monitoring of full blood count, and LFTs is advised. Three patients have had tooth discolouration with loss of enamel in two cases.  This is expected to be a previously unreported side effect.  To date it has only occurred with glibenclamide but this is the treatment used in most patients. 


Please click on the links below to download copies of the protocols for transferring patients with Kir6.2 mutations from insulin to sulphonylureas. 

***Please note these are clinical guidelines but as this is a new treatment area in diabetes - all parts of this process should be critically assessed in your patient.  If you have any questions do not hesitate to contact me:

Professor Andrew Hattersley
Office number +44 (0)1392 406806 
Secretary  +44 (0)1392 406807