Aims/hypothesis  

In type 2 diabetic patients we compared 9 months of combination therapy with insulin glargine and metformin with 9 months of NPH insulin combined with metformin. The primary focus was changes in HbA_1c; secondary focus was diurnal glucose profiles and symptomatic hypoglycaemia.

Methods  

In this investigator-initiated open, parallel-group clinical trial involving seven centres, 110 insulin-naive type 2 diabetic patients with poor glycaemic control (HbA_1c ≥8.0%) on oral hypoglycaemic agents (90% using sulfonylurea plus metformin) were randomised to receive bedtime insulin glargine with metformin (G+MET) or bedtime NPH with metformin (NPH+MET) for 36 weeks. The patients were taught how to self-adjust their insulin dose and use a modem to send the results of home glucose monitoring to treatment centres. The goal was to achieve a fasting plasma glucose (FPG) of 4.0 to 5.5 mmol/l in both groups.

Results  

During the last 12 weeks, FPGs averaged 5.75±0.02 and 5.96±0.03 mmol/l (p<0.001) and insulin doses were 68±5 and 70±6 IU/day (0.69±0.05 and 0.66±0.04 IU kg^–1 day^–1, NS) in the G+MET and NPH+MET groups, respectively. At 36 weeks, mean HbA_1c was 7.14±0.12 and 7.16±0.14%, respectively (NS). Symptomatic, but not confirmed symptomatic, hypoglycaemia was significantly lower during the first 12 weeks in the G+MET group (4.1±0.8 episodes/patient-year) than in the NPH+MET group (9.0±2.3 episodes/patient-year, p<0.05), but not significantly different thereafter. Glucose levels before dinner were higher in the NPH+MET group (10.1±0.3 mmol/l) than in the G+MET group (8.6±0.3 mmol/l, p=0.002) throughout the 36-week study. With regard to baseline characteristics such as initial glycaemia or C-peptide, there was no difference between patients who achieved good glycaemic control (HbA_1c <7.0%) and those who did not. Differences were seen in the following: between study centres, weight gain during the run-in period and insulin therapy, and FPG during the last 12 weeks (5.7±0.2 vs 6.7±0.3 mmol/l for patients reaching vs those not reaching target, p<0.01).

Conclusions/interpretation  

Good glycaemic control can be achieved with both G+MET and NPH+MET. Use of G+MET reduces symptomatic hypoglycaemia during the first 12 weeks and dinnertime hyperglycaemia compared with NPH+MET.