Research design and methods  This was an investigator-initiated, double-blind, randomized, parallel-group, study in five centres. Patients with Type 2 diabetes (n = 88, age 56.0 ± 0.9 years, duration of diabetes 9.4 ± 0.5 years, HbA_1c 7.8 ± 0.1%, body mass index 32.4 ± 0.5 kg/m²) treated with basal insulin and metformin entered a 24-week period, during which basal insulin was titrated to optimize glucose control. Thereafter, the patients were randomized to receive either nateglinide (120 mg three times daily) or placebo before their main meals for 24 weeks.

Results  During the optimization period, HbA_1c decreased by −0.3 ± 0.1 and −0.4 ± 0.2% (NS) and insulin doses increased by 10.0 IU (2.0-32.0) [0.09 IU/kg (0.02-0.34)] and 10.0 IU (0.0-19.0) [0.11 IU/kg (0.0-0.25)] (NS) in the nateglinide and placebo groups. Mean postprandial glucose during weeks 20-24 averaged 9.0 ± 0.3 and 10.0 ± 0.3 mmol/l in the nateglinide and placebo groups (P = 0.025) and mean PPGE averaged 2.4 ± 0.2 and 3.1 ± 0.2 mmol/l, respectively (P = 0.019). At 24 weeks as compared with 0 weeks, mean HbA_1c had decreased by 0.41 ± 0.12% in the nateglinide group and by 0.04 ± 0.12% in the placebo group (P = 0.023). The frequency of confirmed, symptomatic hypoglycaemia was 7.7 episodes/patient-year vs. 4.7 episodes/patient-year in the nateglinide and placebo groups (P = 0.031).

Conclusions  Addition of a short-acting insulin secretagogue at main meals improves postprandial hyperglycaemia during combination therapy with basal insulin and metformin, but increases the frequency of hypolycaemia.