OBJECTIVE This 3-month study compared the effect on overall glycemic control of adding biphasic insulin aspart 30 (BIAsp30) and premixed human insulin 30/70 (BHI30) to metformin (met) in insulin-naïve, obese patients (30 males/20 females) with Type 2 diabetes (T2DM). MATERIAL/SUBJECTS: At baseline, patients had a mean age of 58.7 yr, glycated hemoglobin (HbA1c) 9.5%, and body mass index 34+/-2 kg/m2. Patients received either twice-daily BIAsp30 (no.=20) or twice-daily BHI30 (no.=30), and continued to receive maximal doses (2000 mg) of met for the duration of the study, but sulphonylurea oral antidiabetic drugs were discontinued. Primary efficacy endpoint was the change in HbA1c in both groups at study end. Safety endpoints included hypoglycemic episodes and weight gain. RESULTS Both groups reduced HbA1c by end of trial: BIAsp30 + met by 2.5% [2.16-2.86%; 95% confidence interval (CI)]; BHI30 + met by 1.18% (0.98- 1.39%; 95% CI), giving a significantly better HbA1c reduction with BIAsp30 + met (1.33%; p<0.05). Post-prandial plasma glucose decreased in both groups, by 6.38 mmol/l in patients treated with BIAsp30 + met, and by 4.34 mmol/l in those treated with BHI30 + met (p<0.05). Fasting plasma glucose also decreased in both groups, with a slightly larger decrease seen in BIAsp30 patients than in BHI30 patients (7.36 mmol/l at end of study vs 7.82 mmol/l; p=ns). Subjects treated with BIAsp30 gained less weight than those receiving BHI30 (0.3+/-0.1 kg vs 1.2+/-0.4 kg). There was no significant difference in the frequency or number of hypoglycemic episodes between groups. CONCLUSIONS Adding BIAsp30 to met in obese patients with T2DM results in better glycemic control and less weight gain than adding BHI30.