In this prospective study we aimed to compare insulin plus acarbose with insulin plus gliclazide with respect to their effect on insulin requirement, lipid profiles and body mass index (BMI) while achieving good glycemic control. Forty patients with type 2 diabetes mellitus who were on conventional insulin therapy (subcutaneous insulin therapy consisting of regular and NPH insulin, two times a day) were included in the study. They were randomized to double blind treatment with insulin in combination with gliclazide or acarbose for 6 months. For both groups, acceptable glycemic control was achieved at the end of study period. The mean HbA(1c) levels decreased from 8.32+/-0.26 to 7.13+/-0.18% in acarbose group and 8. 6+/-0.15 to 7.48+/-0.21% in the gliclazide group. The difference between groups was not significant (P 0.29). In the acarbose group, total cholesterol and LDL concentration decreased significantly while other parameters did not change. In the gliclazide group, HDL levels decreased significantly from 46.6+/-2.48 mg/dl to 41.3+/-2.09 mg/dl (P 0.001) BMI increased significantly from 27.60+/-1.21 kg/m(2) to 28.69+/-1.26 kg/m(2). (P 0.003) Total daily insulin dose was not changed in the acarbose group significantly, but increased from 42.6+/-2.73 to 49.27+/-3.58 U/day, which was significant in gliclazide group of (P 0.016). In the acarbose group, there were no significant differences between responders and nonresponders with respect to fasting and stimulated C-peptide, HbA(1c) levels and baseline BMI values. But in the gliclazide group, baseline BMI values were significantly higher in the nonresponding group compared to responders (P 0.02). In conclusion, combination of insulin with acarbose can be a good alternative for type 2 diabetic patients on insulin therapy; seems more beneficial than combination with gliclazide; may have advantage of achieving good glycemic control without increasing insulin dose and BMI; also may have the advantage of providing a decrease in LDL level, which are all important to prevent atherosclerosis.