OBJECTIVE This study tested a simple algorithm for beginning insulin for obese patients with type 2 diabetes after sulfonylurea failure, comparing suppertime 70/30 insulin plus continued glimepiride with insulin alone.

RESEARCH DESIGN AND METHODS This was a multicenter ambulatory randomized double-masked parallel comparison. There were 208 subjects with secondary failure to sulfonylureas who took glimepiride titrated to 8 mg b.i.d. for 8 weeks; 145 subjects with fasting plasma glucose (FPG) 180–300 mg/dl (10–16.7 mmol/1) on this treatment were randomized to placebo plus insulin (PI) or glimepiride plus insulin (GI) for 24 weeks. A dosage of 70/30 insulin before supper was titrated, seeking fasting capillary blood glucose (FBG) 120 mg/dl (6.7 mmol/1), equivalent to FPG 140 mg/dl (7.8 mmol/1). Outcome measures included FPG, HbA_1c, insulin dosage, weight, serum insulin and lipids, and adverse events.

RESULTS FPG and HbA_1c were equivalent at baseline: 261 vs. 250 mg/dl (14.5 vs. 13.9 mmol/1), and 9.9 vs. 9.7%. At 24 weeks, the FPG target was achieved in both groups (136 vs. 138 mg/dl, 7.6 vs. 7.6 mmol/1), and HbA_1c values were equal (7.7 vs. 7.6%). However, with GI, control improved faster and fewer subjects dropped out (3 vs. 15%, P < 0.01), and less insulin was needed (49 vs. 78 U/d, P < 0.001). The outcomes were alike in other respects. No subject had severe hypoglycemia.

CONCLUSIONS Injection of 70/30 insulin before supper safely restored glycemic control of type 2 diabetes not controlled by glimepiride alone. Control was restored more rapidly and with less injected insulin when glimepiride was continued.