In an open-label, 24-week, parallel-group study, 135 patients inadequately controlled with oral antihyperglycemic medications (OAMs) were treated with maximally tolerated doses of metformin and glibenclamide for at least 8 weeks and then randomized to bedtime neutral protamine Hagedorn (NPH) insulin plus maximally tolerated dose of glibenclamide BID (glib/NPH group) or insulin lispro mix 50 (50% lispro, 50% insulin lispro protamine suspension [ILPS]) pre-breakfast and lispro mix 25 (25% lispro, 75% ILPS) pre-dinner (LM50/LM25 group) (both OAMs discontinued). The LM50/LM25 group had significantly lower 2-hour postprandial BG (both meals combined) compared with glib/NPH after 12 (11.70+/-3.40 mmol/L vs. 13.15+/-2.44 mmol/L, p=0.010) and 24 weeks (11.13+/-3.31 mmol/L vs. 14.46+/-2.93 mmol/L, p =0.0001). Both regimens significantly decreased HbA1c. The reduction was greater with LM50/LM25 (-1.31+/-2% vs. -0.5+/-1.6%; P=0.01). At endpoint, the overall hypoglycemia rate increased with LM50/LM25 and decreased with glib/NPH compared with baseline (0.22+/-0.9 vs. -0.08+/-0.72 episodes/patient/30 days; p =0.037). Treatment with LM50/LM25 compared with glib/NPH in patients with inadequate control on combined OAMs yielded better postprandial and overall glycemic control with a higher rate of hypoglycemia.