OBJECTIVE: Insulin glargine (21A-Gly-30Ba-L-Arg-30Bb-L-Arg-human insulin) is a biosynthetic insulin analog with a prolonged duration of action compared with NPH human insulin. This study compared insulin glargine with NPH human insulin in subjects with type 1 diabetes who had been previously treated with multiple daily injections of NPH insulin and regular insulin. RESEARCH DESIGN AND METHODS: This study was a multicenter randomized parallel-group study in which subjects were randomized to receive premeal regular insulin and either insulin glargine (at bedtime) or NPH insulin (at bedtime for patients on once-daily therapy and at bedtime and in the morning for patients on twice-daily therapy) for up to 28 weeks. Dose titration of both basal insulins was based on capillary fasting whole blood glucose (FBG) levels; the goal was a premeal blood glucose concentration of 4.4-6.7 mmol/l. RESULTS: A total of 534 well-controlled type 1 diabetic subjects (mean GHb 7.7%, mean fasting plasma glucose [FPG] 11.8 mmo/l) were treated. A small decrease in GHb levels was noted with both insulin glargine (-0.16%) and NPH insulin (-0.21%; P > 0.05). Significant reductions in median FPG levels from baseline (-1.67 vs. -0.33 mmol/l with NPH insulin, P = 0.0145) and a trend for a reduction in capillary FBG levels were achieved with insulin glargine. After the 1-month titration phase, significantly fewer subjects receiving insulin glargine experienced symptomatic hypoglycemia (39.9 vs. 49.2%, P = 0.0219) or nocturnal hypoglycemia (18.2 vs. 27.1%, P = 0.0116) with a blood glucose level <2.0 mmol/l compared with subjects receiving NPH insulin. CONCLUSIONS: Lower FPG levels with fewer episodes of hypoglycemia were achieved with insulin glargine compared with once- or twice-daily NPH insulin as part of a basal-bolus regimen in patients with type 1 diabetes.