Insulin lispro is a human insulin analog that dissociates more rapidly than human regular insulin after subcutaneous injection, resulting in higher insulin levels at an earlier point in time and a shorter duration of action. The aim of the study was to evaluate if this pharmacokinetic difference would translate into better postprandial and overall control in 30 IDDM patients (age, 35.1 +/- 1.5 years; male-female ratio, 17:13; BMI, 24.8 +/- 0.5 kg/m2; HbA1c, 8.03 +/- 0.13% at baseline) treated with continuous subcutaneous insulin infusion (CSII; Disetronic H-TRON V100) in a double-blind crossover clinical study. Patients were randomized to insulin lispro or human regular insulin for 3 months before crossing over to the other insulin for another 3 months. All meal boluses were given immediately before breakfast, lunch, and supper. An eight-point blood glucose profile was measured once weekly, and HbA1c levels were measured monthly. At the end of the 3-month treatment period, HbA1c levels were significantly lower with insulin lispro, compared with human regular insulin: 7.66 +/- 0.13 vs. 8.00 +/- 0.16% (P = 0.0041). While preprandial, bedtime, and 2:00 A.M. values for blood glucose were not significantly different, 1-h postprandial blood glucose was significantly improved after breakfast, lunch, and dinner with insulin lispro, compared with human regular insulin: 8.35 vs. 9.79 mmol/l (P = 0.006), 7.58 vs. 8.74 mmol/l (P = 0.049), and 7.85 vs. 9.01 mmol/l (P = 0.03). The incidence of hypoglycemia per 30 days (blood glucose levels, <3.0 mmol/l) was 8.4 +/- 1.3 before randomization, decreasing to 6.0 +/- 0.9 for insulin lispro and to 7.6 +/- 1.3 for regular insulin during the last month of the study. Two patients in each group reported insulin precipitation. We conclude that insulin lispro improves glycemic control in CSII without increasing the risk of hypoglycemia.