OBJECTIVE To establish the effects of the short-acting insulin analog Lispro versus human regular insulin (Hum-R) on postprandial metabolic control in IDDM. RESEARCH DESIGN AND METHODS Four studies were performed in 10 C-peptide-negative IDDM patients. Lispro or Hum-R (0.15 U/kg) or Lispro + NPH (0.07 U/kg) or Hum-R + NPH were injected subcutaneously 30 min (Hum-R) or 5 min (Lispro) before lunch. Preprandial plasma glucose (PG) was maintained on all four occasions at ∼ 7.3 mmol/l by intravenous insulin. RESULTS After subcutaneous Lispro injection, plasma free insulin (FIRI) was greater between 0 and 2 h (233 ± 22 pmol/l) than after Hum-R (197 ± 25 pmol/l) but lower between 2.25 and 7 h (81 ± 10 vs. 104 ± 13 pmol/l, P < 0.05). After Lispro, PG was lower versus Hum-R for 3 h (7.4 ± 0.6 vs. 8.3 ± 0.9 mmol/l) but subsequently increased more than after Hum-R (3.25-7h, 11.3 ± 1 vs. 9.6 ± 1.2 mmol/l), resulting in a 7-h postprandial PG greater than Hum-R (9.4 ± 0.5 vs. 8.8 ± 0.6 mmol/l) (all P < 0.05). Addition of NPH to Lispro increased the 2.5-to 7-h FIRI to 110 ± 11 pmol/l and decreased the 3.25- to 7-h PG to 7.7 ± 0.8 pmol/l, resulting in 0- to 7-h PG (7.3 ± 0.3 mmol/l) lower than after Hum-R + NPH (7.9 ± 0.5 pmol/l) ( P < 0.05). CONCLUSIONS At meals, in order for Lispro to improve postprandial blood glucose not only at 2-h, but also over a 7-h period in C-peptide–negative IDDM, basal insulin must be optimally replaced.