The loss of first-phase insulin secretion is a characteristic feature of type 2 diabetic patients. The fast-acting insulin analog lispro provides a therapeutic tool for assessing the metabolic outcome of restoration of an early rise in plasma insulin levels after the ingestion of an oral glucose load. We studied eight type 2 diabetic patients on two different occasions when they received an oral glucose load (50 g) preceded by either human regular insulin or insulin analog lispro (both 0.075 U/kg lean body mass). Tritiated glucose was infused throughout the studies, and the oral glucose was labeled with [13C6]glucose for monitoring systemic and oral glucose kinetics, respectively. Basal plasma glucose (8.2 +/- 0.9 vs. 7.5 +/- 0.8 mmol/l), insulin (224 +/- 21 vs. 203 +/- 21 pmol/l), and endogenous glucose production (10.4 +/- 1.0 vs. 11.1 +/- 1.1 micromol x kg(-1) x min(-1)) were similar on both occasions. In spite of comparable incremental areas under the curve, the time course of plasma insulin concentration was much different. After injection of regular insulin, plasma insulin peaked at 120 min (368 +/- 42 pmol/l), while with lispro, the peak occurred at 60 min (481 +/- 42 pmol/l). Plasma insulin concentration during the last 3 h of the study, however, was lower with lispro compared with regular insulin. The incremental area under the curve of plasma C-peptide was lower with lispro (0.05 +/- 0.01 vs. 0.13 +/- 0.04 micromol/300 min; P < 0.01). After the ingestion of the oral glucose load, plasma glucose concentration increased by 78% at 80-100 min with regular insulin and by 62% with lispro (P < 0.05) and remained lower for the ensuing 3 h. The incremental area under the curve was 46% lower with lispro (715 +/- 109 vs. 389 +/- 109 pmol/300 min; P < 0.01). There was no difference in the two studies in the rate of appearance of the ingested glucose and in the overall rate of glucose disposal. During the initial 90 min, however, the rate of endogenous glucose production was suppressed in a prompter and more profound manner when lispro was administered (1.39 +/- 0.10 vs. 5.00 +/- 1.22 micromol x kg(-1) x min(-1); P < 0.05), while there was no difference in the late prandial phase. These results show that an early rise in plasma insulin levels after the ingestion of a glucose load is associated with a significant improvement in glucose tolerance due to a prompter, though short-lived, suppression of endogenous glucose production. This amelioration in plasma glucose profile prevents late hyperglycemia and hyperinsulinemia. Therefore, restoration of a more physiologic profile of prandial plasma insulin profile represents a rational approach for treatment of type 2 diabetic patients.