It is known that the ingestion of glucose alone causes a greater increase in plasma glucose levels than ingestion of the same amount of glucose given with other nutrients. Since physiological plasma concentrations of cholecystokinin (CCK) prolong gastric emptying, it is proposed that after a meal, CCK may modify plasma glucose levels by delaying glucose delivery to the duodenum. To evaluate the effect of CCK on oral glucose tolerance, plasma CCK, insulin, and glucose levels and gastric emptying rates were measured in eight normal males before and after the ingestion of 60 g glucose with the simultaneous infusion of either saline or one of two doses of CCK-8 (12 or 24 pmol/kg per h). Gastric emptying rates were measured by gamma camera scintigraphy of technetium 99m sulfur colloid and plasma CCK levels were measured by a sensitive and specific bioassay. Basal CCK levels averaged 1.0 +/- 0.1 pM (mean +/- SEM, n = 8) and increased to 7.1 +/- 1.1 pM after a mixed liquid meal. After glucose ingestion, but without CCK infusion, CCK levels did not change from basal, and the gastric emptying t1/2 was 68 +/- 3 min. Plasma glucose levels increased from basal levels of 91 +/- 3.9 mg/dl to peak levels of 162 +/- 11 mg/dl and insulin levels increased from 10.7 +/- 1.8 microU/ml to peak levels of 58 +/- 11 microU/ml. After glucose ingestion, with CCK infused at 24 pmol/kg per h, plasma CCK levels increased to 8 pM and the gastric emptying t1/2 increased to 148 +/- 16 min. In concert with this delay in gastric emptying, peak glucose levels rose to only 129 +/- 17 mg% and peak insulin levels rose to only 24.2 +/- 4.2 microU/ml. With CCK at 12 pmol/kg per h, similar but less dramatic changes were seen. To demonstrate that endogenous CCK could modify the plasma glucose and insulin responses to oral glucose, oral glucose was given with 50 g of lipid containing long-chain triglycerides. This lipid increased peak CCK levels to 3.7 +/- 0.9 pM. Concomitant with this rise in CCK was a delay in gastric emptying and a lowering of plasma glucose and insulin values. To confirm that CCK reduced hyperglycemia by its effect on gastric motility, 36 g glucose was perfused directly into the duodenum through a nasal-duodenal feeding tube in four subjects. With duodenal perfusion of glucose, there was no change in plasma CCK levels, but plasma glucose levels increased from basal levels of 93+/-5 to 148+/-6 mg/dl and insulin levels rose from 10.6+/-3.5 to 29.5+/-5.2 microU/ml. When CCK was infused at 24 pmol/kg per h, neither the plasma glucose nor insulin responses to the duodenal administration of glucose were modified. Thus we conclude that CCK, in physiological concentrations, delays gastric emptying, slows the delivery of glucose to the duodenum, and reduces postprandial hyperglycemia. These data indicate, therefore, that CCK has a significant role in regulating glucose homeostasis in human.