Stress-induced hyperglycemia can lead to significant deterioration in glycemic control in individuals with diabetes. Previously, we have shown in normal dogs that, after intracerebroventricular (ICV) administration of carbachol (a model of moderate stress), increases in both the metabolic clearance rate (MCR) of glucose and endogenous glucose production (GP) occur. However, in hyperglycemic diabetic dogs subjected to the same stress, the MCR of glucose does not increase and glycemia therefore markedly deteriorates because of stimulation of GP. Our aims were to determine the following: 1) whether insulin-induced acute normalization of glycemia, with or without beta-blockade, would correct glucose clearance and prevent the hyperglycemic effect of stress, and 2) whether hyperinsulinemia per se could correct these abnormalities. Stress was induced by ICV carbachol in 27 experiments in five alloxan-administered diabetic dogs subjected to the following protocols in random order: 1) basal insulin infusion (BI) to restore normoglycemia; 2) basal insulin infusion with beta-blockade (BI+block); 3) normoglycemic-hyperinsulinemic clamp with threefold elevation of insulin above basal (3x BI); and 4) normoglycemic-hyperinsulinemic clamp with fivefold elevation of insulin above basal (5 x BI). The BI+block protocol fully prevented stress-induced hyperglycemia, both by increasing MCR (deltaMCR at peak: 0.72 +/- 0.25 ml x kg(-1) x min(-1) vs. no change in BI, P < 0.05) and by diminishing the stress-induced increment in GP observed in BI (deltaGP at peak: 3.72 +/- 0.09 micromol x kg(-1) x min(-1) for BI+block vs. 14.10 +/- 0.31 micromol x kg(-1) x min(-1) for BI, P < 0.0001). In contrast, 3x BI and 5x BI treatments with normoglycemic-hyperinsulinemic clamps proportionately increased basal MCR at baseline, but paradoxically were not associated with an increase in MCR in response to stress, which induced a twofold increase in GP. Thus, in alloxan-administered diabetic dogs, stress increased GP but not MCR, despite normalization of glycemia with basal or high insulin. In contrast, beta-adrenergic blockade almost completely restored the metabolic response to stress to normal and prevented marked hyperglycemia, both by limiting the rise in GP and by increasing glucose MCR. We conclude that acute normalization of glycemia with basal insulin or hyperinsulinemia does not prevent hyperglycemic effects of stress unless accompanied by beta-blockade, and we speculate that short-term beta-blockade may be a useful treatment modality under some stress conditions in patients with diabetes.