We sought 1) to determine which symptoms of hypoglycemia are reproducible, 2) to pharmacologically distinguish neurogenic (autonomic) from neuroglycopenic symptoms, and 3) to test the hypothesis that awareness of hypoglycemia is the result of perception of neurogenic rather than neuroglycopenic symptoms. Awareness of hypoglycemia and 19 symptoms were quantitated in 10 normal, young adults, each studied on four occasions in random sequence, during 1) clamped euglycemia (approximately 5 mM), 2) clamped hypoglycemia (approximately 2.5 mM), 3) clamped hypoglycemia with combined alpha- and beta-adrenergic blockade (phentolamine and propranolol), and 4) clamped hypoglycemia with pan-autonomic blockade (phentolamine, propranolol and atropine). Significant (ANOVA, P < 0.001) treatment effects on the awareness of hypoglycemia ("blood sugar low") were noted. No change occurred in the score for this during euglycemia, but the mean +/- SE increase was 2.1 +/- 0.4 during hypoglycemia. This increase was not reduced significantly by adrenergic blockade (1.6 +/- 0.5), but was reduced significantly and substantially (approximately 70%) by pan-autonomic blockade (0.6 +/- 0.3). Significant neurogenic symptoms included shaky/tremulous (P < 0.001), heart pounding (P < 0.001), and nervous/anxious (P = 0.002), all adrenergic; and sweaty (P < 0.001), hungry (P < 0.001), and tingling (P = 0.009), all cholinergic. Significant neuroglycopenic symptoms, those produced by hypoglycemia but not reduced by pan-autonomic blockade, included warm (P < 0.001), weak (P = 0.011), difficulty thinking/confused (P = 0.004), and tired/drowsy (P = 0.003). We conclude that muscarinic cholinergic mechanisms mediate an important and previously uncharacterized component of the neurogenic symptoms of hypoglycemia and awareness of hypoglycemia.(ABSTRACT TRUNCATED AT 250 WORDS)