We tested the hypothesis that as few as two weekly brief episodes of superimposed hypoglycemia (i.e., doubling the average frequency of symptomatic hypoglycemia) would reduce physiological and behavioral defenses against developing hypoglycemia and reduce detection of clinical hypoglycemia in patients with type 1 diabetes mellitus (T1DM). Compared with nondiabetic controls, six patients with well-controlled T1DM (HbA1c, 7.5 +/- 0.7% [mean +/- SD]) exhibited absent glucagon responses and reduced epinephrine (P = 0.0027), norepinephrine (P = 0.0007), pancreatic polypeptide (P = 0.0030), and neurogenic symptom (P = 0.0451) responses to hypoglycemia as expected. In these patients, 2 h of induced hypoglycemia (50 mg/dl, 2.8 mmol/l) twice weekly for 1 month, compared in a random-sequence crossover design with an otherwise identical 2 h of induced hyperglycemia (150 mg/dl, 8.3 mmol/l) twice weekly for 1 month, further reduced the epinephrine (P = 0.0001) and pancreatic polypeptide (P = 0.0030) responses, tended to further reduce the norepinephrine and neurogenic symptom responses to hypoglycemia, and reduced cognitive dysfunction during hypoglycemia (P = 0.0271), all assessed in the investigational setting. In the clinical setting, induced hypoglycemia did not alter overall glycemic control, but did reduce the total number of symptomatic hypoglycemic episodes detected by the patients from 49 to 30 per month and lowered the mean +/- SE self-monitored blood glucose level during symptomatic hypoglycemia from 51 +/- 2 mg/dl (2.8 +/- 0.1 mmol/l) to 46 +/- 3 mg/dl (2.6 +/- 0.2 mmol/l) (P < 0.01). It also reduced the proportion of low regularly scheduled self-monitored values that were symptomatic by approximately 33%. Thus as little as doubling the frequency of symptomatic hypoglycemia further reduced both the key epinephrine response and clinical awareness of developing hypoglycemia, changes reasonably expected to increase the risk of severe iatrogenic hypoglycemia in T1DM.