To test the hypothesis that insulin regulates leptin, we measured the plasma leptin concentration before and during treatment of diabetic ketoacidosis (DKA), a condition characterized by extreme insulin deficiency. The study included 17 patients with type 1 diabetes (7 males and 10 females), aged 10+/-1 yr (mean +/- SE), with a body mass index of 17.6+/-1.9 kg/m2. Patients were treated with continuous insulin infusion and fluid and electrolyte replacement. Plasma leptin was measured every 6 h in the first 24 h, during which patients received a total insulin dose of 0.6-2.0 U/kg. Plasma leptin concentrations were also measured in a control group of 29 stable type 1 diabetic children (12 males and 17 females) and 25 healthy children (11 males and 14 females), aged 11+/-1 yr, with a body mass index of 18.5+/-1.1 kg/m2. Before treatment, plasma leptin concentrations were significantly lower in patients with DKA than those in diabetic and healthy controls (4.9+/-1.2 vs. 9.0+/-1.8 and 11.2+/-2.1 ng/mL, respectively; P < 0.05). In the DKA patients, plasma leptin increased to 6.4+/-1.5, 7.5+/-1.9, 9.1+/-2.7, and 8.9+/-2.5 at 6, 12, 18, and 24 h, respectively, after starting treatment (P = 0.001). Thus, leptin levels increased by 38+/-10% and 92+/-38% within 6 and 24 h of starting treatment. There was no difference in the change in plasma leptin by 24 h between subjects who could eat (n = 7) and those who could not (n = 10). The plasma leptin increase was paralleled by a rise in insulin level and a decline in glucose and cortisol levels at 6 and 24 h. In conclusion, DKA was associated with decreased plasma leptin concentrations. Treatment resulted in a significant increase in plasma leptin, which may be due to the effect of insulin on leptin production. Our data lend support to the hypothesis that insulin is the link between caloric intake and plasma leptin.