The serum levels of 34K insulin-like growth factor (IGF)-binding protein were measured by RIA in 88 type 1 diabetic patients, 9 patients with type 2 diabetes, 7 patients with insulinoma, 19 normal subjects (all after an overnight fast), and 82 normal subjects after a breakfast meal. In addition, the effect of 2- to 3-h euglycemic steady state hyperinsulinemia on serum IGF-binding protein and IGF-1 levels was studied in some subjects in each of the fasted groups. Compared with normal subjects, the mean serum IGF-binding protein levels were 4-fold (P less than 0.001) higher in type I diabetic patients treated with conventional insulin injections, 2.5-fold (P less than 0.001) higher in those treated with continuous sc insulin infusion, and 2-fold (P less than 0.05) higher in patients with type 2 diabetes. In the patients with insulinoma, the mean IGF-binding protein level was 63% below normal (P less than 0.001), and it normalized after removal of the tumor. There was a slight negative correlation between the IGF-binding protein level and insulin dose in the diabetic patients (r = -0.22; P less than 0.05). In normal subjects, serum insulin concentrations were 2-fold higher (P less than 0.001) and the IGF-binding protein level was 29% lower after a meal (P less than 0.05) than in the fasting state. Serum IGF-I concentrations were virtually identical in the type 1 and 2 diabetic patients, insulinoma patients, and normal subjects. During steady state euglycemic hyperinsulinemia, the IGF-binding protein level fell by 40-70% in each group (P less than 0.001), whereas the IGF-I level declined (P less than 0.05) in the type 2 diabetic patients only. The decline of binding protein was closely related to the baseline level (r = 0.94; P less than 0.001). No correlation was found between serum IGF-I and binding protein levels in any group. In conclusion, 1) serum 34K IGF-binding protein levels are elevated in type 1 and 2 diabetic patients and decreased in patients with insulinoma; 2) the serum binding protein, but not IGF-I concentration is decreased by acute hyperinsulinemia; and 3) these data suggest that the serum insulin concentration plays a role in regulation of the serum 34K IGF-binding protein concentration.

Insulin is known to increase expression of the ob gene product leptin in adipose tissue of rodents. We determined whether insulin increases circulating leptin concentrations in humans, and whether this effect might be altered in patients with noninsulin-dependent diabetes mellitus (NIDDM). Plasma leptin concentrations were determined during an 8.5-h hyperinsulinaemic clamp (serum free insulin approximately 480 pmol/l) and during an 8.5-h infusion of physiological NaCl solution (saline) in eight normal subjects (age 51 +/- 3 years, BMI 26.3 +/- 0.6 kg/m2, fasting plasma glucose 5.6 +/- 0.2 mmol/l) and seven patients with NIDDM (age 54 +/- 2 years, 27.0 +/- 0.9 kg/m2, 11.1 +/- 0.8 mmol/l). Fasting serum insulin level correlated with plasma leptin (r = 0.72, p < 0.005), even after adjusting for the percentage of body fat (p < 0.005). During the insulin infusion, a significant increase in the plasma leptin concentration was observed after 6 h (37 +/- 14%; 5.2 +/- 0.8. vs 3.9 +/- 0.6 ng/ml, 6 vs 0 h, p < 0.05) in the normal subjects and after 8.5 h (38 +/- 11%; 7.1 +/- 1.0 vs 5.5 +/- 0.9 ng/ml, 8.5 vs 0 h, p < 0.05) in the patients with NIDDM. During the saline infusion, plasma leptin concentrations decreased significantly in the normal subjects by 11 +/- 1% (p < 0.005) and in the patients with NIDDM by 14 +/- 1% (p < 0.01) after 2 h. During the infusion of insulin as compared to saline, plasma leptin concentrations were 32 +/- 13 (p < 0.05), 53 +/- 14 (p < 0.001), 106 +/- 15 (p < 0.001) and 165 +/- 21 (p < 0.001) % higher at 2, 4, 6 and 8.5 h in the normal subjects, and 11 +/- 9 (p < 0.05), 27 +/- 10 (p < 0.05), 58 +/- 7 (p < 0.001) and 106 +/- 13 (p < 0.001) % higher in the patients with NIDDM, respectively. No differences were observed in plasma leptin concentrations between the normal subjects and patients with NIDDM, under any conditions. We conclude that prolonged exposure to insulin increases plasma leptin concentrations in humans implying a role for insulin in chronic but not acute regulation of plasma leptin concentrations. The decrease in plasma leptin concentrations during saline infusion was greater than that expected on the basis of change in serum insulin concentrations, suggesting that factors other than insulin also contribute to regulation of plasma leptin concentrations.

No abstract

Among 196 patients examined in 1972-1985 because of pernicious anemia (PA), 105 patients under the age of 76 years were invited for gastroscopic screening, and 71 patients (68%) participated. Gastroscopy revealed carcinoid tumor(s) in 5 patients (7%) but no case of carcinoma. In addition, one patient with gastric carcinoid, 5 patients with adenocarcinomas and one with a concomitant carcinoid and carcinoma had already been diagnosed earlier in the overall PA group on the basis of clinical symptoms. Thus, within a follow-up of 0-20 years (mean 7 years) the total frequency of gastric carcinoid tumors was 4% and that of carcinoma 3%. Patients with carcinoid tumors had a long duration of PA and young age of onset; these cases were not necessarily those with the highest serum gastrin levels. Even though most gastric carcinoids are small subclinical tumors of uncertain clinical significance, their unexpectedly high frequency, combined within the risk of carcinoma in PA, might indicate the need for gastroscopic follow-up, at least in cases of juvenile onset PA.