Endogenous hyperinsulinemic hypoglycemia: Diagnostic strategies, predictive features of malignancy and long-term survival

Abstract

Diagnostic strategies, malignancy predictors and long-term survival were retrospectively evaluated in patients with hyperinsulinemic hypoglycemia (64 insulinomas). Lower median glycemia was 30 (range 20-53) mg/dl [1.6 (1.1-2.9) mmol/l] with concurrent insulin of 48 (13.2-217) microU/ml and 15 (2-46) microU/ml measured by radioimmunoassay (RIA) and immunofluorimetric assay (IFMA), respectively. All patients with insulinomas had a positive prolonged fast within 48 h. Sensitivity of localization methods was: ultrasonography (US) 23%, computed tomography (CT) 28%, magnetic resonance imaging (MRI) 65%, endoscopic US 75%, arteriography 38%, portal venous sampling 67%, selective arterial calcium stimulation 67%, intraoperative US 94% and palpation 92%. Nine patients (14%) had malignant insulinomas. Age at diagnosis (mean+/-SD, 53.8+/-19 vs 39.4+/-16.3 yr; p=0.03), insulin (1372+/-730 vs 785+/-659% (percentage of the method's diagnostic cut-off; 6 and 3 microU/ml for RIA and IFMA, respectively; p=0.007) and C-peptide levels (9.8+/-2.9 vs 3.9+/-2.8 ng/ml (3.2+/-0.9 vs 1.3+/-0.9 nmol/l; p=0.006), and tumor size (6.2+/-4.1 vs 1.5+/-0.6 cm; p=0.0002) were increased in malignant insulinomas. C-peptide level above 6.1 ng/ml (2.0 nmol/l) had a 100% sensitivity and 96% specificity, and tumor size above 2.6 cm yielded a sensitivity of 88% and specificity of 100% in predicting malignancy. Survival of patients with malignant insulinomas was significantly impaired (16 vs 100% at 5 yr; p=0.0000001). The diagnosis of insulinoma can be made within 48 h of fasting. The association between intraoperative US and palpation evidenced the tumor in 95% of the patients. C-peptide and tumor size were reliable malignancy predictors.