A new classification plot for the C-peptide suppression test

Abstract

Context and objective: To evaluate the C-peptide suppression test as a screening test in patients with symptoms of hypoglycemia as compared to the standard fasting test.

Design: Retrospective discriminant analysis of data from C-peptide suppression tests.

Setting: Clinical study.

Patients: Patients with insulinomas and patients without insulinomas but having symptoms compatible with hypoglycemia.

Interventions: The results from C-peptide suppression tests of 26 patients with insulinomas and 100 patients without insulinomas were compared.

Main outcome measures: A classification plot which introduces two discriminant parameters for the C-peptide suppression test: the ratio of [blood glucose]/[C-peptide] at the lowest C-peptide concentration and mean glycemia during insulin infusion.

Results: In patients with insulinomas, minimal serum C-peptide levels were higher (1.81+/- 0.87 ng/mL; median 1.83 ng/mL; maximal suppression 37 +/- 24% of basal C-peptide levels) as compared to patients without insulinoma (0.40 +/- 0.15 ng/mL; median 0.30 ng/mL; maximal suppression of 75 +/- 9%; P<0.001). Mean glycemia during the test was lower in patients with insulinomas (30.8 +/- 3.3 vs. 47.5 +/- 8.3 mg/dL; P<0.001) as was the [blood glucose]/[C-peptide] ratio (21.9 +/- 14.6 vs. 139.2 +/- 43.8; P<0.001). Discriminant analysis revealed a specificity of 96% to rule out the diagnosis of 'insulinoma' at a 1% probability threshold with a sensitivity of 100%.

Conclusions: We developed a new classification plot for the C-peptide suppression test in order to accurately identify those patients whose symptoms of hypoglycemia are not due to endogenous hyperinsulinemia/insulinomas. Thus, the need for fasting tests and hospitalization costs can be reduced.