Introduction: The gold standard for insulinoma diagnosis is still 72-hours fasting trial with the aim to trigger Whipples triad. In this context, biological diagnosis of endogenous hyperinsulinism relies on the occurrence of a hypoglycemia, concomitant with inadequate high insulin and C-peptide levels. However, diagnostic cut-offs are not consensual among the different learned societies (Endocrine Society 2009, NANETS 2010, ENETS 2012). The objective of this work was thus to propose optimized cut-offs for these three parameters for the diagnosis of endogenous hyperinsulinism.
Methods: All the patients having performed a fasting trial in Cochin Hospital Endocrinology Department between February 2012 and August 2022 were included. The results of glycemia, insulin and C-peptide levels during fasting trial were collected and analyzed.
Results: One hundred and fifty-nine patients were included: 26 with endogenous hyperinsulinism and 133 without endogenous hyperinsulinism. ROC analysis of glycemia nadir during fasting trial identified the value of 2.3 mmol/l as the optimal cut-off, ensuring a sensitivity of 100% associated with a specificity of 81%. The median time to glycemia nadir was 22.5 (min 8 max 66.5) hours in the group of patients with endogenous hyperinsulinism in comparison to a median of 62.5 (min 24 max 84) hours in the group of hypoglycemic controls. ROC analysis of insulin and C-peptide levels concomitant with hypoglycemia < 2.3 mmol/l showed very good diagnostic performances of both parameters with respective cut-offs of 3.1 mUI/l (sensitivity=96%; specificity=92%) and 0.30 nmol/l (sensitivity=96%; specificity=100%). Insulin to glycemia ratio as well as C-peptide to glycemia ratio (in pmol/mmol) at the time of glycemia nadir did not show better diagnostic performances than C-peptide alone.
Conclusion: A C-peptide level 0.3 nmol/l concomitant with a hypoglycemia < 2.3 mmol/l appears as the best criterion to make the diagnosis of endogenous hyperinsulinism. Insulin level can be underestimated on hemolyzed blood samples, frequently observed in fasting trial, and thus shows lower diagnostic performances.
13 May 2023 - 16 May 2023