ESPE Abstracts

Background: Insulin resistance is predominantly associated with obesity. The euglycemic hyperinsulinemic clamp test, the gold standard for diagnosis of insulin resistance, has limited practical application. Although oral glucose tolerance test (OGTT) is generally used instead, there remains a need for more practical methods in outpatient clinics.

Aim: The objective of this study is to assess the outcomes of the OGTT in non-obese adolescent girls and their relationship with clinical and biochemical characteristics.

Subjects and Methods: We retrospectively analyzed the clinical and 2-hour standard OGTT data of non-obese adolescent female patients with irregular menstruation and/or clinical hyperandrogenism, in whom other endocrine causes were excluded. Insulin resistance was considered when total insulin levels were >300 mIU/L and/or peak insulin level>100 mIU/L. Subjects were further divided to subgroups according to presence of insulin resistance, total insulin level >300 mIU/L, peak insulin level >100mIU/L and clinical features.

Results: Among the 30 participants, 25 (83%) had irregular menstruation and 16 (53.3%) had clinical hyperandrogenism. Median (IQR) age was 15.15 years (2.3), age at first menstruation 12.5 years (1.13), median weight SD -0.08 (1.77), median height SD 0.55 (2.25) and BMI percentile was 39.93 (51.44). Insulin resistance was identified in 18 subjects (60%). Patients with total insulin levels >300 mIU/L (n = 16) had higher BMI percentile values compared with the subjects with lower total insulin levels [65.38 (54.01) vs 24.3 (34.44), P = 0.013]. BMI percentiles were similar among other subgroups. Among all patients, strong positive correlations were found between total insulin levels and insulin values at the 30th minute (r_s=0.757, P <0.01), 60th minute (r_s=0.941, P <0.01), 90th minute (r_s=0.793, P <0.01), and 120th minute (r_s=0.823, P <0.01). ROC curve analysis indicated that the best cut-off insulin value for predicting insulin resistance was 66.05 mIU/L at 60th minute with 88% sensitivity and 91% specificity. On the other hand, the best cut-off insulin value for predicting total insulin levels >300 mIU/L was 74.47 mIU/L at 60th minute (93% sensitivity and 100% specificity), while it predicted peak insulin levels >100 mU/L with 82% sensitivity and 92% specificity.

Conclusion: Insulin resistance should be investigated in non-obese adolescent girls with irregular menstruation and/or clinical hyperandrogenism. Obtaining only 60th minute insulin level after oral glucose can simplify the diagnosis of insulin resistance in such cases.