Endocrine Abstracts

Introduction: Given the importance of establishing a reliable diagnosis of impaired glucose metabolism and diabetes mellitus (DM), knowledge on accurate and reproducible parameters is needed. However, little is known on the reproducibility of fasting and oral glucose tolerance test (OGTT) derived parameters, especially with respect to the 2-hour derived parameters.

Objectives: To determine reproducibility of parameters of glucose metabolism in a non-diabetic postmenopausal population.

Materials and methods: Thirty seven non-diabetic postmenopausal women (54±3 years, BMI 24.9±4.5 kg/m²) underwent two 75 g OGTTs with a 6 week interval. Serum C-peptide and insulin were determined using the immunoanalyzer COBAS e411 (Roche). Glucose was analysed by the hexokinase method (COBAS, Roche). Fructosamine was determined on the P-modular (Roche). To investigate reproducibility, coefficient of variation (CV), reported as median (Q1-Q3), was calculated and compared using a Mann-Whitney U test. The least significant difference (LSD) was calculated as t0.05/2.DFw*(MSw*2/n)ˆ(0.5).

Results: Fructosamine (CV=1.98 (0.87–4.73)%, LSD=8.53 μmol/l) and fasting glucose (CV=3.33(0.88–6.38)%, LSD=5.67 mg/dl) showed high reproducibility. In contrast, fasting C-peptide (CV=9.27(5.07–13.14)%, LSD=0.40 μg/l) and fasting insulin (CV=16.64(7.21–26.67)%, LSD=3.46 mU/l) had low reproducibility compared to fructosamine and fasting glucose (all P<0.001). Reproducibility of 2 h-glucose (CV=8.53(3.79–17.21)%, LSD=19.4 mg/dl), 2 h-C-peptide (CV=17.46(5.18–23.73)%, LSD=1.78 μg/l) and 2 h-insulin (CV=25.13(17.51–49.19)%, LSD=27.54 mU/l) were lower compared to fasting values (all P<0.05). Time-dependent parameters, such as early insulinogenic (CV=23.34(13.50–42.86)%) or early peptidogenic index (CV=13.17(7.19–38.12)%), generally showed lower reproducibility.

Conclusion: Whereas fasting parameters of glucose metabolism, especially fructosamine and glucose, show high reproducibility in a non-diabetic postmenopausal population, this is not the case for fasting insulin and particularly the 2-hour OGTT derived parameters. This is clearly reflected in their respective LSD, which should be kept in mind in both clinical practices as well as in interventional studies.