Insulin resistance was estimated using HOMA-IR, which CHIR98014 mw was defined as follows: (FPI (μU/mL) × FPG (mmol/L))/22.5. In addition, we estimated insulin sensitivity in the subjects using the three most extensively validated OGTT insulin sensitivity indices against the euglycemic clamp technique in a relatively large numbers of subjects (ISIcomp [13], MCRest [14], and OGIS [15]). To estimate β-cell

function, HOMA-B% was calculated as follows: (20 × FPI)/(FPG − 3.5). The insulinogenic index was defined as the ratio of insulin change to plasma glucose change 30 min after a 75-g oral glucose load (Δ insulin, 0–30 min/Δ plasma glucose, 0–30 min) and was used to estimate early phase insulin secretion. In addition, the area under the curve (AUC) of glucose or insulin levels during the OGTT was calculated by the trapezoidal rule, and the ratio of the total AUC insulin to the total AUC glucose (total AUC insulin/glucose) was used to measure the summation of the total insulin secretory capacity [16]. The disposition index

was defined as the product of the insulinogenic index and Matsuda’s index and was used for estimating the insulin secretory capacity adjusted for insulin resistance. The plasma glucose levels were determined using the hexokinase method in an autoanalyzer (Hitachi, Tokyo, Japan), which had a CV of 1.7%. The plasma insulin (Biosource, Nivelles, Belgium) and C-peptide levels (Immunotech, Czech Republic) were determined using immunoradiometric assays with intra- Adriamycin mouse and inter-assay CVs of 1.6–2.2% and 6.1–6.5% and 2.3–3.0% and 3.5–5.1%, respectively. The plasma total osteocalcin was measured with an IRMA method using an Osteo-RIACT kit from Cis Bio International (Saclay, France), which had intra- and inter-assay CVs of 1.2–2.8% and 3.6–5.2%, respectively. Total plasma adiponectin and leptin levels were measured by ELISA kits (R&D Systems, Minneapolis, MN, USA), as recommended by the manufacturer. Statistical methods All data are presented as the means ± SDs or proportions, except for skewed variables, which were presented as the median why (interquartile range, 25–75%). Because the

distributions of fasting and 2-h plasma insulin levels, AUC insulin, AUC insulin/glucose, HbA1c level, HOMA values, insulinogenic index, disposition index, adiponectin level, and leptin level were skewed as assessed by the Kolmogorov–Smirnov test, the natural logarithmic transformation was applied in the statistical analysis. In the interests of simplicity, nontransformed median Ku-0059436 concentration values are presented in the tables and text. One-way ANOVA, followed by Turkey’s post hoc test, was used to compare the means between the tertiles of osteocalcin levels. Pearson correlation coefficients were calculated to evaluate the associations between osteocalcin and age, body mass index (BMI), and metabolic parameters (glucose, insulin, and insulin secretory and insulin sensitivity indices).