Methods
We cross-sectionally studied 65 premenopausal middle-aged women and 59 middle-aged men, whose details have been reported elsewhere.21 22 They were Japanese and biological parents of students of Mukogawa Women’s University. Subjects who reported that they were under treatment for clinically diagnosed acute or chronic inflammatory diseases, endocrine, cardiovascular, hepatic, renal diseases, hormonal contraception, and unusual dietary habits were excluded. Nobody reported receiving any medications or having regular supplements. The study was in accordance with the Declaration of Helsinki. All subjects were recruited as volunteers and gave written consent after the experimental procedure had been explained.
After a 12-hour overnight fast, at 8:30, they underwent a standard 75 g OGTT with measurement of glucose and insulin at 0 min (fasting), 30 min, 1 hour, and 2 hours after glucose ingestion. Prediabetes (impaired fasting glucose (IFG) and/or impaired glucose tolerance (IGT)) and diabetes were diagnosed based on glycemia criteria (fasting and 2-hour glucose concentrations) of the American Diabetes Association.23 OGTT revealed prediabetes in 14 women and 23 men, and newly identified diabetes in 4 men. Among 37 subjects with prediabetes, 12 had isolated IFG, 14 isolated IGT and 11 IFG/IGT.
Plasma glucose (PG), hemoglobin A1c (HbA1c), serum insulin, triglyceride, high-density lipoprotein (HDL) cholesterol, adiponectin, leptin, high-sensitivity CRP (hsCRP) and tumor necrosis factor-α (TNF-α) were measured as previously reported.24 ORM concentrations were measured in fasted samples by an immunoturbidimetric method using a commercially available kit (N Antiserum to Human α1-acid Glycoprotein, Siemens Healthcare Diagnostics, Tokyo, Japan) and an autoanalyzer (JCA-BM6010, JEOL, Tokyo, Japan). Intra-assay and interassay coefficients of variation at 87 mg/dL were 1.4% and 1.7%, respectively. The area under the concentration curves of PG and insulin (AUCg and AUCi, respectively) were calculated by the trapezoid method. Homeostasis model assessment-insulin resistance (HOMA-IR), the Matsuda Index and the Insulinogenic Index (IGI) were calculated as previously reported.25–27 The Oral Disposition Index (ODI) was calculated as the product of IGI and Matsuda Index.
Fat mass, bone mass and lean mass for arms, legs, trunk and the total body were measured using whole-body dual-energy X-ray absorptiometry (DXA) (Hologic QDR-2000, software V.7.20D, Waltham, Massachusetts, USA) as previously reported.24 The leg region included the entire hip, thigh and leg. General adiposity was assessed using body mass index (BMI), the percentage of body fat (% body fat) and Fat Mass Index (FMI), the last of which was calculated as body fat mass in kilograms divided by height in square meters. Abdominal fat accumulation was assessed by waist circumference and the ratio of trunk to leg fat.28 Skeletal muscle mass was assessed using appendicular Skeletal Muscle Mass Index (SMI) calculated as lean mass in extremities in kilograms dividing by height in square meters.
Data were presented as mean±SD unless otherwise stated. Due to deviation from normal distribution, IGI and ODI were logarithmically transformed for analyses. Comparisons between two groups were made with two-sample t-test. Differences among three groups were analyzed using analysis of variance and then Bonferroni’s multiple comparison procedure. Although there were gender differences in associations between body composition and cardiometabolic risk factors,22 middle-aged men and women did not differ in ORM and glycemia and insulinemia during OGTT (online supplemental tables 1 and 2), therefore, two groups were combined for analyses. Correlations between ORM and glycemic and other variables were investigated by Pearson’s analysis. Stepwise multivariate linear regression analyses were performed to identify most important determinants of ORM. Independent variables included were all variables that showed a significant association with ORM. Because of multicollinearity, glucose values at four time points and AUCg were analyzed in separate models. A two-tailed value of p<0.05 was considered statistically significant. Statistics were performed with SPSS system V.17.0 (SPSS, Chicago, Illinois, USA).