Clinically useful measures of insulin sensitivity and β-cell responsiveness to a glucose challenge: Differences between glucose tolerant women with polycystic ovary syndrome and normally cycling controls

Abstract

To explore the clinical utility of various indices of insulin sensitivity (IS) and β-cell function (βCF), derived from fasting levels and an oral glucose tolerance test (OGTT) either separately or in combination, in glucose tolerant non-obese and obese women with and without polycystic ovary syndrome (PCOS). Prospective cross-sectional study. Twenty-three women , 8 non-obese (BMI 21–27 kg/m2) and 15 obese (BMI 29–48) with PCOS, identified by means of elevated free androgen levels, and oligomenorreha, and normal glucose tolerance were evaluated. (M age = 23). Twelve age-matched healthy women (BMI 19–26) with normal menstrual cycles (NC), were assessed during the early follicular phase. After an overnight fast, all women were administered 75 g glucola for assessment of glucose (G) and insulin (I) at 0, 30, 60 and 120 minutes. Indices of whole-body insulin sensitivity (WBISI), pancreatic β-cell function (insulinogenic index; IGI) and disposition index (compensatory response of β-cells to insulin resistance; DI) were derived from the OGTT. Homeostasis model assessments of insulin resistance index (HOMA-IR) and β-cell function (HOMA-%B) were calculated from the fasting G and I levels. Levels of testosterone, SHBG, and LH were assayed from baseline samples. Other patient assessment included height, weight, and body fat distribution (waist-to-hip ratio[WHR], and dual-energy X-ray absorptiometry[DEXA]-calculated fat distribution index[FDI]). For all women, OGTT-derived WBISI was highly negatively correlated with fasting HOMA-IR (r=-0.72). In contrast, measures of βCF (IGI and HOMA-%B)) were highly correlated in NC but not in PCOS. A strong negative association between IGI and WBISI (r = -0.9; P < 0.0001) and positive correlation with HOMA-IR (r=0.74, P <0.001) was found in NC whereas early phase insulin secretion was not correlated with WBISI (r = -0.05) or HOMA-IR (r = -0.02) in PCOS. When PCOS was compared with NC, lower values of OGTT-derived DI values (WBISI × IGI) were found, especially in obese subjects. For both groups, FDI and WHR were highly correlated (r=0.85) whereas BMI was only mildly related (r =0.48). BMI correlated negatively with WBISI (r=-0.6) and positively with HOMA-IR (r=0.66) in PCOS whereas there was no relationship in NC women. Neither of the βCF measures were related to BMI or body habitus in both patient populations. Indices of IS and βCF derived from the OGTT most accurately assessed metabolic function in a diverse group of women with and without PCOS. Discordance between fasting and post-glucose challenge insulin levels was observed in PCOS that resulted in fasting-based indices that underestimated insulin resistance and impaired β-cell function. Poor β-cell compensation for insulin resistance was evident in both non-obese and obese women with PCOS that had a strong family history of diabetes.