Insulin resistance in a large cohort of women with polycystic ovary syndrome: a comparison between euglycaemic-hyperinsulinaemic clamp and surrogate indexes

STUDY QUESTION: Could surrogate indexes identify insulin resistant individuals among women with polycystic ovary syndrome (PCOS)? SUMMARY ANSWER: Surrogate indexes may be able to rule in, but not rule out, insulin resistance in women with PCOS. WHAT IS KNOWN ALREADY: Insulin resistance is a typical finding of women with PCOS and most clinical information on this issue is based upon surrogate indexes of insulin resistance. However, data on the performance of these indexes in PCOS women are very limited. STUDY DESIGN SIZE, DURATION: A retrospective analysis of 406 women referred to our outpatient clinic for hyperandrogenism and/or menstrual dysfunction and submitted to hyperinsulinemic euglycaemic clamp between 1998 and 2015. PARTICIPANTS/MATERIALS, SETTING, METHODS: In total, 375 of these women had PCOS by the Rotterdam criteria and were included in the study. Six surrogate indexes of insulin sensitivity were calculated from glucose and insulin levels, either at fasting (homeostasis model assessment (HOMA), glucose/insulin (G/I) ratio and quantitative insulin sensitivity check index (QUICKI)) or after oral glucose load (Gutt, Stumvoll0-120 and Matsuda). MAIN RESULTS AND THE ROLE OF CHANCE: Overall, insulin resistance, as identified by the M-clamp value, was found in 74.9% of these women. The percentage was 59.3% in normal-weight vs 77.5% in overweight and 93.9% in obese subjects. All surrogate indexes were highly correlated with the M-clamp values. However, their ability to identify insulin resistant individuals was limited, in terms of sensitivity and especially in normal-weight subjects. ROC analysis showed similar performances of these indexes (AUC values 0.782-0.817). LIMITATIONS REASONS FOR CAUTION: Potential referral bias of PCOS patients may have caused overestimation of the prevalence of insulin resistance in these women. WIDER IMPLICATIONS OF THE FINDINGS: By using surrogate indexes many subjects with PCOS may be erroneously diagnosed as insulin sensitive, especially among normal-weight women. These indexes can be used to rule in, but not rule out, insulin resistance in PCOS.