The original diagnostic criteria of polycystic ovary syndrome (PCOS) were simply hyperandrogenism and menstrual irregularity. Subsequently, the association with insulin resistance and its role in the pathophysiology became recognised. Ten years ago, surveys of Australian specialists identified many who would not even assess glucose metabolism as part of PCOS assessment. Now the pendulum seems to have swung the other way with the pathophysiology of PCOS being almost solely attributed to insulin resistance. Does this ignore the heterogeneity of PCOS, and does it overlook the fact that the interaction between insulin resistance and hyperandrogenism works in both directions?