ABSTRACT
The evidence that miscarriage might be more common in women with PCOS compared to average started to appear in the 1980s soon after the application of high-resolution ultrasound. In reviewing this literature we need to consider the strategies that can be used to explore the association. The studies differ in their starting clinical source. One experiment would be to identify a large group of women planning to conceive and to classify them according to PCOS status, and then to compare this diagnosis with the miscarriage rate. This type of design was used in an early paper using serum luteinizing hormone (LH) rather than PCOS as a clinical marker.1 Of course, large numbers of women would be required if we estimate that 5% have PCOS (as opposed to the percentage who show polycystic ovaries on ultrasound), of whom 20% may miscarry. Thus a study group of 20 women with PCOS and miscarriage would require a starting population of 2000 women.
