Background: The aetiology of polycystic ovary syndrome (PCOS) is unknown and contested. While it has been suggested that PCOS could have origins in perturbed development, epidemiological findings have been inconclusive. We aimed to examine potential fetal origins of PCOS.
Methods: A retrospective birth cohort of 948 singleton female babies born at one hospital in South Australia in 1973-1975 was assembled. Birth characteristics were obtained from hospital records and PCOS symptoms were identified through interview and clinical examination when women were ~30 years old. Based on the combination of PCOS symptoms, women formed seven outcome groups. A multinomial logistic regression analysis was used to investigate associations between birth characteristics and these outcome groups.
Results: After adjusting for gestational age, two distinct birth characteristics were associated with two PCOS symptom groups. Each 100 g increase in birthweight increased the risk of hyperandrogenism (as a single symptom) in adulthood by 5% [relative risk ratio: 1.05, 95% confidence interval (CI): 1.01-1.09]. In contrast, each one unit increase in the ponderal index at birth decreased the risk of all three key PCOS symptoms (hyperandrogenism, menstrual dysfunction and polycystic ovaries) by 21% (0.79, 95% CI: 0.66-0.93).
Conclusions: These results suggest two discrete fetal programming pathways (related to high birthweight and to thinness at birth) are operating. Our findings point to differing aetiologies for symptom clusters, and inform the debate over symptoms that best represent the disorder.