Prevalence and correlates of diagnosed and probable polycystic ovary syndrome (PCOS) in a cohort of parous women

Abstract

Correlates of diagnosed and probable polycystic ovary syndrome (PCOS) among parous women were assessed in this study. A total of 557 women were recruited from multi-specialty clinics in eastern Massachusetts. The women were categorized as being diagnosed with PCOS based on medical records and self-reported clinician-diagnoses. A category of "probable PCOS" was created for women without a diagnosis but with ≥ 2 of the following: ovulatory dysfunction (cycle length < 21 or ≥ 35 days), hyperandrogenism (free testosterone concentration > 75th percentile), or elevated anti-Müllerian hormone (AMH) concentration (> 75th percentile). The remaining participants were placed in the "no PCOS" category, and characteristics were compared across groups. Of the total cohort, 9.7% had diagnosed and 9.2% had probable PCOS. The frequency of irregular cycles was similar for diagnosed and probable PCOS. Free testosterone and AMH levels were higher in women with probable than with diagnosed PCOS. Frequency of irregular cycles and both hormones were higher for the 2 PCOS groups vs the no PCOS group. Obesity prevalence for diagnosed PCOS was twice that of probable PCOS (43.9% vs 19.6%), yet the 2 groups had similar HbA1c and adiponectin values. Women with probable PCOS are leaner but have comparable glycemic traits to those with a formal diagnosis, highlighting the importance of assessing biochemical profiles among women with irregular cycles, even in the absence of overweight/obesity.

Keywords: Rotterdam criteria; anti-Mullerian hormone; cohort; hyperandrogenism; irregular menstrual cycles; oligo-anovulation; parous; polycystic ovary syndrome (PCOS).