The additional value of ultrasound markers in the diagnosis of polycystic ovary syndrome

Objective: To study the value of current definitions for follicle number per ovary and ovarian volume in the diagnosis of polycystic ovary syndrome (PCOS). Design: Cross-sectional study. Subjects: Women diagnosed with PCOS after standardized screening were eligible for inclusion in the PCOS group. Women without PCOS who underwent the same screening, had regular menstrual cycles, normal hormonal values, and no other endocrine pathology were eligible for inclusion. Exposure: Not applicable. Main Outcome Measure(s): Follicle number per ovary and ovarian volume in women with PCOS, stratified by age. Linear regression models to investigate the influence of body mass index (BMI) on follicle number per ovary and ovarian volume. Differences in follicle number per ovary and ovarian volume between the PCOS phenotypes and the additional value of ovarian volume compared with follicle number per ovary. Result(s): A total of 2,492 women (16–50 years) with PCOS and 152 women without PCOS were included. Most women with PCOS up to age of 35 exhibit a follicle number per ovary ≥20 (87.8%–100%) (using an ultrasound transducer ≥8 MHz) or ≥12 (95.1%–98.6%) (using a transducer <8 MHz), followed by a decline in follicle number per ovary >35 years. Median ovarian volume was below the 10 mL cutoff in every age group, for both ultrasound transducers. Follicle number per ovary and ovarian volume were higher in women with PCOS compared with women without PCOS in every age category. In our cohort, 13/2,297 women with PCOS (0.6%) would not have received the diagnosis if ovarian volume was not considered a marker for polycystic ovarian morphology. For both ultrasound transducers, women with phenotype A (ovulatory dysfunction + hyperandrogenism + polycystic ovarian morphology) exhibited the highest follicle number per ovary and ovarian volume, followed by phenotype D (ovulatory dysfunction + polycystic ovarian morphology), then phenotype C (hyperandrogenism + polycystic ovarian morphology), and then phenotype B (ovulatory dysfunction + hyperandrogenism). No clinically significant correlation between BMI and follicle number per ovary or ovarian volume was observed. Conclusion(s): Criteria to define follicle number per ovary should be established per age category, as follicle number per ovary decreases with age. Ovarian volume shows a less clear decline with age and has a lower discriminative power, and therefore could be excluded from the diagnostic criteria. Follicle number per ovary does not need to be stratified by BMI.