OBJECTIVE: To evaluate revised criteria for polycystic ovarian morphology (PCOM) in the diagnosis of polycystic ovary syndrome (PCOS) in anovulatory infertility.

DESIGN: Prospective cohort study.

PATIENTS: WHO Group II anovulatory infertile women (n = 75).

MEASUREMENTS: Clinical, sonographic and endocrine parameters, including anti-Müllerian hormone (AMH).

RESULTS: The Rotterdam criteria for PCOM (antral follicle count (AFC) ≥12 and/or ovarian volume >10 ml) were fulfilled in 93% of the women. The PCOM prevalence was 68% when increasing the threshold to AFC >20 and 76% according to an AMH-based threshold of >35 pmol/l. The most recently proposed AFC ≥ 25 threshold reduced the PCOM prevalence to 52% (n = 39), leaving 48% (n = 36) without features of PCOM. Comparing the 36 women with non-PCOM with the 39 women in the PCOM group according to AFC ≥ 25, 22% vs 59% (P = 0·001) had serum LH >10 IU/l, 11% vs 41% (P = 0·003) had an LH/FSH ratio >2 and 19% vs 41% (P = 0·04) had hirsutism and/or elevated total testosterone, free testosterone, and/or androstenedione. The non-PCOM group included significantly more women with secondary infertility. The median AMH in the non-PCOM group was 47 pmol/l, which was twofold lower than in the PCOM group but above the upper limit of normo-ovulatory women.

CONCLUSIONS: According to a revised threshold of 25 follicles, almost half the anovulatory infertile women do not have PCOM. The characteristics of these women may be compatible with hypothalamic anovulation, but according to AMH levels, the ovaries remain multifollicular.

PERSPECTIVES: A better distinction between hypothalamic amenorrhoea and PCOS could improve treatment strategies for anovulatory infertility.