Basal androgen status as a modifiable predictor of poor ovarian response in controlled ovarian hyperstimulation

Ovarian response to controlled ovarian hyperstimulation (COH) is commonly predicted by age, ovarian reserve markers, and outcomes of previous stimulation cycles. Androgens play a key role in follicular recruitment, development, and atresia. This study aimed to evaluate circulating androgens as additional, potentially modifiable prognostic markers of ovarian response.

We conducted a retrospective cohort study analyzing oocyte yield after COH in relation to early follicular phase androgen levels. Expected ovarian response was classified according to the Bologna criteria. Odds ratios (ORs) for androgen levels and oocyte yield were calculated, and receiver operating characteristic (ROC) curves were generated to assess the predictive accuracy of androgen levels for normal versus poor ovarian response (NOR/POR).

A total of 299 stimulation cycles were analyzed (216 NOR and 83 POR). The POR group was characterized by significantly higher median age, lower anti-Müllerian hormone (AMH) levels, reduced oocyte yield, and significantly lower androgen concentrations. A significant positive correlation was observed between the number of retrieved oocytes and both AMH and testosterone levels. The odds of retrieving more than three oocytes were increased for testosterone levels >0.21 ng/mL (OR: 2.67) and for DHEAS levels >0.93 ng/mL (OR: 1.37). While AMH demonstrated the strongest predictive performance for an unfavorable outcome (oocyte yield ≤3; AUC: 0.877), androgen levels showed a moderate predictive ability for poor ovarian response (testosterone AUC: 0.659; DHEAS AUC: 0.610).

Considering the observed association between androgen levels and ovarian response, the assessment of androgens prior to COH may provide additional prognostic value. Androgens could represent modifiable biomarkers to identify patients who may benefit from pre-treatment androgen supplementation in the presence of low baseline levels.