ART outcomes in lean compared to obese phenotypes of polycystic ovarian syndrome.

RESEARCH QUESTION: To investigate differences in reproductive outcomes among IVF patients with lean compared to obese polycystic ovarian syndrome (PCOS) phenotypes. DESIGN: A retrospective cohort study of patients with PCOS who underwent IVF in a single, academically affiliated infertility center in the USA between December 2014 and July 2020. The diagnosis of PCOS was assigned based on Rotterdam criteria. Patients were designated as lean (< 25) or overweight/obese (≥ 25) PCOS phenotype based on BMI (kg/m2) at cycle start. Baseline clinical and endocrinologic laboratory panel, cycle characteristics, and reproductive outcomes were analyzed. The cumulative live birth rate included up to 6 consecutives cycles. A Cox proportional hazards model and Kaplan-Meier curve for estimating live birth rates were used to compare the two phenotypes. RESULTS: A total of 1395 patients who underwent 2348 IVF cycles were included. The mean (SD) BMI was 22.7 (2.4) in the lean and 33.8 (6.0) in the obese group (p < 0.001). A number of endocrinological parameters were similar between lean and obese phenotypes: total testosterone 30.8 ng/dl (19.5) vs 34.1 (21.9), p > 0.02 and pre-cycle hemoglobin A1C 5.33% (0.38) vs 5.51% (0.51) p > 0.001, respectively. The CLBR was higher in those with a lean PCOS phenotype: 61.7% (373/604) vs 54.0% (764/1414) respectively. Miscarriage rates were significantly higher for O-PCOS patients (19.7% (214/1084) vs 14.5% (82/563) p < 0.001) and the rate of aneuploids was similar (43.5%, 43.8%, p = 0.8). A Kaplan-Meier curve estimating the proportion of patients with a live birth was higher in the lean group (log-rank test p = 0.013). After adjusting for potential confounders, the lean phenotype was associated with an increased hazard ratio for live birth: HR = 1.38 p < 0.001. CONCLUSIONS: Lean PCOS phenotype is associated with a significantly higher CLBR compared to their obese counterparts. Miscarriage rates were significantly higher among obese patients, despite comparable pre-cycle HBA1C and similar aneuploidy rates in patients who underwent PGT-A.