Cumulative live birth rate following ovarian stimulation with freeze-all in women with polycystic ovaries: does the PCOS phenotype have an impact?

ABSTRACT Research question Do cumulative live birth rates (CLBR) differ between polycystic ovary syndrome (PCOS) phenotypes when a freeze-all strategy is used to prevent ovarian hyperstimulation syndrome (OHSS) after ovarian stimulation (OS)? Design This is a single-centre, retrospective cohort study including 422 women with PCOS as defined by Rotterdam criteria or with PCO-like ovarian morphology-only (PCOM) in whom, between January 2015 and December 2019, a freeze-all strategy was applied after GnRH agonist triggering in the context of hyper-response defined as ≥18 follicles of ≥11 mm in their first or second IVF/ICSI cycle in a tertiary referral hospital. The primary outcome was CLBR resulting from the transfer of all cryopreserved embryos from the same IVF/ICSI cycle. Analysis was performed by multivariate logistic regression analysis adjusting for confounders. Results In total, 91/422 (21.6%) patients had PCOS phenotype A (hyperandrogenism + ovulation disorder + PCOM = HOP), 33/422 (7.8%) had phenotype C (hyperandrogenism + PCOM = HP), 161/422 (38.2%) had phenotype D (ovulation disorder + PCOM = OP) and 137/422 (32.5%) had PCOM. Unadjusted CLBR was similar among the groups (69.2%, 69.7%, 79.5% and 67.9%, respectively; p= 0.11). The multivariate logistic regression model adjusting for age, BMI, number of oocytes, fertilization rate and embryo stage at vitrification confirmed that the PCOS/PCOM phenotype did not have any impact on CLBR [OR 0.72, CI 0.24-2.14 (phenotype C); OR 1.55, CI 0.71-3.36 (phenotype D); OR 0.84, CI 0.39-1.83 (PCOM); p= 0.2, with phenotype A as reference]. Conclusions In subfertile women with PCOS eligible for ART, hyper-response after OS confers excellent cumulative live birth rates of ∼70%, irrespective of the phenotype, when a freeze-all strategy is used. This contrasts with the unfavourable clinical outcomes that have previously been observed in hyperandrogenic PCOS women who underwent mild OS targeting normal ovarian response and fresh embryo transfer. The data should however be interpreted with caution as the retrospective nature of the study holds the possibility of unmeasured confounding factors. The results cannot be generalized to all ART cycles in women with polycystic ovaries as they pertain to those cycles in which hyper-response is observed after OS.