Abstract / Summary
Abstract Objective: To explore the clinical heterogeneity among patients with polycystic ovary syndrome (PCOS) classified according to the Rotterdam criteria, and to evaluate the impact of personalized ovarian stimulation protocols tailored to each phenotype on ovarian response, the inflammatory microenvironment, and pregnancy outcomes in assisted reproductive technology (ART).
Method: This study is a retrospective cohort study that included 181 PCOS patients and 60 control patients who underwent ART treatment at our hospital from January 1, 2024, to January 1, 2026. The PCOS patients were classified into four phenotypes according to the Rotterdam criteria: Phenotype A (hyperandrogenism + ovulatory dysfunction + polycystic ovaries), Phenotype B (hyperandrogenism + ovulatory dysfunction), Phenotype C (hyperandrogenism + polycystic ovaries), and Phenotype D (ovulatory dysfunction + polycystic ovaries). Each phenotype group received an individualized ovarian stimulation protocol (such as the antagonist protocol, pre-treatment followed by controlled ovarian hyperstimulation, metformin combined with ovarian stimulation, mild stimulation protocol, etc.), while the control group received the standard antagonist protocol. We compared baseline characteristics, ovarian stimulation parameters, embryo quality, pregnancy outcomes, and serum levels of inflammatory cytokines (IL-6, TNF-α, CRP) among the groups, and analyzed independent risk factors influencing pregnancy rates. Baseline characteristic differences: Significant differences were observed among the groups in terms of BMI, AMH, AFC, and testosterone levels (P<0.05). Among these, phenotype B exhibited a unique profile characterized by "high metabolic risk and low ovarian reserve." It had the highest BMI (27.9 ± 2.0 kg/m²) and testosterone level (3.3 ± 1.2 nmol/L), yet its AMH level (6.2 ± 2.0 ng/mL) and AFC (4.4 ± 0.9) were significantly lower than those in other PCOS phenotypes (P<0.001). Ovarian stimulation response: Phenotype D showed the lowest gonadotropin dosage (1729.9 ± 557.3 IU); phenotype B had significantly fewer oocytes retrieved (10.1 ± 2.5) and fewer mature oocytes (MII oocytes) (7.5 ± 2.1) compared to other groups (P<0.001). Pregnancy outcomes: Phenotype D had the highest clinical pregnancy rate (63.2%) and live birth rate (54.4%), as well as the lowest incidence of OHSS (2.9%); phenotype B had the lowest live birth rate (29.4%). Phenotype A had the highest risk of OHSS (16.4%). Inflammatory markers and regression analysis: Serum levels of IL-6, TNF-α, and CRP in phenotypes A and B were significantly higher than those in other groups (P<0.001). Multivariate logistic regression analysis revealed that elevated CRP levels were an independent risk factor for clinical pregnancy rate (OR=0.672, P=0.002). After adjusting for confounding factors, the live birth rate in phenotype A was significantly lower than that in the control group (OR=0.401). P = 0.011).