Development and temporal validation of a prediction model for live birth in infertile women with polyendocrine metabolic ovarian syndrome: a retrospective cohort study.

OBJECTIVE: To develop and temporally validate a prediction model for delivery of a live-born infant within 365 days after the initial infertility consultation in infertile women with a documented diagnosis of polycystic ovary syndrome (PCOS), which was renamed in 2026 as polyendocrine metabolic ovarian syndrome (PMOS), using routine clinical and basal endocrine variables. METHODS: This single-center retrospective cohort study included infertile women with a documented diagnosis of PCOS and an index date between December 1, 2021, and December 31, 2024. The index date was the first infertility consultation at which baseline clinical assessment and basal endocrine testing were completed. Patients were assigned to the development cohort (December 1, 2021-December 31, 2023) or the same-center temporal validation cohort (January 1-December 31, 2024). Outcome ascertainment ended on December 31, 2025. Multivariable logistic regression was used to predict delivery of a live-born infant within 365 days under routine individualized care. RESULTS: Among 610 women, 389 were in the development cohort and 221 in the temporal validation cohort. Live-birth proportions within 365 days were 44.7% and 44.3%, respectively. Complete-case model development and validation included 349 and 203 women, respectively. The final model included age, body mass index, infertility duration, gravidity, luteinizing hormone/follicle-stimulating hormone ratio, tubal/pelvic factors, and uterine factors. Older age (OR = 0.91, 95% CI: 0.85-0.98), higher body mass index (OR = 0.92, 95% CI: 0.86-0.98), longer infertility duration (OR = 0.85, 95% CI: 0.77-0.94), higher luteinizing hormone/follicle-stimulating hormone ratio (OR = 0.71, 95% CI: 0.55-0.91), and uterine factors (OR = 0.37, 95% CI: 0.15-0.80) were associated with lower live-birth probability. The AUCs were 0.735 and 0.738, and the Brier scores were 0.208 and 0.205, respectively. CONCLUSION: This model showed moderate discrimination and acceptable calibration for predicting delivery of a live-born infant within 365 days after initial infertility consultation. Multicenter external validation and impact studies are needed before broader clinical application.