Linlin Cui
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Linlin Cui has a total of 192 co-authors. Here are the top co-authors by number of shared publications:
Zi-Jiang Chen
State Key Laboratory of Reproductive Medicine and Offspring Health, Center for Reproductive Medicine, Institute of Women, Children and Reproductive Health, Shandong University, Jinan, China
19 shared publicationsJingmei Hu
State Key Laboratory of Reproductive Medicine and Offspring Health, Center for Reproductive Medicine, Institute of Women, Children and Reproductive Health, Shandong University, Jinan, China
15 shared publicationsJingyu Li
State Key Laboratory of Reproductive Medicine and Offspring Health, Center for Reproductive Medicine, Institute of Women, Children and Reproductive Health, Shandong University, Jinan, China
10 shared publicationsYiyuan Zhang
Center for Reproductive Medicine, The Second Qilu Hospital of Shandong University, Jinan, Shandong, 250012, China
10 shared publications- Zi-Jiang Chen
Department of Reproductive Medicine, Ren Ji Hospital, Shanghai Jiao Tong University School of Medicine
10 shared publications Bingqian Zhang
8 shared publications
Latest Publications
- Open Access
- 19 October 2026
- Published by Springer Nature in BMC Pregnancy and Childbirth
Low-molecular-weight heparin for frozen-thawed embryo transfer in women with adenomyosis: a parallel randomized pilot study
- 19 October 2026
- Published by Springer Nature in BMC Pregnancy and Childbirth
Open AccessLow-molecular-weight heparin for frozen-thawed embryo transfer in women with adenomyosis: a parallel randomized pilot study
- Open Access
- 18 May 2026
- Published by Springer Nature in BMC Pregnancy and Childbirth
- Vol. 26 (1)
Do the children born after in vitro fertilization or intracytoplasmic sperm injection have poor intellectual performance? A prospective cohort study
Assisted reproductive technologies (ART) have been proven to be associated with the increased risk of pregnancy complications such as gestational hypertens
- 18 May 2026
- Published by Springer Nature in BMC Pregnancy and Childbirth
- Vol. 26 (1)
Open AccessDo the children born after in vitro fertilization or intracytoplasmic sperm injection have poor intellectual performance? A prospective cohort study
Assisted reproductive technologies (ART) have been proven to be associated with the increased risk of pregnancy complications such as gestational hypertens
- 13 May 2026
- Published by Springer Nature in Phenomics
- Vol. 6 (3) , 321-331
Cohort Profile: The Chinese Ovarian Aging Cohort (COACH)
- 13 May 2026
- Published by Springer Nature in Phenomics
- Vol. 6 (3) , 321-331
Cohort Profile: The Chinese Ovarian Aging Cohort (COACH)
- 29 April 2026
- Published by Elsevier in Journal of Reproductive Immunology
- Vol. 176, 104902
Age-dependent effects of low-molecular-weight heparin in recurrent implantation failure: A retrospective cohort study
- 29 April 2026
- Published by Elsevier in Journal of Reproductive Immunology
- Vol. 176, 104902
Age-dependent effects of low-molecular-weight heparin in recurrent implantation failure: A retrospective cohort study
- Open Access
- 1 March 2026
- Published by Taylor & Francis in Clinical Epidemiology
- Vol. ume 18, 1-14
Development and External Validation of Machine Learning Model to Predict Live Birth Following Assisted Reproductive Technology in Women with Ovarian Endometriomas: A Decision-Support Tool
Background: Ovarian endometriomas damage the ovarian structure, alter ovarian inflammation, and impair ovarian reserve. Given the conflicting results, determining an optimal reproductive strategy for women with endometriomas―whether expectant management, medication, surgery, or assisted reproductive technology (ART)―remains challenging. Objective: This study aims to preliminarily develop and validate clinically applicable decision-support tools by training, testing, and validating an automated machine learning (ML) model to predict the likelihood of live birth following ART in women with endometriomas. Methods: The derivation and testing cohort included 1705 women, and the external validation cohort included 1475 women with ovarian endometriomas following ART retrospectively. Two ML models were developed and validated to predict the probability of live birth. Model performance was evaluated using the area under the curve (AUC), accuracy, sensitivity, specificity, F1 score, and Brier score. The SHapley Additive exPlanations (SHAP) method was employed to interpret feature importance. Results: Comparing seven ML algorithms, the Extreme Gradient Boosting (XGBoost) demonstrated superior predictive performance both in model-1 and model-2, achieving an AUC of 0.90 [95% confidence interval (CI): 0.88– 0.92] and 0.88 (95% CI: 0.86– 0.89) in test-datasets and 0.80 (95% CI: 0.76– 0.83) and 0.69 (95% CI: 0.65– 0.73) in external validation cohort. The SHAP analysis revealed that the age and features associated with ovarian reserve had strong predictive power and the ovarian endometriomas had limited predictive power. Conclusion: Model-2, which uses only pre-ART variables, can support reproductive strategy selection prior to ART initiation. Conversely, Model-1 is designed to support embryo transfer strategy option after oocyte retrieval, incorporating post-ART data. Although both models show promise as decision-support tools for personalizing infertility treatment in women with endometriomas, their clinical implementation awaits confirmation from prospective, multicenter validation.
- 1 March 2026
- Published by Taylor & Francis in Clinical Epidemiology
- Vol. ume 18, 1-14
Open AccessDevelopment and External Validation of Machine Learning Model to Predict Live Birth Following Assisted Reproductive Technology in Women with Ovarian Endometriomas: A Decision-Support Tool
Background: Ovarian endometriomas damage the ovarian structure, alter ovarian inflammation, and impair ovarian reserve. Given the conflicting results, determining an optimal reproductive strategy for women with endometriomas―whether expectant management, medication, surgery, or assisted reproductive technology (ART)―remains challenging. Objective: This study aims to preliminarily develop and validate clinically applicable decision-support tools by training, testing, and validating an automated machine learning (ML) model to predict the likelihood of live birth following ART in women with endometriomas. Methods: The derivation and testing cohort included 1705 women, and the external validation cohort included 1475 women with ovarian endometriomas following ART retrospectively. Two ML models were developed and validated to predict the probability of live birth. Model performance was evaluated using the area under the curve (AUC), accuracy, sensitivity, specificity, F1 score, and Brier score. The SHapley Additive exPlanations (SHAP) method was employed to interpret feature importance. Results: Comparing seven ML algorithms, the Extreme Gradient Boosting (XGBoost) demonstrated superior predictive performance both in model-1 and model-2, achieving an AUC of 0.90 [95% confidence interval (CI): 0.88– 0.92] and 0.88 (95% CI: 0.86– 0.89) in test-datasets and 0.80 (95% CI: 0.76– 0.83) and 0.69 (95% CI: 0.65– 0.73) in external validation cohort. The SHAP analysis revealed that the age and features associated with ovarian reserve had strong predictive power and the ovarian endometriomas had limited predictive power. Conclusion: Model-2, which uses only pre-ART variables, can support reproductive strategy selection prior to ART initiation. Conversely, Model-1 is designed to support embryo transfer strategy option after oocyte retrieval, incorporating post-ART data. Although both models show promise as decision-support tools for personalizing infertility treatment in women with endometriomas, their clinical implementation awaits confirmation from prospective, multicenter validation.
- Open Access
- 1 February 2026
- Published by Taylor & Francis in International Journal of Women's Health
- Vol. ume 18, 1-10
Causal Association Between Sleep Traits and Endometriosis: A Mendelian Randomization Study
Previous studies suggested that irregular sleep-wake rhythms increased the risk of endometriosis. Mechanistically, sleep pattern dysregulation may promote the pathogenesis of endometriosis via neuroendocrine disturbances. However, the causal relationship between sleep traits and endometriosis remains unclear. We aim to investigate the potential causal associations between sleep traits and endometriosis, including its specific anatomical subtypes. Two-sample Mendelian Randomization (MR) analyses were conducted. The primary exposure was insomnia, and the secondary exposures were getting up in the morning, nap during day, sleep disorders, chronotype, snoring, short sleep duration, and long sleep duration. The outcomes were endometriosis along with its anatomical subtypes: uterine, ovarian, vaginal, intestinal, peritoneal, and unspecified endometriosis. We applied inverse variance weighted (IVW), weighted median, simple mode, weighted mode, and MR-Egger regression to estimate causal relationships between sleep traits and the risk of endometriosis. Additionally, a series of sensitivity and validation analyses were performed to ensure robustness. For the primary exposure, IVW results suggested that insomnia was significantly associated with elevated risks of endometriosis (OR=1.83, 95% CI: 1.34–2.50), uterine endometriosis (OR=2.22, 95% CI: 1.30–3.80), ovarian endometriosis (OR=1.61, 95% CI: 1.03–2.54), vaginal endometriosis (OR > 1.00, 95% CI: 1.03–3.79) and unspecified endometriosis (OR=2.70, 95% CI: 1.45–5.02). For the secondary exposure, most sleep traits showed no significant associations with endometriosis risk. Our study demonstrated a causal relationship between insomnia and elevated risks of endometriosis across multiple anatomical subtypes. While no significant associations were observed for other sleep traits, the robust link with insomnia highlights the potential clinical value of sleep-focused interventions in mitigating the risk of endometriosis.
- 1 February 2026
- Published by Taylor & Francis in International Journal of Women's Health
- Vol. ume 18, 1-10
Open AccessCausal Association Between Sleep Traits and Endometriosis: A Mendelian Randomization Study
Previous studies suggested that irregular sleep-wake rhythms increased the risk of endometriosis. Mechanistically, sleep pattern dysregulation may promote the pathogenesis of endometriosis via neuroendocrine disturbances. However, the causal relationship between sleep traits and endometriosis remains unclear. We aim to investigate the potential causal associations between sleep traits and endometriosis, including its specific anatomical subtypes. Two-sample Mendelian Randomization (MR) analyses were conducted. The primary exposure was insomnia, and the secondary exposures were getting up in the morning, nap during day, sleep disorders, chronotype, snoring, short sleep duration, and long sleep duration. The outcomes were endometriosis along with its anatomical subtypes: uterine, ovarian, vaginal, intestinal, peritoneal, and unspecified endometriosis. We applied inverse variance weighted (IVW), weighted median, simple mode, weighted mode, and MR-Egger regression to estimate causal relationships between sleep traits and the risk of endometriosis. Additionally, a series of sensitivity and validation analyses were performed to ensure robustness. For the primary exposure, IVW results suggested that insomnia was significantly associated with elevated risks of endometriosis (OR=1.83, 95% CI: 1.34–2.50), uterine endometriosis (OR=2.22, 95% CI: 1.30–3.80), ovarian endometriosis (OR=1.61, 95% CI: 1.03–2.54), vaginal endometriosis (OR > 1.00, 95% CI: 1.03–3.79) and unspecified endometriosis (OR=2.70, 95% CI: 1.45–5.02). For the secondary exposure, most sleep traits showed no significant associations with endometriosis risk. Our study demonstrated a causal relationship between insomnia and elevated risks of endometriosis across multiple anatomical subtypes. While no significant associations were observed for other sleep traits, the robust link with insomnia highlights the potential clinical value of sleep-focused interventions in mitigating the risk of endometriosis.
- Open Access
- 6 January 2026
- Published by Oxford University Press (OUP) in Human Reproduction Open
- Vol. 2026 (1)
Lack of difference in thyroid hormone profile between offspring conceived naturally and through ART
STUDY QUESTION: Does the thyroid function of children conceived through ART differ from that of naturally conceived (NC) children? SUMMARY ANSWER: Our study results indicate that both children conceived through fresh embryo transfer and frozen embryo transfer exhibit a thyroid function profile similar to that of NC children. WHAT IS KNOWN ALREADY: Few studies have compared thyroid function between ART-conceived and NC children, and conflicting conclusions have been reported. STUDY DESIGN, SIZE, DURATION: This study was based on a cohort at a tertiary health centre in China. The prospective cohort study included 9450 children born between 2005 and 2021 and followed up until June 2023. The outcomes of interest were the levels of the children’s thyroid-stimulating hormone (TSH), free triiodothyronine (FT3), free thyroxine (FT4), thyroglobulin autoantibody (A-TG), and thyroid peroxidase antibody (A-TPO). PARTICIPANTS/MATERIALS, SETTING, METHODS: The participants were 3940 fresh embryo transfer conceived children (ET group), 4730 frozen embryo transfer conceived children (FET group), and 780 NC children. Their ages ranged from 1.5 to 10 years. Generalized estimating equations (GEE) was used to compare the main outcomes. MAIN RESULTS AND THE ROLE OF CHANCE: No significant between-group differences were observed in FT4 levels or the odds of TSH, FT3, FT4, A-TG, or A-TPO levels beyond the standard range. Further subgroup analysis indicated that the ET group exhibited a slightly higher TSH level than the NC group during the toddler age period (mean TSH: 2.93 vs. 2.72 µIU/ml). However, this difference was no longer significant among preschool and school-aged children. Additionally, there were no significant differences in FT4 levels between the ART and NC groups across all age subgroups. LIMITATIONS, REASONS FOR CAUTION: Due to the inability to acquire urinary iodine data, we cannot adequately assess the influence of iodine intake on the study results. Both the unavoidable loss to follow-up in the cohort and the imbalance in the proportion of girls (higher in the NC group than in the ART group) may potentially introduce selection bias. WIDER IMPLICATIONS OF THE FINDINGS: The study results indicate that ART-conceived children exhibit a thyroid function profile comparable to that of NC children. Furthermore, ART appears to be safe for long-term thyroid function in children. STUDY FUNDING/COMPETING INTEREST(S): National Key Research and Development Program of China (no. 2024YFC2706700), the National Natural Science Foundation of China (NSFC) Regional Innovation and Development Joint Fund (U24A20664), the National Key Technology Research and Developmental Program of China (2024YFC2706902), CAMS Innovation Fund for Medical Sciences (2021-I2M-5-001), National Special Support Programme for High-level Talents, Taishan Scholars Programme for Young Experts of Shandong Province (tsqn202507388), Shandong Provincial Postdoctoral Innovation Project (no. SDCX-ZG-202502002), General Programme of Shandong Provincial Natural Science Foundation of China (ZR2022MH087), Young Talent of Lifting engineering for Science and Technology in Shandong, China (SDAST2024QTA071), Chinese Red Cross Foundation (2220025007). The authors declare no competing interests. TRIAL REGISTRATION NUMBER: N/A.
- 6 January 2026
- Published by Oxford University Press (OUP) in Human Reproduction Open
- Vol. 2026 (1)
Open AccessLack of difference in thyroid hormone profile between offspring conceived naturally and through ART
STUDY QUESTION: Does the thyroid function of children conceived through ART differ from that of naturally conceived (NC) children? SUMMARY ANSWER: Our study results indicate that both children conceived through fresh embryo transfer and frozen embryo transfer exhibit a thyroid function profile similar to that of NC children. WHAT IS KNOWN ALREADY: Few studies have compared thyroid function between ART-conceived and NC children, and conflicting conclusions have been reported. STUDY DESIGN, SIZE, DURATION: This study was based on a cohort at a tertiary health centre in China. The prospective cohort study included 9450 children born between 2005 and 2021 and followed up until June 2023. The outcomes of interest were the levels of the children’s thyroid-stimulating hormone (TSH), free triiodothyronine (FT3), free thyroxine (FT4), thyroglobulin autoantibody (A-TG), and thyroid peroxidase antibody (A-TPO). PARTICIPANTS/MATERIALS, SETTING, METHODS: The participants were 3940 fresh embryo transfer conceived children (ET group), 4730 frozen embryo transfer conceived children (FET group), and 780 NC children. Their ages ranged from 1.5 to 10 years. Generalized estimating equations (GEE) was used to compare the main outcomes. MAIN RESULTS AND THE ROLE OF CHANCE: No significant between-group differences were observed in FT4 levels or the odds of TSH, FT3, FT4, A-TG, or A-TPO levels beyond the standard range. Further subgroup analysis indicated that the ET group exhibited a slightly higher TSH level than the NC group during the toddler age period (mean TSH: 2.93 vs. 2.72 µIU/ml). However, this difference was no longer significant among preschool and school-aged children. Additionally, there were no significant differences in FT4 levels between the ART and NC groups across all age subgroups. LIMITATIONS, REASONS FOR CAUTION: Due to the inability to acquire urinary iodine data, we cannot adequately assess the influence of iodine intake on the study results. Both the unavoidable loss to follow-up in the cohort and the imbalance in the proportion of girls (higher in the NC group than in the ART group) may potentially introduce selection bias. WIDER IMPLICATIONS OF THE FINDINGS: The study results indicate that ART-conceived children exhibit a thyroid function profile comparable to that of NC children. Furthermore, ART appears to be safe for long-term thyroid function in children. STUDY FUNDING/COMPETING INTEREST(S): National Key Research and Development Program of China (no. 2024YFC2706700), the National Natural Science Foundation of China (NSFC) Regional Innovation and Development Joint Fund (U24A20664), the National Key Technology Research and Developmental Program of China (2024YFC2706902), CAMS Innovation Fund for Medical Sciences (2021-I2M-5-001), National Special Support Programme for High-level Talents, Taishan Scholars Programme for Young Experts of Shandong Province (tsqn202507388), Shandong Provincial Postdoctoral Innovation Project (no. SDCX-ZG-202502002), General Programme of Shandong Provincial Natural Science Foundation of China (ZR2022MH087), Young Talent of Lifting engineering for Science and Technology in Shandong, China (SDAST2024QTA071), Chinese Red Cross Foundation (2220025007). The authors declare no competing interests. TRIAL REGISTRATION NUMBER: N/A.
- Open Access
- 6 January 2026
- Published by Springer Nature in Lipids in Health and Disease
- Vol. 25 (1)
Offspring BMI and lipid profiles following assisted reproductive technology: a comparative study of underweight and normal-weight mothers
Background: The early developmental environment is associated with long-term metabolic health outcomes in children. A large proportion of underweight women tend to become pregnant via assisted reproductive technology (ART). However, the correlation between maternal prepregnancy underweight status and metabolic health in ART-conceived offspring remains poorly understood. The present study aimed to elucidate the associations of maternal underweight status with offspring development and lipid profiles, as well as to examine ART’s potential interaction in this process. Methods: The prospective cohort study included children conceived via ART at the Hospital for Reproductive Medicine Affiliated to Shandong University between April 2007 and August 2014. A total of 2,506 mother–offspring pairs, encompassing 3,496 visits, were analysed. Participants were divided into the following two groups based on the maternal body mass index (BMI): the normal weight group (18.5 ≤ maternal BMI < 24 kg/m 2 ) and the underweight group (maternal BMI < 18.5 kg/m 2 ). The BMI z-scores and serum lipid profiles of offspring aged 4–10 years were assessed. Results: Compared with those of controls, the offspring of underweight mothers had significantly lower BMI z-scores (mean difference: -0.55; 95% confidence interval [CI]: -0.71 to -0.38). Interaction analysis revealed no significant interaction effect between ART treatment (fertilization mode, embryo transfer stage, or embryo freezing) and maternal BMI status on offspring BMI and lipid profiles. However, a significant interaction effect was observed between maternal BMI status and offspring sex on offspring lipid profiles (interaction P values: 0.026 for total cholesterol and 0.021 for low-density lipoprotein cholesterol [LDL-c]). Compared with the control offspring, the female offspring of underweight mothers had higher total cholesterol levels (mean difference: 0.16; 95% CI: 0.02 to 0.29) and LDL-c levels (mean difference: 0.13; 95% CI: 0.02 to 0.25), while no such effects were found for male offspring (total cholesterol mean difference: -0.05; 95% CI: -0.17 to 0.09; LDL-c mean difference: -0.04; 95% CI: -0.15 to 0.07). Conclusions: In the present study, there was an association between prepregnancy maternal underweight status and reduced BMI among ART-conceived offspring, and significant sex-specific differences in offspring lipid profiles were identified. Moreover, the analysis excluded any interaction effects between ART procedures and maternal BMI status on these outcomes. These findings emphasize the association of prepregnancy maternal underweight status with lower BMI and altered lipid profiles in ART-conceived offspring, providing crucial insights for preconception care and targeted postnatal prevention strategies.
- 6 January 2026
- Published by Springer Nature in Lipids in Health and Disease
- Vol. 25 (1)
Open AccessOffspring BMI and lipid profiles following assisted reproductive technology: a comparative study of underweight and normal-weight mothers
Background: The early developmental environment is associated with long-term metabolic health outcomes in children. A large proportion of underweight women tend to become pregnant via assisted reproductive technology (ART). However, the correlation between maternal prepregnancy underweight status and metabolic health in ART-conceived offspring remains poorly understood. The present study aimed to elucidate the associations of maternal underweight status with offspring development and lipid profiles, as well as to examine ART’s potential interaction in this process. Methods: The prospective cohort study included children conceived via ART at the Hospital for Reproductive Medicine Affiliated to Shandong University between April 2007 and August 2014. A total of 2,506 mother–offspring pairs, encompassing 3,496 visits, were analysed. Participants were divided into the following two groups based on the maternal body mass index (BMI): the normal weight group (18.5 ≤ maternal BMI < 24 kg/m 2 ) and the underweight group (maternal BMI < 18.5 kg/m 2 ). The BMI z-scores and serum lipid profiles of offspring aged 4–10 years were assessed. Results: Compared with those of controls, the offspring of underweight mothers had significantly lower BMI z-scores (mean difference: -0.55; 95% confidence interval [CI]: -0.71 to -0.38). Interaction analysis revealed no significant interaction effect between ART treatment (fertilization mode, embryo transfer stage, or embryo freezing) and maternal BMI status on offspring BMI and lipid profiles. However, a significant interaction effect was observed between maternal BMI status and offspring sex on offspring lipid profiles (interaction P values: 0.026 for total cholesterol and 0.021 for low-density lipoprotein cholesterol [LDL-c]). Compared with the control offspring, the female offspring of underweight mothers had higher total cholesterol levels (mean difference: 0.16; 95% CI: 0.02 to 0.29) and LDL-c levels (mean difference: 0.13; 95% CI: 0.02 to 0.25), while no such effects were found for male offspring (total cholesterol mean difference: -0.05; 95% CI: -0.17 to 0.09; LDL-c mean difference: -0.04; 95% CI: -0.15 to 0.07). Conclusions: In the present study, there was an association between prepregnancy maternal underweight status and reduced BMI among ART-conceived offspring, and significant sex-specific differences in offspring lipid profiles were identified. Moreover, the analysis excluded any interaction effects between ART procedures and maternal BMI status on these outcomes. These findings emphasize the association of prepregnancy maternal underweight status with lower BMI and altered lipid profiles in ART-conceived offspring, providing crucial insights for preconception care and targeted postnatal prevention strategies.
- Open Access
- 20 November 2025
- Published by Springer Nature in Nature Medicine
- Vol. 31 (12) , 4314
Author Correction: Data-driven subtypes of polycystic ovary syndrome and their association with clinical outcomes
- 20 November 2025
- Published by Springer Nature in Nature Medicine
- Vol. 31 (12) , 4314
Open AccessAuthor Correction: Data-driven subtypes of polycystic ovary syndrome and their association with clinical outcomes
- Top Cited PapersOpen Access
- 29 October 2025
- Published by Springer Nature in Nature Medicine
- Vol. 31 (12) , 4214-4224
Data-driven subtypes of polycystic ovary syndrome and their association with clinical outcomes
Polycystic ovary syndrome (PCOS) is a common and heterogeneous endocrine disorder that affects 11%–13% of women worldwide, with profound implications for fertility and long-term metabolic health. Here we identify four reproducible subtypes—PCOS with hyperandrogen, with obesity, with high-sex hormone-binding globulin and with high-luteinizing hormone–anti-Müllerian hormone—through unsupervised clustering of 9 clinical variables in 11,908 affected women, validated across 5 international cohorts. Prospective 6.5-year follow-up and in vitro fertilization treatment data revealed distinct reproductive and metabolic trajectories: hyperandrogenic PCOS showed the highest risk of second trimester pregnancy loss and dyslipidemia incidence; PCOS with obesity exhibited the most severe metabolic complications, lowest live birth rates and highest PCOS remission rate; PCOS with high-sex hormone-binding globulin demonstrated favorable reproductive outcomes and the lowest incidence of diabetes and hypertension; and PCOS with high-luteinizing hormone–anti-Müllerian hormone had the greatest risk of ovarian hyperstimulation and the lowest PCOS remission rate. These findings advance understanding of PCOS heterogeneity and provide a framework for subtype-based risk stratification and personalized management.
- 29 October 2025
- Published by Springer Nature in Nature Medicine
- Vol. 31 (12) , 4214-4224
Top Cited PapersOpen AccessData-driven subtypes of polycystic ovary syndrome and their association with clinical outcomes
Polycystic ovary syndrome (PCOS) is a common and heterogeneous endocrine disorder that affects 11%–13% of women worldwide, with profound implications for fertility and long-term metabolic health. Here we identify four reproducible subtypes—PCOS with hyperandrogen, with obesity, with high-sex hormone-binding globulin and with high-luteinizing hormone–anti-Müllerian hormone—through unsupervised clustering of 9 clinical variables in 11,908 affected women, validated across 5 international cohorts. Prospective 6.5-year follow-up and in vitro fertilization treatment data revealed distinct reproductive and metabolic trajectories: hyperandrogenic PCOS showed the highest risk of second trimester pregnancy loss and dyslipidemia incidence; PCOS with obesity exhibited the most severe metabolic complications, lowest live birth rates and highest PCOS remission rate; PCOS with high-sex hormone-binding globulin demonstrated favorable reproductive outcomes and the lowest incidence of diabetes and hypertension; and PCOS with high-luteinizing hormone–anti-Müllerian hormone had the greatest risk of ovarian hyperstimulation and the lowest PCOS remission rate. These findings advance understanding of PCOS heterogeneity and provide a framework for subtype-based risk stratification and personalized management.