The Predictive Value of the Metabolic Score for Insulin Resistance (Mets-IR) for Metabolism-Related Disorders and Fertility Outcomes in Chinese Women with Polycystic Ovary Syndrome

Fengjuan Lu¹Hong Yu²Baichao Shi¹Jiaxing Feng³Yang Liu²Muxin Guan¹Jiannan Yu¹Zhuwei Gao¹Jingshu Gao²Hongli Ma³Yu Wang³Jing Cong³Xiaoke Wu^2*

• ¹Heilongjiang University of Chinese Medicine, Harbin, China
• ²Zhejiang Provincial Hospital of Chinese Medicine, Hangzhou, China
• ³First Affiliated Hospital of Heilongjiang University of Traditional Chinese Medicine, Harbin, China

Objective: This study aimed to evaluate the predictive value of the Metabolic Score for Insulin Resistance (Mets-IR) for metabolism-related disorders and its association with hormonal status and fertility outcomes in Chinese women with Polycystic Ovary Syndrome (PCOS). Methods: This secondary analysis included 957 women from the PCOSAct trial. Participants were stratified by Mets-IR quartiles. Linear regression analyzed correlations between Mets-IR and metabolic/hormonal parameters. ROC curves assessed Mets-IR's predictive performance for metabolic disorders. Multivariable logistic regression evaluated associations with fertility outcomes. Results: Significant linear correlations were observed between Mets-IR and key metabolic parameters (e.g., HOMA-IR, TG, WHR) and hormonal parameters (PG, E2, FT, LH, FSH, LH/FSH ratio, SHBG, FAI, AMH) (all P < 0.05). Specifically, Mets-IR was negatively correlated with PG, E2, LH, SHBG and AMH, and positively correlated with FT and FAI. After adjusting for age and BMI, Mets-IR remained significantly negatively associated with LH, FSH, SHBG and AMH (all P < 0.01), while it showed a significant positive association with FAI (P < 0.001) and a significant negative association with TT (P < 0.05). Mets-IR exhibited superior independent predictive ability compared to BMI for key hormonal parameters, including SHBG (ΔR² = 17.2% vs 13.7%), FAI (ΔR² = 16.8% vs 14.9%), and AMH (ΔR² = 3.9% vs 2.8%). ROC analysis demonstrated high predictive performance of Mets-IR for IR (AUC=0.814), MetS (AUC=0.878), and NAFLD (AUC=0.818). In predicting ovulation, Mets-IR demonstrated moderate predictive performance comparable to BMI (AUC=0.606), with an optimal cut-off value of 23.46. Mets-IR was negatively associated with ovulation but not with other fertility outcomes (conception,clinical pregnancy,live birth and pregnancy loss). Conclusion: In women with PCOS, Mets-IR demonstrates significant associations with both metabolic and hormonal parameters and exhibits superior predictive ability over BMI for key hormonal markers (SHBG, FAI, AMH). This index serves as an effective non-invasive predictor for IR, MetS, and NAFLD.For the assessment of ovulatory dysfunction, its predictive performance is comparable to that of BMI, yet it demonstrates no significant association with other fertility outcomes.