Zhang You, Qin Ningxin, Hu Jing, Bai Jie, Pan Mengjia, Xu Yan, Huang Xin, Wang Ke
Information Center, Shanghai Key Laboratory of Maternal Fetal Medicine, Shanghai Institute of Maternal-Fetal Medicine and Gynecologic Oncology, Shanghai First Maternity and Infant Hospital, School of Medicine, Tongji University, Shanghai, China.
Center of Reproductive Medicine, Shanghai Key Laboratory of Maternal Fetal Medicine, Shanghai Institute of Maternal-Fetal Medicine and Gynecologic Oncology, Shanghai First Maternity and Infant Hospital, School of Medicine, Tongji University, Shanghai, China.
Front Endocrinol (Lausanne). 2025 Aug 21;16:1585144. doi: 10.3389/fendo.2025.1585144. eCollection 2025.
To establish and validate a nomogram model for the quality of sleep in patients with recurrent implantation failure (RIF) and to evaluate its performance.
From January 2023 to June 2023, 484 RIF patients who underwent ART fertilization treatment at the Reproductive Medicine Center of Tongji University-affiliated Obstetrics and Gynecology Hospital were selected as the modeling set and internal validation. Additionally, from July to September 2023, 223 RIF patients who underwent ART fertilization treatment at the Reproductive Medicine Center of Tongji University-affiliated Obstetrics and Gynecology Hospital were chosen as the external validation set. Their clinical data was collected. Lasso regression was used to screen potential predictive variables and multifactor logistic regression analysis was conducted to determine the final predictors. A nomogram model was established, and the model was evaluated using methods such as plotting receiver operating characteristic (ROC) curves, calibration curves, Hosmer-Lemeshow goodness of fit test, and decision curve analysis.
Through Lasso regression and multifactor logistic regression, 7 predictors were identified, including FSH, E2, depression mood (moderate, severe), daily exercise time, sun exposure, caffeine intake, and shift work (>16h/w) for constructing the nomogram model. The AUC for the modeling set was 0.971 (95%CI:0.952∼0.989), for the internal validation set was 0.960 (95%CI:0.937∼0.979), and for the external validation set was 0.850 (95%CI:0.739∼0.960), indicating good predictive performance of the model.
This study established and validated a nomogram model composed of 7 clinical indicators for sleep disorders in RIF patients. The predictors include both physiological indicators and daily lifestyle habits, demonstrating significant predictive value and clinical application efficiency. It can be used for early identification of potential sleep disorders in RIF patients, providing certain reference significance for clinical work.
建立并验证复发性植入失败（RIF）患者睡眠质量的列线图模型，并评估其性能。
选取2023年1月至2023年6月在同济大学附属第一妇婴保健院生殖医学中心接受辅助生殖技术（ART）受精治疗的484例RIF患者作为建模集和内部验证集。此外，选取2023年7月至9月在同济大学附属第一妇婴保健院生殖医学中心接受ART受精治疗的223例RIF患者作为外部验证集。收集他们的临床资料。采用Lasso回归筛选潜在预测变量，并进行多因素逻辑回归分析以确定最终预测因素。建立列线图模型，并使用绘制受试者操作特征（ROC）曲线、校准曲线、Hosmer-Lemeshow拟合优度检验和决策曲线分析等方法对模型进行评估。
通过Lasso回归和多因素逻辑回归，确定了7个预测因素，包括促卵泡生成素（FSH）、雌二醇（E2）、抑郁情绪（中度、重度）、每日运动时间、日照时间、咖啡因摄入量和轮班工作（每周>16小时），用于构建列线图模型。建模集的曲线下面积（AUC）为0.971（95%置信区间：0.952～0.989），内部验证集为0.960（95%置信区间：0.937～0.979），外部验证集为0.850（95%置信区间：0.739～0.960），表明模型具有良好的预测性能。
本研究建立并验证了一个由7个临床指标组成的RIF患者睡眠障碍列线图模型。预测因素包括生理指标和日常生活习惯，具有显著的预测价值和临床应用效率。可用于早期识别RIF患者潜在的睡眠障碍，为临床工作提供一定的参考意义。
