Wang X, Fan Y Y, Li L, Zhang S D, Zhang C L
Reproductive Medicine Institute, Henan Provincial People's Hospital, Zhengzhou 450003, China.
Zhonghua Fu Chan Ke Za Zhi. 2022 Feb 25;57(2):110-116. doi: 10.3760/cma.j.cn112141-20210930-00561.
To explore the related factors of poor ovarian response (POR) in patients receiving controlled ovarian stimulation (COS) and to establish the nomogram for predicting POR in patients who received in vitro fertilization or intracytoplasmic sperm injection (IVF/ICSI). In this retrospective research, clinical data of 17 164 cycles of patients who received IVF/ICSI treatment at Henan Provincial People's Hospital from September 1st, 2016 to September 1st, 2020 were analyzed. Independent correlative factors affecting the occurrence of POR were screened by logistic regression, which were the model enrollment variables in the prediction model. Totally 13 266 cycles with well-record of enrollment variables were screened, and these data were randomly divided into model group (9 896 patients) and validation group (3 370 patients) according to 3∶1. The nomogram was established according to the regression coefficient of the relevant variables. The prediction accuracy of the nomogram was evaluated by calculating area under the receiver operating characteristic curve (AUC). Multivariate logistic regression analysis showed age, infertility type, body mass index, anti-Müllerian hormone, basal follicle stimulating hormone, basal estrogen, antral follicle number, previous times of POR, history of ovarian surgery, ovulation stimulation protocol and average amount of gonadotropin were independent correlative factors affecting the occurrence of POR (all <0.05). In the model group, according to the above factors, the prediction model and nomogram of POR risk were constructed and the validation group verified the model. The AUC of the model group was 0.893 (95%: 0.885-0.900), and the AUC of the validation group was 0.890 (95%: 0.878-0.903). The influencing factors of POR after COS in patients treated by IVF/ICSI are screened, and the nomogram for predicting POR established in this study is proved to be effective, simple, intuitive and clear in predicting the occurrence of POR.
探讨接受控制性卵巢刺激（COS）患者卵巢反应不良（POR）的相关因素，并建立预测接受体外受精或卵胞浆内单精子注射（IVF/ICSI）患者POR的列线图。在这项回顾性研究中，分析了2016年9月1日至2020年9月1日在河南省人民医院接受IVF/ICSI治疗的17164个周期患者的临床资料。通过逻辑回归筛选影响POR发生的独立相关因素，这些因素为预测模型中的模型纳入变量。共筛选出13266个纳入变量记录完整的周期，这些数据按照3∶1随机分为模型组（9896例患者）和验证组（3370例患者）。根据相关变量的回归系数建立列线图。通过计算受试者操作特征曲线下面积（AUC）评估列线图的预测准确性。多因素逻辑回归分析显示，年龄、不孕类型、体重指数、抗苗勒管激素、基础卵泡刺激素、基础雌激素、窦卵泡数、既往POR次数、卵巢手术史、促排卵方案和促性腺激素平均用量是影响POR发生的独立相关因素（均P<0.05）。在模型组中，根据上述因素构建了POR风险预测模型和列线图，并在验证组中对模型进行了验证。模型组的AUC为0.893（95%CI：0.8850.900），验证组的AUC为0.890（95%CI：0.8780.903）。筛选出IVF/ICSI治疗患者COS后POR的影响因素，本研究建立的预测POR的列线图在预测POR发生方面有效、简单、直观且清晰。
