探讨早孕期孕妇药物流产失败的影响因素及构建相关预测模型，评估其对早孕期孕妇药物流产成功的预测价值。

选择2020年1月至2022年12月于盐城市中医院采取药物流产的323例早孕期孕妇为研究对象。采取回顾性分析方法，根据药物流产结局将其分为失败组(n＝35)及成功组(n＝288)。采用单因素分析及多因素非条件logistic回归分析法，分析早孕期孕妇药物流产失败的影响因素，并构建相关预测模型，采用受试者工作特征(ROC)曲线，评估该模型对早孕期孕妇药物流产成功的预测价值。本研究符合2013年修订的《世界医学协会赫尔辛基宣言》相关要求。受试者对诊疗均知情同意，并签署知情同意书。

①单因素分析结果显示，受试者孕产次(≤1次、>1次)，既往流产史，子宫位置(水平位、非水平位)，孕囊直径(孕囊3径线平均值)(<20 mm、≥20 mm)及慢性盆腔炎史，均为导致早孕期孕妇药物流产失败的可能影响因素。②多因素非条件logistic回归分析结果显示，既往流产史(OR＝7.791，95%CI：3.234～18.767，P<0.001)，子宫位置非水平位(OR＝4.614，95%CI：1.990～10.702，P<0.001)及孕囊直径≥20 mm(OR＝6.564，95%CI：2.709～15.905，P<0.001)，均为导致早孕期孕妇药物流产失败的独立危险因素。③基于上述导致早孕期孕妇药物流产失败的3项独立危险因素，构建预测模型，评估其对于早孕期孕妇药物流产成功的ROC的曲线下面积(AUC)、敏感度、特异度分别为0.846、95.7%、90.6%，均高于模型中3项因素任意一项预测的AUC、敏感度及特异度，并且该预测模型的拟合度较好(Hosmer-Lemeshow χ²＝2.02，P＝0.732)。

基于既往流产史、子宫位置、孕囊直径，3项导致早孕期孕妇药物流产失败的独立危险因素，构建的相关预测模型，对于评估早孕期孕妇药物流产成功具有一定预测价值。

To explore the influencing factors of medical abortion failure of pregnant women in early pregnancy and constructing a related prediction model, and to evaluate its predictive value for medical abortion success of pregnant women in early pregnancy.

A total of 323 women with medical abortion in early pregnancy who were treated in the outpatient clinic department of Yancheng TCM Hospital from January 2020 to December 2022, were included in the study. Took a retrospective analysis method, based on outcomes of medical abortion, they were divided into failure group (n=35) and success group (n=288). Univariate analysis and multivariate unconditional logistic regression analysis were utilized to analyze the influencing factors of medical abortion failure of pregnant women in early pregnancy, and to construct a related predictive model for the success of medical abortion of pregnant women in early pregnancy. Receiver operating characteristic (ROC) curve analysis was conducted to evaluate the predictive value of this model for the success of medical abortion of pregnant women in early pregnancy. This study was in line with World Medical Association Declaration of Helsinki revised in 2013. Participants were granted their informed consent for diagnosis and treatments, substantiated by signing informed consent forms.

①The results of univariate analysis showed that the number of pregnancies and childbirth (≤1 time, >1 time) of subjects, history of previous abortions, uterine position (horizontal, non-horizontal), gestational sac diameter (average of the three diameters of the gestational sac)(<20 mm, ≥20 mm), and history of chronic pelvic inflammatory disease could be potential influencing factors of the failure of medical abortion of pregnant women in early pregnancy. ②The results of multivariate unconditional logistic regression analysis showed that history of previous abortions (OR=7.791, 95%CI: 3.234-18.767, P<0.001), non-horizontal uterine position (OR=4.614, 95%CI: 1.990-10.702, P<0.001), and gestational sac diameter ≥20 mm (OR=6.564, 95%CI: 2.709-15.905, P<0.001) were all independent risk factors for the failure of medical abortion of pregnant women in early pregnancy. ③ The area under curve (AUC) of ROC, sensitivity and specificity of the prediction model for the success of medical abortion of pregnant women in early pregnancy based on the above three independent risk factors were 0.846, 95.7% and 90.6% respectively, which were higher than those of predicted by any one of the three factors, and this prediction model has a good fitting degree (Hosmer-Leme showed χ²=2.02, P=0.732).

The prediction model constructed based on three independent risk factors for the failure of medical abortion of pregnant women in early pregnancy including the history of previous abortions, uterine position and gestational sac diameter, holds certain predictive value for the success of medical abortion of pregnant women in early pregnancy.