基于胎儿头围、腹围和产程进展角建立简单模型对阴道助产术难度的预测价值

A Simple Model Based on Fetal Head Circumference,Fetal Abdominal Circumference and Angle of Progression to Predict the Difficulty of Operative Vaginal Delivery

目的:探讨联合使用胎儿头围、腹围和产程进展角(angle of progression,AOP)预测阴道助产术(胎头吸引术和产钳助产术)难度的价值。方法:采用前瞻性观察性研究,选取2017年5月—2022年7月在广西壮族自治区人民医院产科行阴道助产术的足月单胎初产妇共64例。收集孕妇入院后的临床资料,超声测量胎儿头围和腹围,助产前采用经会阴超声测量AOP(宫缩间期和宫缩期AOP均值分别记为AOP1、AOP2)。按助产结局分组,有下列1种及以上情况纳入困难或失败组(A组):需要3次及以上牵引;中途更换助产器械后助产牵引成功(胎头吸引术失败转产钳助产术);阴道助产失败转剖宫产终止妊娠;术者主观评估助产失败可能性高,行剖宫产终止妊娠;新生儿产伤;会阴Ⅲ度或Ⅳ裂伤。其余纳入成功且容易组(B组)。将胎儿头围、腹围、AOP1和AOP2进行不同组合建立4个多因素Logistic回归模型,使用受试者操作特征(receiver operator characteristic,ROC)曲线评估4个模型对助产难度的区分度,选出ROC曲线下面积(area under the curve,AUC)最大者。Hosmer-Lemeshow检验评估最优模型的校准度,并利用Bootstrap重复取样(1000次)的方法进行内部验证。结果:2组间阴道指检胎头位置、末次牵引时长、末次牵引不同初始牵引方向占比、胎儿腹围、AOP1、AOP2、脐动脉血气pH值和新生儿窒息率比较差异有统计学意义(均P<0.05)。同时纳入胎儿头围、胎儿腹围、AOP1和AOP2建立的模型AUC最大(AUC=0.917,95%CI:0.821~0.972,Z=11.676,P<0.001),Hosmer-Lemeshow检验显示χ2=5.873,P=0.661,模型区分度和校准度均良好。内部验证结果表明模型有较好的一致性。结论:联合应用胎儿头围、腹围、AOP1和AOP2建立的简单模型能较好地预测阴道助产术难度,为临床决策提供一定依据。

Objective:To investigate the predictive value of joint use of fetal head circumference,fetal abdominal circumference and angle of progression(AOP)for the difficulty of operative vaginal delivery(forceps and vaccum).Methods:This prospective observational study was carried out in Department of Obstetrics of the People′s Hospital of Guangxi Zhuang Autonomous Region from May 2017 to July 2022.A total of 64 nulliparous women with singleton term pregnancies who required instrumentation to expedite vaginal delivery during the second stage of labor were included.Clinical data of pregnant women were collected after admission to the hospital.Fetal head circumference and fetal abdominal circumference were measured.Before performing operative vaginal delivery,AOP at rest and that during uterine contraction were measured by transperineal ultrasound(AOP1:mean of AOPs at rest,AOP2:mean of AOPs during uterine contraction).According to the outcomes of operative vaginal delivery,a case was included into"difficult or failed group"(Group A)when one or more of the following situations occurred:three or more tractions needed;successful operative vaginal delivery requiring sequential instruments(change vacuum to forceps);failed operative vaginal delivery;cesarean delivery due to a subjective impression of a difficult or failed traction;birth injury and the third-or the fourth-degree perineal tear.The rest cases were included into"successful and easy group"(Group B).Fetal head circumference,fetal abdominal circumference,AOP1 and AOP2 were used to establish four different predictive models by employing multivariable logistic regression.Receiver operator characteristic(ROC)curve was used to evaluate the discrimination of four models and the model with the largest area under ROC curve(AUC)was considered as the best one.Hosmer-Lemeshow test assessed the calibration of the optimal model.The optimal one was internally validated by using Bootstrap replicated sampling(1000 times)method.Results:Statistically significant differences were revealed with regard to fetal head station by vaginal digital examination,duration of the last traction,initial direction of pulling in the last traction,fetal abdominal circumference,AOP1,AOP2,umbilical cord artery pH and rate of neonatal asphyxia(all P<0.05).A predictive model basing on the combination of fetal head circumference,fetal abdominal circumference,AOP1 and AOP2 showed the largest AUC(AUC=0.917,95%CI:0.821-0.972,Z=11.676,P<0.001).The Hosmer-Lemeshow test showed thatχ2=5.873,P=0.661.Both of discrimination and calibration were good.The model showed good consistency in the internal validation.Conclusions:The simple model based on the combination of fetal head circumference,fetal abdominal circumference,AOP1 and AOP2 showed a predictive value for the difficulty of operative vaginal delivery and provided some basis for clinical decision-making.