Background: Emergency cervical cerclage (ECC) is of potential value in twin pregnancy, when the cervix is dilated to >1cm. McDonald and Shirodkar were two main techniques of transvaginal cerclage at present. As ECC at extremely high risk of spontaneous preterm birth (sPTB) especially for twins with cervical dilated ≥ 3cm and prolapsed membranes, so which technique has more advantages is still uncertain. Objectives: The aim of our study was to evaluate the effectiveness of ECC performed with combined McDonald-Shirodkar technique in twin pregnancies between 18–26 weeks with painless cervical dilation 1-6cm. Methods: A retrospective, cohort study matched with the degree of cervical dilation was conducted. The study group (case group) included twin pregnancies who underwent combined McDonald-Shirodkar approach with cervical dilation ≥1 cm between 18–26 weeks of gestation at four institutions, from December 2015 to December 2022. To minimize confounding factors, we elucidated the causality structure using a DAG (Figure 1) and performed 1:1 case-control Matching. A control group performed McDonald approach. The primary outcome was gestational age (GA) at delivery. The secondary outcomes were pregnancy latency, the rates of sPTB at <28, <30, <32, <34 weeks, and neonatal outcomes. Additional sub-analysis was performed by dividing the patients into two subgroups of cervical dilation ≥ 3cm and < 3cm. Results: 84 twin pregnancies were managed with either combined McDonald-Shirodkar approach (case group: n=42) or McDonald approach (control group: n=42). Demographic characteristics were not significantly different in two groups(p>0.05). After adjusting for confounders which were represented by a directed acyclic graph (DAG, Figure 1), median GA at delivery was significantly higher (30.5 vs 27 weeks, Bate: 3.40, 95% confidence interval (CI): 2.13-4.67, p<0.001) and median pregnancy latency was significantly longer (56 vs 28 days, Bate: 24.04, 95% CI: 13.31-34.78, p<0.001) in the case group compared with the control group. Rates of sPTB at <28, <30, <32, and <34 weeks were significantly lower in the case group than in the control group. For neonatal outcomes, there were higher birth weight (BW) (1543.75 vs 980g, Bate: 420.08, 95%CI: 192.18-647.98, p<0.001) and significantly lower overall perinatal mortality (7.1% vs 31%, aOR: 0.16, 95% CI: 0.04-0.70, p=0.014) in the case group compared with the control group. And when cervical dilation ≥ 3cm, combined McDonald-Shirodkar procedure can significantly reduce perinatal mortality (8.3% vs 46.7%, aOR:0.09, 95%CI: 0.01-0.77, p=0.028), significantly decrease the risk of delivery at <28, <30weeks, prolong GA at delivery and pregnancy latency compared with McDonald procedure. Conclusions: ECC performed with the combined McDonald-Shirodkar procedure in twin pregnancies with cervical dilation 1-6 cm in mid-trimester pregnancy may reduce the rate of sPTB and improve perinatal and neonatal outcomes compared with McDonald procedure, especially for twins with cervical dilation of 3-6 cm and prolapsed membranes.

It is important to monitor labor progress and identify the possibility of vaginal birth. Based on intrapartum sonographic findings, we describe the prediction models for evaluating fetal occiput rotation and vaginal delivery. We performed serial intrapartum ultrasonography in nulliparous women with singleton cephalic presentation at term in the latent phase and every three hours after that. Six hundred and fourteen women were included in our study, of whom 524 underwent vaginal delivery, and 90 required cesarean section. The percentage of women with fetuses in non-occiput anterior(non-OA) position at the latent phase was 53.9% (331 cases), as 257(77.6%) women underwent spontaneous rotation to OA position before delivery, 74 (22.4%) women were with persistent occiput posterior (OP) position or occiput transverse (OT) position. We developed a model based on the maternal height and middle angle to predict the spontaneous fetal occiput rotation, with the C statistic was 0.667 (95%CI 0.583-0.751). Moreover, a prediction model based on the maternal height and angle of progression to evaluate whether women underwent vaginal delivery was also developed, of which the C statistic was 0.738(95% CI: 0.763-0.793). Both models showed satisfactory calibration. Our approach provides helpful information for predicting vaginal delivery and internal fetal occiput rotation.